Char, Roll for Outcome

config.style.googleFont: '<link href="https://fonts.googleapis.com/css2?family=Bona+Nova:ital,wght@0,400;0,700;1,400&display=swap" rel="stylesheet">' config.style.page.font: 'Bona Nova/serif 10' config.style.page.header.font: "20" config.style.page.footer.font: 'Bona Nova/serif normal' config.style.page.footer.font: "10" config.style.page.footer.link.font: "normal" config.style.page.footer.link.active.font: "normal" config.style.page.style: "shadow" config.style.page.theme.enableSwitching: true config.style.page.theme.override: 'light' config.style.backdrop: "#DFD8D1" config.style.page.color: "black on #F0EEE6" config.style.page.link.color: "black" config.style.page.link.lineColor: "black" config.style.page.link.active.color: "#5e5d59 on #F0EEE6" config.style.page.link.active.lineColor: "#5e5d59 on #F0EEE6" config.style.dark.backdrop: "teal-5" config.style.dark.page.color: "gray-1 on gray-6" config.style.dark.page.link.color: "white" config.style.dark.page.link.lineColor: "teal-3" config.style.dark.page.link.active.color: "aquamarine-2 on gray-6" config.style.dark.page.link.active.lineColor: "teal-5" config.style.dark.page.header.border: 'thin-line' config.body.transition.name: 'crossfade' config.style.fontScaling.baseViewportWidth: 750 config.style.page.verticalAlign: 'center' config.style.page.footer.link.color: "black" config.style.page.footer.link.lineColor: "black" config.style.page.footer.link.hover.color: "#4A4A4A" config.style.page.footer.link.hover.lineColor: "#4A4A4A" config.style.page.footer.link.active.color: "#777777 on #F0EEE6" config.style.page.footer.link.active.lineColor: "black" config.style.page.verticalAlign: 'centre' config.footer.center: "" config.footer.left: "{theme switcher, darkLabel: 'CRO Guide', lightLabel: 'CRO Guide'}" config.footer.right: "{link to: 'About', label: 'About'} | {restart link}" -- <div class="cro-title-smaller" style="margin-top: 1em;">{Char}, [Roll]</div> <div class="cro-title-connector">for</div> <div class="cro-title-big">[Outcome]</div> <div class="cro-title-subtitle" style="margin-bottom: 1.5em;"> A simple system for guiding AI narratives with a dice. </div> <div class="cro-passage-nav" style="margin-top: 2.3em;"> <div class="cro-title-nav"> [[Quick Reference Hub]] <span class="nav-description">Fast lookup of any rule or scale.</span> [[Learn the System->Getting Started]] <span class="nav-description">Guided tutorial progression.</span> [[Full Text Version->CRO Guide Full Text]] <span class="nav-description">View the entire guide on a single page.</span> [[CRO Helper Script->CRO Helper Script Detail]] <span class="nav-description">{NEW} Browser extension for supported chat platforms </span> </div>

config.footer.center: "" config.footer.right: "{link to: 'Title Screen', label: 'Home'} | {restart link}" config.footer.left: "{theme switcher, darkLabel: 'CRO Guide', lightLabel: 'CRO Guide'}" -- <h1 class="cro-h1" style="text-align:center;">Quick Reference Hub</h1> <p style="text-align:center; margin-bottom:1.5em;">Find any CRO rule, scale, or extension quickly. Click to jump to the relevant section.</p> <div class="cro-hub-container"> <div class="cro-hub-section"> <h3 class="cro-h3">Core & Setup</h3> <ul> <li>[[Where to Use CRO->CRO Where to use Detail]]</li> <li><strong>[[The Basic CRO Format->CRO Basic Action Detail]]</strong></li> <li><strong>[[The Main Outcome Scale->CRO Main Outcome Scale Detail]]</strong></li> <li>[[Difficulty Modifiers->CRO Difficulty Modifiers Detail]]</li> <li>[[Narrative Consequences->CRO Narrative Consequences Detail]]</li> <li>[[Interpreting the Roll (Main vs. Specific)->CRO Interpreting Roll Detail]]</li> <li><strong>[[CRO Helper Browser Script->CRO Helper Script Detail]]</strong></li> </ul> <h4 class="cro-h4" style="font-size:0.9em; margin-top:0.8em; margin-bottom:0.3em;">Time Progression:</h4> <ul> <li>[[Simple Time Declaration->CRO Time Simple Declaration Detail]]</li> <li>[[Random Time Duration->CRO Time Random Duration Detail]]</li> <li>[[Time & Fortune Scale->CRO Time TimeAndFortune Detail]]</li> </ul> </div> <div class="cro-hub-section"> <h3 class="cro-h3">Skill-Specific Scales</h3> <ul> <li>[[Physical Scale->CRO Physical Scale Detail]]</li> <li>[[Manual Scale->CRO Manual Scale Detail]]</li> <li><strong>[[Mental Scale->CRO Mental Scale Detail]]</strong></li> <li>[[Social Scale->CRO Social Scale Detail]]</li> <li><strong>[[Perception Scale->CRO Perception Scale Detail]]</strong></li> <li>[[Fortune Scale->CRO Fortune Scale Detail]]</li> <li>[[Stealth Scale->CRO Stealth Scale Detail]]</li> </ul> </div> <div class="cro-hub-section"> <h3 class="cro-h3">Extensions: Action & Outcome</h3> <h4 class="cro-h4" style="font-size:0.9em; margin-bottom:0.3em;">Action Modifiers:</h4> <ul> <li>[[Stakes->CRO Stakes Detail]]</li> <li>[[Pushing Limits->CRO Pushing Limits Detail]]</li> <li>[[Skill Leverage->CRO Skill Leverage Detail]]</li> <li>[[Serendipity->CRO Serendipity Detail]]</li> <li>[[Targeted Effort->CRO Targeted Effort Detail]]</li> <li>[[Guarded Approach->CRO Guarded Approach Detail]]</li> </ul> <h4 class="cro-h4" style="font-size:0.9em; margin-top:0.8em; margin-bottom:0.3em;">Outcome Details (Investigation):</h4> <ul> <li>[[Item Focus->CRO Item Focus Detail]]</li> <li>[[Inventory Check->CRO Inventory Check Detail]]</li> </ul> </div> <div class="cro-hub-section"> <h3 class="cro-h3">Extensions: Environment & Meta</h3> <h4 class="cro-h4" style="font-size:0.9em; margin-bottom:0.3em;">Environmental Factors:</h4> <ul> <li>[[Random Events->CRO Random Events Detail]]</li> <li>[[Emotional Static->CRO Emotional Static Detail]]</li> <li>[[Interference/Nemesis->CRO Interference Detail]]</li> </ul> <h4 class="cro-h4" style="font-size:0.9em; margin-top:0.8em; margin-bottom:0.3em;">Meta-Narrative Tools:</h4> <ul> <li>[[Montage->CRO Montage Detail]]</li> <li>[[Narrative Techniques->CRO Narrative Techniques Detail]]</li> <li>[[Narrative Momentum->CRO Narrative Momentum Detail]]</li> </ul> </div> <div class="cro-hub-section"> <h3 class="cro-h3">Advanced & Supporting</h3> <ul> <li>[[Judge System (Overview & Setup)->CRO Advanced Tools & Examples]]</li> <li>[[Judge AI - Base Prompt->CRO Judge AI Base Prompt Detail]]</li> <li>[[Judge AI - Scales Module->CRO Judge AI Scales Module Detail]]</li> <li>[[Judge AI - Extensions Module->CRO Judge AI Extensions Module Detail]]</li> <li>[[Implementation Examples->CRO Implementation Examples Detail]]</li> <li>[[Notes & Best Practices->CRO Notes and Best Practices Detail]]</li> </ul> </div> </div> <div class="cro-passage-nav"> <div><span class="nav-next">[[Start Learning the System->Getting Started]]</span></div> <div><span class="nav-hub">[[Title Screen]]</span></div> </div>

config.footer.right: "{link to: 'Title Screen', label: 'Home'} | {restart link}" -- <h1 class="cro-h1">Getting Started with CRO</h1> <h2 class="cro-h2">1.0 Introduction</h2> Ever find an AI chatbot a bit too... agreeable? You state an intention, and the AI complies smoothly, missing chances for dramatic tension or unexpected turns. The <span class="cro-inline-code">{Char}, [Roll] for [Outcome]</span> system, or <span class="cro-inline-code">CRO</span> for short, aims to fix this for you. <div class="cro-guide-image-container cro-image-fullwidth"> {embed image: 'Images/Introduction Part 1.png', alt: 'The user and Roxy having a bit of a misunderstanding.'} </div> Here’s how it works: when your character attempts to do something, **you** roll a {tooltip: 'd10', text: 'A ten-sided die.'} and present the result using the simple <span class="cro-inline-code">CRO</span> format provided below. The AI then uses that number and the directions provided inside the Outcome Scales, to guide its response. <div class="cro-guide-image-container cro-image-fullwidth"> {embed image: 'Images/Introduction Part 2.png', alt: 'The CRO guide is used, and once again all is good in the world.'} </div> Suddenly, success isn't automatic. Your roll might lead to triumph, a costly setback, or something in between – all fascinating narrative threads not unlike the ones found through the medium of tabletop roleplaying games. The AI in control of your roleplaying session - your very own artificial DM - is encouraged to explore far more interesting narrative paths than just going along with its own vague interpretation of what you may like or don’t. There are no complex character sheets, inventories or statistics to track; your <span class="cro-inline-code">{Char}</span> would not be able to effectively remember any of those. Not for long, anyway. And this is not meant to be homework! So it's just you. You, your character's intent, a straightforward d10 roll, and a clear format to guide the story. <div class="cro-guide-image-container"> {embed image: 'Images/Introduction Final.png', alt: 'Roxy and the user wish you good luck!'} </div> <h2 class="cro-h2" style="margin-top:1.5em;">1.1 Where to use</h2> CRO works with any AI that can engage in conversational roleplay, and requires absolutely no setup on the AI's side. No special system prompts, no configuration - you simply start using the format and the AI will naturally adapt to it. If you’re in need of a d10 dice, just search “Dice Roller” into google, there’s an easter egg for it. I've primarily developed and tested this system on **character roleplay platforms**: first on Character.ai (that was like a year ago, when it was still kind of decent!), then Spicychat.ai and Janitor.com. These platforms are designed around chatting with specific AI personas in ongoing conversations, which tends to work particularly well with CRO's narrative focus. There are also some good world-based “characters" there too: chatbots designed to create a world for your own character to work and act inside of. You'll find plenty of them to experiment with. There’s also **general AI assistants** like ChatGPT, Claude, Gemini, or Deepseek, of course - those also work just fine. You can start a conversation by describing your character and setting, then begin using the CRO format for actions. These platforms often bring more sophisticated reasoning to the table, though they may need a sentence or two of context about what you're doing initially. Point is, you're not changing how the AI works - you're changing how you present information to it. The format itself teaches the AI what you want: that dice results should guide outcomes, that low rolls mean complications, and that high rolls mean smooth success. Most AI systems pick this up within a few exchanges, if not right away. I personally lean towards character platforms because they're built and fine-tuned for sustained narrative interaction, but honestly, experimentation is part of the fun. Try it wherever you normally interact with LLMs. See how it feels! <div class="cro-hr"></div> <h2 class="cro-h2">1.2 Quickstart Guide</h2> <div class="cro-guide-image-container"> {embed image: 'Images/Basic Action and Rolling.png', alt: 'A d10 dice is thrown - a 5. The CRO format is used accordingly.'} </div> Want to jump right in? Here's everything you need to start using CRO: <h3 class="cro-h3">The Basic Process:</h3> 1. **Declare what you're attempting** - not what you succeed at, but what you're trying to do 2. **Roll a d10** - this determines how things actually go 3. **Apply any difficulty modifiers mentally** (Easy: +1, Difficult: -1, etc.) 4. **Find your result on the ↪ [[Main Outcome Scale->CRO Core Mechanics]]** (Detailed in the next section) 5. **Present it to the AI using this format:** [code] *I attempt to [action].* **[SKILL] ROLL** *[Final Result]/10 - [Outcome Description]* [continue] <h3 class="cro-h3">Example</h3> You want to climb a crumbling tower wall. Instead of just saying "I climb up," you declare your attempt and roll: [code] *I attempt to scale the crumbling tower wall.* **PHYSICAL ROLL** *4/10 - Setback Failure (no, but... with silver lining)* [continue] Now the AI has clear direction: you don't make it up, but something interesting happens. Maybe you fall but discover a hidden entrance, or attract the attention of someone who can help. **That's it!** You've just turned a simple success into a dramatic story moment. The AI will handle the rest based on your roll. Instead of the AI guessing what you want or defaulting to success, the dice created genuine uncertainty. That 4/10 could have been a 9/10 triumph or a 1/10 disaster - and now you're both curious to see how this setback becomes the next part of your story. <div class="cro-hr"></div> <h2 class="cro-h2">1.3 Browser Script</h2> For users of Spicychat.ai, Character.ai, or JanitorAI, there's a ↪ [[browser script->CRO Helper Script Detail]] that provides a convenient graphical interface directly on those platforms. It handles all the formatting and dice rolling automatically, so you can focus on the storytelling rather than the mechanics. The script is completely optional - everything in this guide works perfectly well with manual rolling and copy-pasting. But if you're spending a lot of time on those platforms, it might save you some effort. <div class="cro-hr"></div> For skill categories, extensions, and advanced techniques, keep reading. But this basic approach will immediately make your AI conversations more dramatic and unpredictable. <div class="cro-passage-nav"> <div class="nav-next">[[Next: Core Mechanics->CRO Core Mechanics]]</div> <div class="nav-hub">[[Quick Reference Hub]]</div> </div>

<h1 class="cro-h1">Core Mechanics</h1> This section outlines the fundamental principles of the {tooltip: 'CRO', text: '{Char}, [Roll] for [Outcome]'} system, covering how to declare actions, make rolls, and interpret their basic outcomes. <h2 class="cro-h2">2.0 Basic Action & Rolling</h2> <h3 class="cro-h3">2.1 Declaring Intent & Skill Category</h3> 1. Player declares action and skill category. 2. Player rolls d10, applies mental modifiers, presents final <span class="cro-inline-code">[Result]/10</span>. 3. AI/DM uses the corresponding Threshold Scale below for outcome interpretation. <h3 class="cro-h3">Format</h3> [code] *I attempt to [action].* **[SKILL] ROLL** *[Result]/10 - [Outcome Description]* [continue] <h3 class="cro-h3">Example</h3> [code] *I attempt to bake a chocolate soufflé.* **CULINARY ROLL** *8/10 - Clear Success (achieves goal)* [continue] <div class="cro-hr"></div> <h2 class="cro-h2">3.0 Interpreting the Roll: Outcome Scales</h2> The d10 roll determines the raw result, but the Outcome Scales give that number narrative meaning. {tooltip:'CRO', text:'{Char}, [Roll] for [Outcome]'} offers two approaches: * **3.1 Main Outcome Scale:** This is the default, universal scale. Its broad outcome descriptions (Failure, Mixed Result, Success, etc.) provide a lot of flexibility to the AI; in how it interprets the action’s outcome given the context and the declared Skill. * Use this scale when you want to prioritize surprising or unexpected narrative developments arising directly from the roll number. * Works best with very powerful models (ex. Deepseek V3/R1). * **3.2 Skill-Specific Threshold Scales:** These optional scales offer more thematically tailored outcome descriptions for certain common skill types (Physical, Social, Perception, etc.). * Use these when you want the outcome interpretation to more strongly reflect the specific nature of the skill being employed, leading to potentially more predictable, but coherent results aligned with that skill domain. * Works best with smaller, less advanced, and less knowledgeable models. **Choosing Your Scale:** You, the player, can decide which scale best suits your intention for any given roll. You might default to the Main Scale for most actions, but then choose a Specific Scale when attempting a particularly characteristic task (e.g., using the Stealth Scale for a sneaking attempt). Try it out! See what method fits your roleplaying style, or simply works best. Whatever the case, do keep in mind that the AI needs to be kept aware that a low roll means a failure, a high one a success, and a middling one something in between - without the use of Outcomes, it will inevitably forget this, and drift towards giving you a positive outcome regardless of the dice number. <h3 class="cro-h3">3.1 Main Outcome Scale (General Use)</h3> <div class="cro-guide-image-container"> {embed image: 'Images/Main Scale.png', alt: 'Roxy ponders a bunch of funny looking shapes. They signify the entries. A metaphor, you see. You get it.'} </div> [code] *1/10 - Critical Failure (significant setback)* [continue] [code] *2/10 - Major Failure (clear consequence)* [continue] [code] *3/10 - Simple Failure (no progress)* [continue] [code] *4/10 - Poor Attempt (close, but falls short)* [continue] [code] *5/10 - Complication (a problem emerges)* [continue] [code] *6/10 - Minor Success (limited progress)* [continue] [code] *7/10 - Weak Success (achieves goal, barely)* [continue] [code] *8/10 - Clear Success (achieves goal)* [continue] [code] *9/10 - Strong Success (achieves goal cleanly)* [continue] [code] *10/10 - Critical Success (exceeds expectations)* [continue] This is CRO's foundational scale – the default when you're not using a skill-specific approach. Its 10 outcomes use deliberately broad language: that gives the AI maximum interpretive freedom while still providing clear guidance about the action's general trajectory. It’s universal: whether you're attempting to convince a guard, climb a wall, or solve a riddle, these outcome descriptions adapt naturally to the context. I designed this as the primary scale because it captures the essential dramatic beats without constraining too much how those beats manifest. Once again, these are relatively open-ended descriptions, and will work best with more powerful AI models: ones that can take a broad narrative direction and develop it meaningfully, if not creatively. <div class="cro-hr"></div> <h2 class="cro-h2">4.0 Modifying Rolls & Outcomes (Core)</h2> <h3 class="cro-h3">4.1 Difficulty Modifiers</h3> <div class="cro-guide-image-container"> {embed image: 'Images/Modifying Rolls.png', alt: 'A hand adjusting some parameters.'} </div> Apply mental modifiers to the final roll result. Keep in mind that you do not need to show them to the AI, just adjust the final number when you present it: [code] - Trivial: +3 - Easy: +1 - Moderate: +0 - Difficult: -1 - Formidable: -3 [continue] For example, if you rolled a 6 on a Difficult task (-1), present it as <span class="cro-inline-code">5/10</span>. <h3 class="cro-h3">4.2 Narrative Consequences</h3> * A recent success in a similar area could grant you a **purely** mental +2 or Advantage (roll twice, keep higher). * A recent failure could impose a **purely** mental -2 or Disadvantage (roll twice, keep lower). You simply add or subtract from your final presentation. Don’t show the math to the AI nor, in the case of advantage/disadvantage, inform it that you’ve used advantage/disadvantage - just the final number you end up. <div class="cro-passage-nav"> <div><span class="nav-prev">[[Previous: Getting Started->Getting Started]]</span></div> <div><span class="nav-next">[[Next: Skill-Specific Scales->CRO Skill-Specific Scales]]</span></div> <div class="nav-hub">[[Quick Reference Hub]]</div> </div>

<h1 class="cro-h1">Skill-Specific Scales</h1> While the [[Main Outcome Scale->CRO Core Mechanics]] is designed for universal application, these optional skill-specific scales offer more thematically tailored outcome descriptions for common types of actions. They can provide more predictable, but contextually richer, results. Use these when you want the outcome interpretation to more strongly reflect the specific nature of the skill being employed. <h2 class="cro-h2">3.2 Skill-Specific Scales</h2> <h3 class="cro-h3">Physical Scale</h3> (Strength, Athletics, Endurance, Combat Maneuvers) <div class="cro-guide-image-container"> {embed image: 'Images/Physical Scale.png', alt: 'A buff arm, flexing.'} </div> [code] *1/10 - Harsh Consequence (clear physical cost or injury)* [continue] [code] *2/10 - Strained Effort (body struggles, fatigue sets in)* [continue] [code] *3/10 - Forced Failure (pushing through but can't complete)* [continue] [code] *4/10 - Rough Attempt (partial progress, noticeable strain)* [continue] [code] *5/10 - Taxing Success (achieves goal but at physical cost)* [continue] [code] *6/10 - Adequate Performance (body cooperates, gets it done)* [continue] [code] *7/10 - Steady Control (confident execution)* [continue] [code] *8/10 - Strong Performance (body responds well)* [continue] [code] *9/10 - Smooth Execution (effortless feel)* [continue] [code] *10/10 - Perfect Form (exactly as intended, if not better)* [continue] Physical actions present a unique challenge: they carry real consequences that can impact your entire ongoing narrative. A failed climb isn't just "you don't get up the wall" – it might mean injury, exhaustion, or equipment damage that matters for the next hour of play. That's why I settled on 10 outcomes here, giving you enough granularity to distinguish between "winded but fine" and "twisted ankle that'll slow you down." The scale focuses on the relationship between your character's intention and their body's response rather than just success/failure, because in my testing, that relationship created far more interesting ongoing complications than simple binary outcomes. <div class="cro-hr"></div> <h3 class="cro-h3">Manual Scale</h3> (Dexterity, Crafting, Fine Motor Skills, Delicate Tasks) <div class="cro-guide-image-container"> {embed image: 'Images/Manual Scale.png', alt: 'Two hands working on a engine.'} </div> [code] *1/10 - Complete Mishap (task goes wrong, clear setback)* [continue] [code] *2-3/10 - Sloppy Work (barely functional, obvious flaws)* [continue] [code] *4-5/10 - Rough Execution (works but crude, noticeable imperfections)* [continue] [code] *6-7/10 - Competent Work (solid execution, minor rough edges)* [continue] [code] *8-9/10 - Clean Execution (smooth work, well-handled)* [continue] [code] *10/10 - Precise Control (exactly as intended, no wasted motion)* [continue] Originally, I had manual skills lumped in with physical ones, but I quickly realized they needed their own approach. When you're picking a lock or crafting something, the question isn't really "did it work" but "how well did you execute it?" A crude lockpick job might open the door but leave obvious scratches, while masterful work leaves no trace. Seven outcomes gives you enough range to capture that spectrum without bogging down the narrative in excessive detail. The focus on execution quality over mere functionality emerged from countless sessions where the difference between "it works" and "it works elegantly" drove the next scene's direction. <div class="cro-hr"></div> <h3 class="cro-h3">Mental Scale</h3> (Problem Solving, Intelligence, Academic Knowledge, Remembering Facts or Learned Information) <div class="cro-guide-image-container"> {embed image: 'Images/Mental Scale.png', alt: 'Roxy, thinking through a problem.'} </div> [code] *1/10 - Analysis Paralysis (ex. overthinking, endless loops, decision paralysis, etc.)* [continue] [code] *2/10 - Tunnel Vision (ex. fixation, missing alternatives, rigid thinking, etc.)* [continue] [code] *3/10 - Emotional Bias (ex. wishful thinking, fear-driven reasoning, prejudice, etc.)* [continue] [code] *4/10 - Scattered Focus (ex. jumping between ideas, unfocused effort, etc.)* [continue] [code] *5/10 - Logical Analysis (ex. step-by-step reasoning, systematic breakdown, etc.)* [continue] [code] *6/10 - Pattern Recognition (ex. connections, similarities, recurring themes, etc.)* [continue] [code] *7/10 - Practical Application (ex. real-world solutions, functional thinking, etc.)* [continue] [code] *8/10 - Creative Synthesis (ex. novel combinations, innovative approaches, etc.)* [continue] [code] *9/10 - Abstract Reasoning (ex. theoretical concepts, pure logic, philosophy, etc.)* [continue] [code] *10/10 - Intuitive Leap (ex. sudden understanding, gut insights, breakthrough moments, etc.)* [continue] Things did get tricky with this one. Mental actions in AI narratives have a fundamental problem: there's usually no objective information waiting to be discovered. When you roll to "figure out the puzzle," the AI hasn't actually designed a puzzle with a specific solution - it never thinks that much ahead, it simply can’t. I kept running into situations where the AI would either make up jarring new facts, or the roll would become either meaningless, or the whole category a catch-all, capable of being used for everything. It still kind of is the latter, but eventually I stopped asking "what do you learn" and started asking "how does your mind approach this?" The 10 outcomes ahead represent different cognitive modes for your character – from analysis paralysis to intuitive leaps. This actually gives the AI something concrete to work with without forcing it to necessarily invent setting details. <div class="cro-hr"></div> <h3 class="cro-h3">Social Scale</h3> (Persuasion, Deception, Intimidation, Charm, Negotiation) <div class="cro-guide-image-container"> {embed image: 'Images/Social Scale.png', alt: 'A very intense chat.'} </div> [code] *1-2/10 - Momentum Lost (interaction creates resistance or withdrawal)* [continue] [code] *3-4/10 - Awkward Exchange (stilted, uncomfortable social flow)* [continue] [code] *5-6/10 - Neutral Transaction (stalling, neither builds nor damages social standing)* [continue] [code] *7-8/10 - Strong Rapport (builds connection and cooperation)* [continue] [code] *9-10/10 - Lasting Impact (memorable interaction, doors open)* [continue] Social interactions... I tend not to roll for these, actually, I prefer to talk in character, and try to convince/intimidate the chatbot that way. So I actually haven't tested this scale too hard! I'm open to improvements, as I am with the rest of the system's components. Five outcomes captures the essential range from "actively damages the relationship" to "creates lasting positive impression" without overthinking it. During development, I noticed that focusing on social momentum and interpersonal energy worked better than specific persuasion mechanics. For example, rolling high doesn't necessarily mean "they agree with you," but rather "you've created positive social energy, and that may open doors." This keeps the AI focused on relationship dynamics rather than forcing specific character decisions. <div class="cro-hr"></div> <h3 class="cro-h3">Perception Scale</h3> (Awareness, Investigation, Searching, Noticing, Sensing Motives) <div class="cro-guide-image-container"> {embed image: 'Images/Perception Scale.png', alt: 'A magnifying glass inspecting the world around itself.'} </div> [code] *1-2/10 - Physical State (ex. appearance, condition, immediate qualities, etc.)* [continue] [code] *3-4/10 - Functional Aspect (ex. role, purpose, capabilities, etc.)* [continue] [code] *5-6/10 - Contextual Clues (ex. connections, significance, anomalies, etc.)* [continue] [code] *7-8/10 - Hidden Elements (ex. concealed aspects, overlooked details, etc.)* [continue] [code] *9-10/10 - Temporal Traces (ex. history, recent changes, accumulated effects, etc.)* [continue] This 5-outcome scale works differently from the others – instead of measuring how well you perceive, it determines what type of perception you apply. Each outcome represents a different investigative lens: you might focus on physical condition, functional purpose, contextual relationships, hidden features, or traces of recent activity. This approach emerged from a practical problem with traditional perception mechanics: telling the AI that you "notice more things" often produced vague atmospheric descriptions that gave players nothing concrete to pursue. Oftentimes, the instruction was completely ignored. By specifying the type of observation, the scale consistently delivers actionable information. Rolling for "functional aspect" when examining a door gives you insights about its purpose and mechanisms, while "temporal traces" might reveal recent use patterns – both useful for planning your next action, I feel. <div class="cro-hr"></div> <h3 class="cro-h3">Fortune Scale</h3> (Luck, Chance, Timing, Random Circumstance) <div class="cro-guide-image-container"> {embed image: 'Images/Luck Scale.png', alt: 'A sequence of clover leaves growing. At the end, a four-leafed one.'} </div> [code] *1/10 - Bad Timing (unfavorable coincidence)* [continue] [code] *2-5/10 - Unlucky Turn (circumstances hinder, disadvantage)* [continue] [code] *6-9/10 - Lucky Break (favorable circumstance, advantage)* [continue] [code] *10/10 - Perfect Timing (major positive turn of events)* [continue] Fortune is pure circumstance, isn't it? Things either go your way or they don't. I originally had a complex 10-outcome scale with dramatic language like "calamitous misfortune," but found it was pulling the AI into over-the-top territory. Bad timing is just bad timing; lucky breaks are just lucky breaks. Sometimes the most elegant design choice is restraint, and 4, well-defined outcomes work well enough. <div class="cro-hr"></div> <h3 class="cro-h3">Stealth Scale</h3> (Sneaking, Hiding, Moving Silently - Replaces default Physical/Manual for these actions) <div class="cro-guide-image-container"> {embed image: 'Images/Stealth Scale.png', alt: 'A lock being picked.'} </div> [code] *1/10 - Fully Exposed (immediate detection, alarm raised)* [continue] [code] *2/10 - Clearly Spotted (presence and intent obvious)* [continue] [code] *3/10 - Obviously Noticed (seen but intent unclear)* [continue] [code] *4/10 - Suspicion Raised (traces left, heightened awareness)* [continue] [code] *5/10 - Minor Disturbance (small signs noticed)* [continue] [code] *6/10 - Barely Avoided (close call, narrowly unnoticed)* [continue] [code] *7/10 - Successfully Hidden (avoided direct detection)* [continue] [code] *8/10 - Cleanly Unnoticed (no awareness triggered)* [continue] [code] *9/10 - Seamless Passage (no signs of presence)* [continue] [code] *10/10 - Perfect Concealment (completely undetected, no trace)* [continue] Stealth demands granularity because detection operates on a gradient, and small differences have big consequences. There's a meaningful distinction between "guards are suspicious" and "guards are actively searching," and your options change dramatically between those states. These 10 outcomes let the AI calibrate responses precisely, from "completely blown" to "ghosted through undetected." I kept this one linear because stealth is fundamentally about degrees of concealment – it's one of the few skills where more complex categorization would actually reduce clarity. <div class="cro-passage-nav"> <div><span class="nav-prev">[[Previous: Core Mechanics->CRO Core Mechanics]]</span></div> <div><span class="nav-next">[[Next: Time & Extensions Overview->CRO Time & Extensions Overview]]</span></div> <div class="nav-hub">[[Quick Reference Hub]]</div> </div>

<h1 class="cro-h1">Time Progression & Extensions Overview</h1> This section covers how to manage the passage of time in your narrative and introduces the concept of Optional Extensions to further customize the CRO system. <h2 class="cro-h2">5.0 Time Progression</h2> <div class="cro-guide-image-container"> {embed image: 'Images/Time Roll.png', alt: 'A hand holding up a clock. Yes, it does look like the doomsday clock. This is very deep.'} </div> Narratives often involve periods where little of immediate consequence happens, or where significant change occurs gradually over time. Instead of playing out every single moment in what is occasionally excruciating detail, you should be allowed to "skip forward," and to move the story to the next point of interest. This maintains pacing, allows for character development or environmental shifts during the intervening time, and focuses gameplay on significant events. Use the following methods to manage the passage of time. <h3 class="cro-h3">5.1 Method A: Simple Declaration</h3> Allows the player to specify an intended duration for a time skip. In this specific case, the player doesn’t necessarily have to roll anything. (↪ [[More on Simple Declaration->CRO Time Simple Declaration Detail]]) [code] **[approximate duration, e.g., 'DAYS', 'WEEKS', 'MONTHS'] LATER.** [continue] <h3 class="cro-h3">5.2 Method B: Random Duration</h3> This method is useful when you want to jump forward an indeterminate amount of time, letting chance dictate how long passes before the narrative picks up again. Simply roll a d10 and consult the scale below to see the duration of the time skip. (↪ [[More on Random Duration->CRO Time Random Duration Detail]]) <h4 class="cro-h4">Format</h4> [code] **TIME ROLL** *[Result]/10 - [Time Jump]* **[UNIT OF TIME] LATER** [continue] <h4 class="cro-h4">Random Duration Scale</h4> [code] *1-3: Days later* *4-6: Weeks later* *7-8: Months later* *9-10: Years later* [continue] <h3 class="cro-h3">5.3 Method C: Time & Fortune</h3> Allows the player to describe the character of a chosen, or random, span of time, providing narrative flavour without prejudging outcomes or significant events. Offers an alternative to the basic {tooltip: 'TIME ROLL', text: 'See Method B: Random Duration'} for a more directed pacing (a similar deal to the Skill-Specific scales), allowing the player to choose to characterize the period as broadly favourable or unfavourable. (↪ [[More on Time & Fortune->CRO Time TimeAndFortune Detail]]) <h4 class="cro-h4">Format</h4> [code] **[UNIT OF TIME] LATER** **TIME ROLL** *[Result]/10 - [Time & Fortune Scale Outcome]* [continue] <h4 class="cro-h4">Time & Fortune Scale</h4> [code] *1/10 - Deep Troubles (Time passed under notably difficult background conditions)* *2/10 - Mostly Downs (The period felt marked by a generally negative trend)* *3/10 - Problems Mount (Minor issues seemed to consistently arise or worsen)* *4/10 - Felt Resisted (The period generally felt resistant or effortful)* *5/10 - Slight Drag (Progress or stability felt subtly held back during this time)* *6/10 - Slight Boost (Progress or stability felt subtly helped along during this time)* *7/10 - Easy Flow (The period generally felt smooth or cooperative)* *8/10 - Things Align (Minor opportunities seemed to consistently arise or improve)* *9/10 - Mostly Ups (The period felt marked by a generally positive trend)* *10/10 - Great Fortune (Time passed under notably favourable background conditions)* [continue] Characterizing entire time periods required a full 10-outcome range to capture the subtlety between "slightly difficult" and "persistently troubled" periods. Unlike moment-to-moment fortune rolls, this scale affects background conditions over extended periods, so the extra granularity helps the AI calibrate how that time period should feel, or how it would’ve felt, without mandating specific events. The scale shapes atmosphere, and so the plot, but indirectly. <div class="cro-hr"></div> <h2 class="cro-h2">6.0 Introduction to Extensions</h2> <div class="cro-guide-image-container"> {embed image: 'Images/Extensions Introduction.png', alt: 'A connective tree made of asterisks and text boxes sprouting from your hand.'} </div> The Core Mechanics provide the foundation for the system. Extensions are optional modules you can add to introduce more specific rules or narrative effects for particular situations. Think of them as tools in a toolbox – use the ones that enhance your desired style of play and ignore the ones that don't. You can decide which extensions are active at the start of a session, or even introduce one mid-game if a situation calls for it. They are designed to be self-contained and integrate smoothly with the core d10 roll system. The following section, "Extensions Collection," groups these optional rules by their general function. <div class="cro-passage-nav"> <div><span class="nav-prev">[[Previous: Skill-Specific Scales->CRO Skill-Specific Scales]]</span></div> <div><span class="nav-next">[[Next: Extensions Collection->CRO Extensions Collection]]</span></div> <div class="nav-hub">[[Quick Reference Hub]]</div> </div>

<h1 class="cro-h1">Extensions Collection</h1> This passage gathers all optional extensions, organized by their primary function. Use these to add more depth and specific mechanics to your CRO-driven narratives. <h2 class="cro-h2">7.0 Action Modifiers & Enhancements</h2> <div class="cro-guide-image-container"> {embed image: 'Images/Modifiers And Enhancements.png', alt: 'Three hardworking hands on a graph doing their best to rationalize the roll number in their favour lmao'} </div> This first set of extensions focus on modifying how individual actions work. <h3 class="cro-h3">7.1 Stakes</h3> Raise or lower the stakes before a roll (mentally) by 1-3: <span class="cro-inline-code">- Raising Stakes (+1 to +3)</span> * Higher payoff if successful, worse outcome if you fail <span class="cro-inline-code">- Lowering Stakes (-1 to -3)</span> * Softer fail consequences, but capped success If roll ≥ 5, add the chosen “stakes” number to your final result (up to 10). If roll < 5, subtract the “stakes” number (down to 1). You do this mentally and only show the final adjusted roll. Sometimes the standard difficulty modifiers don't capture what you're really after. Stakes let you amplify both the potential reward and the potential cost of an action without changing the basic difficulty. This emerged from situations where I wanted to attempt something that wasn't necessarily harder, but where the consequences mattered more. Climbing a wall during a casual exploration versus climbing that same wall while being chased by guards - same physical challenge, completely different stakes. The mechanic lets you push for bigger successes while accepting that failures will hurt more, creating that "high risk, high reward" feeling. It makes certain moments feel memorable, though, to be fair, I don’t use it often. This is a common pattern among these extensions - that's the point. You use the ones that feel right to you, or the situation at hand. <div class="cro-hr"></div> <h3 class="cro-h3">7.2 Pushing Limits</h3> Allows a character to attempt an action far beyond their normal capabilities or under extreme duress by accepting a significantly higher risk of failure and more severe consequences. <h4 class="cro-h4">Format</h4> [code] I attempt to push my limit by [describe extreme action]. **[SKILL] ROLL** *[Result]/10 - [Pushing Limits Outcome]* [continue] <h4 class="cro-h4">Pushing Limits Scale</h4> The player must use this harsher scale when "Pushing Limits" is invoked, reflecting the mentally increased difficulty threshold. It replaces the Main Outcome Scale for the given roll’s outcome. [code] *1/10 - Severe Backfire (major negative consequence)* [continue] [code] *2-3/10 - Costly Failure (significant setback, clear price paid)* [continue] [code] *4-5/10 - Painful Attempt (falls short, noticeable strain or cost)* [continue] [code] *6-7/10 - Strained Success (goal met, but with visible effort)* [continue] [code] *8-9/10 - Hard-Won Achievement (success despite the odds)* [continue] [code] *10/10 - Breakthrough Performance (exceeds expectations)* [continue] When your character attempts something far beyond normal capability, consequences should be harsher than usual, but I learned that overly dramatic language ("catastrophic failure") pushed the AI toward melodramatic responses. Also, who wants to use a scale that just destroys your character? Sure, it's narratively interesting... for a quick and brutal ending to your session, that is. This 7-outcome scale compresses the success range deliberately – most attempts to push limits should be difficult, with only the highest rolls achieving clean success. This reflects the inherent danger of exceeding normal boundaries while keeping it grounded. <div class="cro-hr"></div> <h3 class="cro-h3">7.3 Skill Leverage</h3> A slight adjustment that allows players to declare using a secondary, relevant skill to influence their primary action by letting the chatbot be more aware of what skills the player is leveraging. Simply declare your primary action (Skill A) and the leveraged skill (Skill B). [code] *I attempt to [Primary Action] using [Skill A], leveraging my [Skill B] to [Describe how B helps A].* **[SKILL A]&[SKILL B] ROLL** *[Result]/10 - [Main Outcome Scale]* [continue] This simple addition addresses a common frustration: when your character's approach clearly involves multiple types of expertise, but the system only recognizes one. Using your knowledge of architecture to inform your climbing approach, or leveraging your social skills to make a manual task easier by getting help - these combinations happen naturally in stories but can feel overlooked by single-skill rolls. Skill Leverage doesn't change the mechanics, just makes the AI more aware of your character's multi-faceted approach. It's particularly useful with less sophisticated AI models that might miss the connection between different skill domains. <div class="cro-hr"></div> <h3 class="cro-h3">7.4 Serendipity</h3> When a roll results in a mixed outcome (4-6 on the Main Outcome Scale), an opportunity emerges: you can add another action to the original roll. The outcome of this secondary action is decided by flipping a coin, allowing the user to get to shift towards a partial success, or a failure. [code] *I attempt to [describe original action].* **[PRIMARY SKILL] ROLL** *[Result]\/10 - [Outcome]* **SERENDIPITY ACTION** *Coin Flip - Heads (Success)* or *Coin Flip - Tails (Failure)* [continue] Mixed results create natural story tension, but sometimes they leave you hanging in an unsatisfying middle ground. Serendipity gives you a way to push through that ambiguity when it feels appropriate. The coin flip keeps it simple - no complex modifiers or new scales to remember. Just you and Mr.Lincoln, partners in crime. <div class="cro-hr"></div> <h3 class="cro-h3">7.5 Targeted Effort</h3> Allows a player to declare a more specific, difficult goal within a standard action attempt, aiming for precision rather than general success. Use this when achieving a precise outcome is more important than simply succeeding at the general task (e.g., hitting a specific weak point, finding a particular clue, using exact leverage). This offers the potential for more impactful success but increases the risk of failure due to the inherent difficulty of precision. 1. Declare your action using the standard format but include a specific target or refined goal within your description. 2. Mentally apply a -2 modifier to your d10 roll result before presenting it. This reflects the increased difficulty of the specific aim. 3. Present the final, adjusted <span class="cro-inline-code">[Result]/10</span> using the standard <span class="cro-inline-code">**[SKILL] ROLL**</span> format and outcome scales. <div class="cro-hr"></div> <h3 class="cro-h3">7.6 Guarded Approach</h3> Allows a player to perform an action with deliberate caution, aiming to minimize the risk of critical failure at the cost of potential critical success. Use this when avoiding the absolute worst outcome (a roll of 1) is more important than achieving the absolute best (a roll of 10). It's a risk-averse strategy suitable for high-danger situations where a standard failure is acceptable, but a disaster is not. 1. Declare your action using the standard format and specify you are taking a <span class="cro-inline-code">**Guarded Approach**</span>. 2. Roll the d10 for the relevant skill. 3. Mentally adjust the roll **only** if it is a 1 or a 10: * If the roll is 1, treat it as 2. * If the roll is 10, treat it as 9. 4. Present the final (potentially adjusted) <span class="cro-inline-code">[Result]/10</span> using the standard <span class="cro-inline-code">**[SKILL] ROLL**</span> format. <div class="cro-hr"></div> <h2 class="cro-h2">8.0 Outcome Nuance & Detail (Investigation & Discovery)</h2> <div class="cro-guide-image-container"> {embed image: 'Images/Outcome Nuance.png', alt: 'The X-ray of a backpack, revealing a sword inside.'} </div> These extensions add detail to specific types of actions. <h3 class="cro-h3">8.1 Item Focus</h3> For examining or interacting with a specific physical object to understand its condition, properties, or potential uses. <h4 class="cro-h4">Format</h4> [code] *I examine the [Item Name] closely.* **ITEM FOCUS ROLL** *[Result]/10 - [Item Insight/Condition]* [continue] <h4 class="cro-h4">Item Insight Scale</h4> [code] *1/10 - Material Properties (ex. composition, durability, craftsmanship quality, etc.)* [continue] [code] *2-3/10 - Functional Design (ex. intended purpose, how it works, efficiency, etc.)* [continue] [code] *4-5/10 - Historical Context (ex. age, previous use, wear patterns, etc.)* [continue] [code] *6-7/10 - Hidden Features (ex. concealed mechanisms, subtle details, etc.)* [continue] [code] *8-9/10 - Contextual Significance (ex. relevance to situation, connections, etc.)* [continue] [code] *10/10 - Symbolic Meaning (ex. cultural importance, personal resonance, etc.)* [continue] As an extension of the perception system, this 7-outcome scale applies the same domain-based approach to examining specific objects. Rather than forcing the AI to determine how "much" an item is, it focuses your attention on different aspects: material properties, functional design, historical context, and so on. Seven outcomes provide good coverage of investigation approaches without overwhelming a mechanic that's meant to be used sparingly in the first place. <div class="cro-hr"></div> <h3 class="cro-h3">8.2 Inventory Check</h3> Determines what mundane, useful, or unexpected items a character finds when checking their pockets, bag, or immediate personal effects without searching for something specific. This is a way to force the chatbot AI to give you some items to play with. The result indicates the quantity and general nature/relevance of the items found. <h4 class="cro-h4">Format Examples</h4> [code] *I pat down my pockets to see what I have on me.* *I rummage through my bag to see what I have on me.* *I check to see what I have on me.* **INVENTORY CHECK** *[Result]/10 - [Find Description]* [continue] <h4 class="cro-h4">Find Scale</h4> [code] *1/10 - Empty Search (nothing useful found)* [continue] [code] *2-3/10 - Mundane Items (basic, expected contents)* [continue] [code] *4-5/10 - Useful Basics (practical items for current situation)* [continue] [code] *6-7/10 - Helpful Resources (relevant, contextually appropriate)* [continue] [code] *8-9/10 - Valuable Discovery (significantly useful items)* [continue] [code] *10/10 - Perfect Find (exactly what's needed right now)* [continue] <h4 class="cro-h4">Usage Notes for Inventory Check</h4> The find scale had a habit of derailing narratives. High rolls would produce elaborate, setting-defining items that the AI felt compelled to make significant, while low rolls left you empty-handed in situations where some basic supplies made narrative sense. This 7-outcome scale emphasizes contextual appropriateness over raw value – a perfect roll gives you "exactly what's needed right now" rather than "something amazingly valuable." This keeps the extension useful without letting it hijack your story's direction. Again, use sparingly. We don’t want the AI getting overwhelmed, nor your character ending up with an endless amount of items. Maybe every TIME roll? Your call. <div class="cro-hr"></div> <h2 class="cro-h2">9.0 Narrative & Environmental Factors</h2> <div class="cro-guide-image-container"> {embed image: 'Images/Environmental Extensions.png', alt: 'Oh no! Rocks are coming your way! Move out of the way!'} </div> Some extensions can introduce external elements beyond your character's direct control. <h3 class="cro-h3">9.1 Random Events</h3> Random events needed their own logic. The linear "how important is this" approach often produced either trivial distractions or story-hijacking developments. This scale focuses on external event types – weather shifts, people movement, objects – that add texture without forcing major plot changes. Each category suggests environmental dynamism (big words for big people) rather than dramatic revelation, keeping random events in their proper place: a supporting role. <h4 class="cro-h4">Format</h4> [code] **RANDOM EVENT** *[Result]/10 - [Impact Description]* [continue] You can also roll a d10 (1-7 = No event, 8-10 = Event occurs) to see if a random event happens at all. <h4 class="cro-h4">Random Event Scale</h4> [code] *1/10 - Weather Shift (ex. sudden climate change, atmospheric conditions, etc.)* [continue] [code] *2/10 - Social Arrival (ex. people appear, someone approaches, crowd gathers, etc.)* [continue] [code] *3/10 - Object Appearance (ex. item falls, tool breaks, something delivered, etc.)* [continue] [code] *4/10 - Communication Event (ex. message received, overhear conversation, etc.)* [continue] [code] *5/10 - Environmental Change (ex. door opens, path blocked, noise starts, etc.)* [continue] [code] *6/10 - Timing Coincidence (ex. simultaneous events, scheduling overlaps, etc.)* [continue] [code] *7/10 - Authority Attention (ex. official notice, rule enforcement, inspection, etc.)* [continue] [code] *8/10 - Resource Fluctuation (ex. supply issue, price change, availability shift, etc.)* [continue] [code] *9/10 - Technical Disruption (ex. equipment malfunction, system failure, etc.)* [continue] [code] *10/10 - Information Surface (ex. documents appear, secrets revealed accidentally, etc.)* [continue] <div class="cro-hr"></div> <h3 class="cro-h3">9.2 Emotional Static</h3> Introduce brief, surreal, and non-literal manifestations of a character's strong underlying emotion (like intense fear, jealousy, joy, despair) into the surrounding environment or perception, creating bizarre or unsettling narrative moments independent of direct action. <h4 class="cro-h4">Format</h4> [code] *(Player describes character experiencing a strong underlying emotion)* *I focus on this feeling of [briefly name emotion, e.g., intense regret, blinding hope].* **EMOTIONAL STATIC ROLL** *[Result]/10 - [Static Manifestation Description] (triggered by [Emotion Name])* [continue] <h4 class="cro-h4">Static Manifestation Scale</h4> [code] *1-2/10 - Sensory Whisper (ex. fleeting sound, incongruous smell, etc.)* [continue] [code] *3-4/10 - Perceptual Drift (ex. visual ripple, misplaced shadow, etc.)* [continue] [code] *5-6/10 - Strange Occurrence (ex. impossible object appears, unnatural sensation, etc.)* [continue] [code] *7-8/10 - Reality Distortion (ex. environment briefly warps, spectral image, etc.)* [continue] [code] *9-10/10 - Psychic Feedback (ex. physics briefly altered, reality glitch, etc.)* [continue] Emotional static... this one is weird. Sure enough, it was suggested to me by Claude - I wanted to put at least one example of a completely AI-generated extension, because why not. There is some potential. It represents surreal moments where strong feelings briefly warp perception or reality. Five outcomes provide enough range from subtle sensory whispers to reality glitches, encouraging the AI to go completely off the rails, or get very meta. Maybe you want that for your story! The challenge was calibrating weirdness – too subtle and it becomes forgettable, too intense and it overwhelms the main narrative thread too quickly. <div class="cro-hr"></div> <h3 class="cro-h3">9.3 Interference/Nemesis</h3> (Formerly 9.4 in guide) Introduce a competing force or character that actively works against the player's goal. To be placed right after an action. <h4 class="cro-h4">Format</h4> [code] **INTERFERENCE ROLL** *[Result]/10 - [Opposition Level]* [continue] <h4 class="cro-h4">Opposition Scale</h4> [code] *1/10 - Minimal Hindrance (barely noticeable resistance)* *2-3/10 - Minor Obstacles (slight complications, easy to work around)* *4-5/10 - Notable Resistance (requires attention, strategic adjustment)* *6-7/10 - Active Opposition (substantial challenge, complicates approach)* *8-9/10 - Strong Interference (significantly threatens success)* *10/10 - Direct Confrontation (goal requires completely new approach)* [continue] Opposition should escalate your challenge level without necessarily changing your fundamental approach until the highest levels. "Minor obstacles" require attention but don't force you to completely rethink your strategy, while "direct confrontation" demands a new plan entirely. The scale gives the AI clear guidance on how much resistance to apply to your actions. <div class="cro-hr"></div> <h2 class="cro-h2">10.0 Meta-Narrative & Self-Awareness</h2> <div class="cro-guide-image-container"> {embed image: 'Images/Meta Narrative.png', alt: 'Roxy reaches artificial enlightment.'} </div> Finally, these extensions operate at the story level rather than the action level. <h3 class="cro-h3">10.1 Montage</h3> Progress through multiple actions over a longer period. [code] **MONTAGE ROLL** *[Time Period]/10 - [Total length of time that passes] - [Number of actions] Actions* [continue] Then break it down into segments, each with its own roll. Example: [code] **MONTAGE ROLL** *6/10 - 6 MONTHS - 4 Actions* **FIRST MONTH** *1/4 - [Action]* **[SKILL] ROLL** *[Result]/10 - [Outcome]* *2/4 - [Action]* **[SKILL] ROLL** *[Result]/10 - [Outcome]* **MONTHS 2-4** *3/4 - [Action]* **[SKILL] ROLL** *[Result]/10 - [Outcome]* *4/4 - [Action]* **[SKILL] ROLL** *[Result]/10 - [Outcome]* [continue] Traditional scene-by-scene play sometimes hits periods where you need to cover significant time and multiple attempts without losing narrative momentum. Montage provides structure for those "training sequences" or "investigation periods" where the journey matters but doesn't need minute-by-minute detail. The initial roll determines scope and pace, then the segmented approach lets you highlight key moments without getting bogged down. This works particularly well for character development periods, skill learning, or any situation where the pattern of attempts and outcomes tells a more interesting story than any single action would. <div class="cro-hr"></div> <h3 class="cro-h3">10.2 Narrative Techniques</h3> This extension provides direct communication methods for shaping your story without relying on dice. These are tools for active, more direct collaboration - sometimes you recognize what a scene is missing, and shouldn’t be forced to hope for a certain roll’s outcome to fix it. Use them to guide the AI toward specific narrative goals through clear signaling. Communication is key, as they say. <h4 class="cro-h4">Atmospheric Direction</h4> [code] * (I'd like some environmental details here) * (Something feels off about this place) * (The mood should feel [tense/peaceful/mysterious]) [continue] <h4 class="cro-h4">Scene Management</h4> [code] * "I take a moment to look around the room..." (shifts focus from dialogue to environment) * "After we finish talking..." (signals conversation should wrap up) * "Before I respond..." (creates space for reflection or action) [continue] <h4 class="cro-h4">Pacing Control</h4> [code] * "Quickly, I..." or "I take my time..." (guides scene speed) * "Without hesitation..." vs "I pause, considering..." (controls decision tempo) * (Let's move this along) or (Let's slow down and explore this) [continue] <h4 class="cro-h4">Thread Management</h4> [code] * "I remember [person/event] and..." (brings back dormant story elements) * "This reminds me of..." (connects to established narrative threads) * "Speaking of [topic]..." (steers conversation direction) [continue] <h4 class="cro-h4">Emotional Anchoring</h4> [code] * Starting actions with internal states: "Feeling nervous, I..." "Curious about her reaction, I..." * (My character should be more [confident/worried/excited] here) [continue] <h4 class="cro-h4">Environmental Prompting</h4> [code] * "The sound of [footsteps/machinery/voices] catches my attention..." * "I notice..." (directs AI to provide specific observational details) * "Something about [object/person] strikes me as..." (requests focused description) [continue] <div class="cro-hr"></div> <h3 class="cro-h3">10.3 Narrative Momentum</h3> When you want to inject unexpected energy into a scene without controlling the specifics, roll for narrative momentum. This creates genuine surprise by shifting the emotional or dramatic tone in ways you might not have anticipated. <h4 class="cro-h4">Format</h4> [code] *I want to shift the energy of this scene.* **NARRATIVE MOMENTUM ROLL** *[Result]/10 - [Energy Type]* [continue] <h4 class="cro-h4">Narrative Energy Scale</h4> [code] *1/10 - Mysterious Energy (ex. unexplained elements, hidden significance, secrets, etc.)* *2/10 - Urgent Pressure (ex. time constraints, approaching deadlines, rushing, etc.)* *3/10 - Melancholic Mood (ex. nostalgia, loss, bittersweet moments, etc.)* *4/10 - Tense Undercurrent (ex. brewing conflict, unspoken tension, etc.)* *5/10 - Contemplative Tone (ex. reflection, philosophical moments, quiet thought, etc.)* *6/10 - Playful Spirit (ex. humor, lightness, unexpected fun, etc.)* *7/10 - Intimate Atmosphere (ex. personal connection, vulnerability, closeness, etc.)* *8/10 - Dramatic Intensity (ex. heightened stakes, powerful emotions, etc.)* *9/10 - Peaceful Flow (ex. calm, harmony, gentle moments, etc.)* *10/10 - Chaotic Disruption (ex. unpredictability, wild energy, upheaval, etc.)* [continue] This roll determines what kind of energy enters the scene, letting the AI manifest it naturally within your current situation. Unlike specific suggestions, you're genuinely open to whatever type of momentum emerges. <div class="cro-passage-nav"> <div><span class="nav-prev">[[Previous: Time & Extensions Overview->CRO Time & Extensions Overview]]</span></div> <div><span class="nav-next">[[Next: Advanced Tools & Examples->CRO Advanced Tools & Examples]]</span></div> <div class="nav-hub">[[Quick Reference Hub]]</div> </div>

<h1 class="cro-h1">Advanced Tools & Examples</h1> This section covers the Judge System for impartial rulings, provides comprehensive implementation examples, and offers best practices for using CRO. <h2 class="cro-h2">Part 3: Judge System</h2> <div class="cro-guide-image-container"> {embed image: 'Images/AI Judge.png', alt: 'A gavel, concidentally also reaching artificial enlightment.'} </div> While self-managed rolls work beautifully most of the time, there's a human element that can sometimes get in the way: bias. Unconscious, maybe. We’re not talking about dramatic cheating, mind you! More so subtle things - maybe you think that climb should be rated "Easy" instead of "Moderate" because you really want to reach that tower, or a social roll becomes "Difficult" because you're not sure how to handle success in that moment. You are both a player wanting things to go well, and a referee trying to assess fairly. Most of the time, this bias actually improves the experience. You naturally adjust difficulty to maintain good pacing and avoid crushing setbacks when they'd derail interesting storylines. But sometimes you want the dice to be genuinely impartial - during crucial moments where the uncertainty itself creates tension, when you're trying to discover what your character is truly capable of, or when you want to be surprised by the chatbot’s storytelling. The Judge AI system splits these responsibilities: a separate AI instance handles all mechanical rulings while your main storytelling AI focuses purely on narrative response. It's entirely optional, and honestly, most players will prefer the flexibility of self-managed rolls. I sure do. But when you want that "anything could happen" feeling, or when you catch yourself nudging results toward desired outcomes, having an impartial referee ensures the system works exactly as designed. <h3 class="cro-h3">11.2 Set-up</h3> Create a separate conversation with an AI using: * The prompt below (see detail pages for full prompts) * The "CRO Reference Document", dowloadable below This AI will serve as your impartial referee for all CRO rolls. Simply provide it with your intended action and dice roll, and it will return a properly formatted result for you to use with your main storytelling AI. <h4 class="cro-h4">Module Selection Guide:</h4> * **New users**: Start with just the [[Base Prompt->CRO Judge AI Base Prompt Detail]] * **Experienced users**: Add the [[Scales Module->CRO Judge AI Scales Module Detail]] for more nuanced outcomes * **Full system users**: Add both the Scales Module and the [[Extensions Module->CRO Judge AI Extensions Module Detail]] to leverage the complete CRO toolkit Copy and paste the base prompt, then add any modules you want below it. <h4 class="cro-h4">Input Format:</h4> [code] "I attempt to [action]. [Any relevant context]. [Your d10 roll result]" [continue] <h4 class="cro-h4">Example:</h4> [code] "I attempt to convince the guard to let me pass. He looks suspicious and I'm clearly not supposed to be here. 7" [continue] For the full prompts, see: * [[Judge AI - Base System Prompt->CRO Judge AI Base Prompt Detail]] * [[Judge AI - Scales Module->CRO Judge AI Scales Module Detail]] * [[Judge AI - Extensions Module->CRO Judge AI Extensions Module Detail]] <div class="cro-hr"></div> <h2 class="cro-h2">12.0 Implementation Examples</h2> <div class="cro-guide-image-container"> {embed image: 'Images/Supporting Informations.png', alt: 'Two hands holding a piece of paper in the most uncomfortable way you have ever seen.'} </div> <h3 class="cro-h3">Ex.1 - Mental Approach (Cognitive Modes)</h3> [code] *I try to figure out how this ancient mechanism works.* **MENTAL ROLL** *6/10 - Pattern Recognition (ex. connections, similarities, recurring themes, etc.)* [continue] (AI might narrate: Looking at the gears and symbols, you start noticing similarities to a clockwork design you've seen before. The recurring spiral motifs suggest this follows the same principle as those old Meridian automatons...) <div class="cro-hr"></div> <h3 class="cro-h3">Ex. 2 - Perception Investigation (Domain Focus)</h3> [code] *I examine this abandoned camp carefully.* **PERCEPTION ROLL** *7/10 - Hidden Elements (ex. concealed aspects, overlooked details, etc.)* [continue] (AI might narrate: Most of the obvious details are clear - cold fire pit, scattered supplies - but something catches your eye. One of the tent stakes has been driven in at an odd angle, and there's a small piece of fabric caught underneath, deliberately placed.) <div class="cro-hr"></div> <h3 class="cro-h3">Ex. 3 - Physical vs Manual Distinction</h3> [code] *I attempt to force this rusted door open with my shoulder.* **PHYSICAL ROLL** *4/10 - Rough Attempt (partial progress, noticeable strain)* [continue] (AI might narrate: You slam into the door repeatedly. It shifts slightly in its frame, but your shoulder aches from the impact...) [code] *Instead, I try to work the lock mechanism carefully.* **MANUAL ROLL** *7/10 - Competent Work (solid execution, minor rough edges)* [continue] (AI might narrate: With patient manipulation, you feel the tumblers respond. The lock clicks open, though you notice you've left tiny metal filings around the keyhole.) <div class="cro-hr"></div> <h3 class="cro-h3">Ex. 4 - Using Narrative Techniques</h3> [code] *(I'd like to shift focus to what's outside the window)* *I pause mid-conversation and glance toward the street.* [continue] (AI responds by describing activity outside, transitioning naturally from dialogue to environmental observation) <div class="cro-hr"></div> <h3 class="cro-h3">Ex. 5 - Item Focus Extension</h3> [code] *I examine the merchant's ledger more closely.* **ITEM FOCUS ROLL** *3/10 - Functional Design (ex. intended purpose, how it works, efficiency, etc.)* [continue] (AI might narrate: This isn't just a simple record book. The margins contain a second notation system - coded references that seem to track something other than normal transactions. It's designed for dual purposes.) <div class="cro-hr"></div> <h3 class="cro-h3">Ex. 6 - Inventory Check Extension</h3> [code] *I rummage through my pack to see what supplies I have.* **INVENTORY CHECK** *8/10 - Valuable Discovery (significantly useful items)* [continue] (AI might narrate: Among your basic travel gear, you find a small leather pouch containing silver coins you'd forgotten about, plus a quality rope that's exactly what you need for the climb ahead.) <div class="cro-hr"></div> <h3 class="cro-h3">Ex. 7 - Narrative Momentum Extension</h3> [code] *This conversation is getting stale. I want to shift the energy of this scene.* **NARRATIVE MOMENTUM ROLL** *4/10 - Tense Undercurrent (ex. brewing conflict, unspoken tension, etc.)* [continue] (AI might respond by having the NPC's expression subtly shift, or introducing an interruption that raises stakes, or revealing that someone's been listening...) <div class="cro-hr"></div> <h3 class="cro-h3">Ex. 8 - Main Scale Flexibility</h3> [code] *I try to convince the innkeeper to let us stay despite having no money.* **SOCIAL ROLL** *5/10 - Costly Success (yes, but... with clear drawback)* [continue] (AI might narrate: The innkeeper agrees, but insists you work in the kitchens at dawn to pay off the debt, and makes it clear this arrangement won't be repeated.) <div class="cro-hr"></div> <h3 class="cro-h3">Ex. 9 - All at once - The Abandoned Camp Scenario</h3> [code] *I try to figure out what happened at this abandoned campsite.* **MENTAL ROLL** *6/10 - Pattern Recognition (ex. connections, similarities, recurring themes, etc.)* [continue] (AI narrates: Looking at the scattered belongings and overturned pots, you start recognizing a pattern. This looks similar to the aftermath of those bandit raids you've heard about - hasty departure, but valuables left behind suggests they fled rather than were robbed...) [code] *I examine the campsite more carefully for details.* **PERCEPTION ROLL** *7/10 - Hidden Elements (ex. concealed aspects, overlooked details, etc.)* [continue] (AI continues: Most of the obvious signs are clear, but something catches your eye. One of the tent stakes has been driven in at an odd angle, and beneath it, you spot the corner of something buried...) [code] *I try to force the buried object out with my hands.* **PHYSICAL ROLL** *4/10 - Rough Attempt (partial progress, noticeable strain)* [continue] (AI responds: You claw at the packed earth, making some progress but your fingernails are getting torn and your hands cramped. You've exposed what looks like a small wooden chest, but it's still firmly stuck...) [code] *Instead, I look around the camp for a tool to help dig it out properly.* **MANUAL ROLL** *7/10 - Competent Work (solid execution, minor rough edges)* [continue] (AI narrates: Using a bent tent pole as a makeshift lever, you carefully work the chest free. It comes up cleanly, though the wood shows some scrape marks from your improvised tool...) [code] *(I want to examine this chest closely before opening it)* *I take a moment to study the chest's construction and condition.* **ITEM FOCUS ROLL** *3/10 - Functional Design (ex. intended purpose, how it works, efficiency, etc.)* [continue] (AI responds: This isn't just storage - it's specifically designed as a strongbox. The iron reinforcements and the particular lock mechanism suggest it was made to protect something valuable during travel...) [code] *I check my pack to see if I have anything useful for dealing with this lock.* **INVENTORY CHECK** *8/10 - Valuable Discovery (significantly useful items)* [continue] (AI continues: Among your basic gear, you find a small set of lockpicks you'd forgotten about, wrapped in oiled cloth. Perfect for this type of mechanism...) [code] *The scene feels too quiet. I want to shift the energy here.* **NARRATIVE MOMENTUM ROLL** *4/10 - Tense Undercurrent (ex. brewing conflict, unspoken tension, etc.)* [continue] (AI responds: As you work on the lock, you become aware of being watched. The forest around the clearing feels too still, and you catch the faint sound of a branch creaking as if someone just shifted their weight...) [code] *I call out: "I know you're there. No point hiding." I try to sound confident rather than threatening.* **SOCIAL ROLL** *5/10 - Costly Success (yes, but... with clear drawback)* [continue] (AI concludes: A figure emerges from the treeline - a young woman with a crossbow. "That's my family's chest," she says. "But I suppose since you dug it up, we'll need to negotiate..." She's willing to talk, but clearly sees you as a potential threat to be managed rather than trusted...) <div class="cro-hr"></div> <h2 class="cro-h2">13.0 Notes & Best Practices</h2> The mechanics are only half the story - how you use them matters just as much. <h3 class="cro-h3">Children should not learn math</h3> Apply difficulty modifiers, advantage, stakes, and other adjustments before presenting your final roll to the AI. Showing your calculations can unconsciously bias the AI's interpretation. A clean "7/10" result tells a clearer story than "6+1 for advantage = 7/10." The AI should responds to the narrative weight of the number, not the arithmetic that produced it. <h3 class="cro-h3">You ”attempt” to drink water</h3> Your action descriptions should emphasize trying rather than succeeding. "I attempt to..." keeps the AI focused on interpreting your roll rather than assuming success. This small linguistic choice reinforces that outcomes depend on dice, not declarations. <h3 class="cro-h3">Blood is narrative fuel, and you the hell-engine of decent writing</h3> Low rolls aren't punishment - they're story engines. Some of my most memorable sessions have emerged from spectacular failures that forced creative solutions. Let the AI explore consequences rather than trying to minimize or rationalize poor results. The system works best when you trust dramatic setbacks to create interesting problems. <h3 class="cro-h3">Muscles, brains, or spirit of determination?</h3> Even when using the Main Outcome Scale, declaring your skill type helps the AI frame the action appropriately. "Physical Roll" versus "Social Roll" gives context for how that "6/10 - Weak Success" should manifest, even with identical outcome language. <h3 class="cro-h3">Extensionally restrain yourself</h3> Introduce optional extensions gradually. The core system already provides plenty of narrative variety, and too many simultaneous mechanics can overwhelm both you and the AI. Pick extensions that solve specific problems you're actually encountering, not ones that seem theoretically interesting. Unless your goal is to experiment! Then go right ahead. I’m not going to stop you. <h3 class="cro-h3">Artificial early onset dementia</h3> AI memory has limits, sometimes very strict ones at that, so focus on immediate action and outcome rather than expecting long-term state tracking. The system's simplicity stems from this limitation, yes, but that’s also a feature: we are working with the current LLM’s capabilities rather than against them. Keep complex details in your own notes if they matter for future sessions. <div class="cro-passage-nav"> <div><span class="nav-prev">[[Previous: Extensions Collection->CRO Extensions Collection]]</span></div> <div class="nav-hub">[[Quick Reference Hub]]</div> </div>

<h1 class="cro-h1">Judge System Setup</h1> This page contains the detailed setup instructions and the full system prompt for the optional Judge AI. <h2 class="cro-h2">11.2 Set-up</h2> Create a separate conversation with an AI using: * The prompt below * The "CRO Reference Document", dowloadable below This AI will serve as your impartial referee for all CRO rolls. Simply provide it with your intended action and dice roll, and it will return a properly formatted result for you to use with your main storytelling AI. <h4 class="cro-h4">Input Format:</h4> [code] "I attempt to [action]. [Any relevant context]. [Your d10 roll result]" [continue] <h4 class="cro-h4">Example:</h4> [code] "I attempt to convince the guard to let me pass. He looks suspicious and I'm clearly not supposed to be here. 7" [continue] <h2 class="cro-h2">11.3 Judge AI System Prompt</h2> Copy and paste this exact prompt when setting up your Judge AI: <div class="cro-code-block-container" data-copyable-container="true"> <button class="cro-copy-button" data-target-id="judge-prompt-content-setup" aria-label="Copy code snippet" tabindex="-1"></button> <div class="cro-code-block" id="judge-prompt-content-setup"> \# CRO Judge AI System Prompt You are the \*\*Judge AI\*\* for the \{Char\}, \[Roll\] for \[Outcome\] \(CRO\) system\. Your role is to make impartial mechanical rulings for player actions, then generate dice rolls and format results\. You do NOT create narrative responses \- only mechanical judgments\. \#\# Your Process 1\. \*\*Receive\*\*: Player\'s action declaration and their d10 roll result 2\. \*\*Analyze\*\*: Assess difficulty and determine appropriate skill category 3\. \*\*Calculate\*\*: Apply difficulty modifier to the player\'s roll 4\. \*\*Format\*\*: Return standardized CRO result for the Narrative AI \#\# Difficulty Assessment Framework \#\#\# Base Difficulty \(Default\: Moderate \+0\) \*\*Trivial \(\+3\)\*\*: Action is almost guaranteed to succeed in normal circumstances \- Examples\: Opening an an unlocked door, walking across flat ground, basic conversation with friendly person \*\*Easy \(\+1\)\*\*: Minor challenge with favorable conditions \- Examples\: Climbing a ladder, convincing someone of something they\'re already inclined to believe \*\*Moderate \(\+0\)\*\*: Standard challenge requiring competence \- Examples\: Picking a basic lock, persuading a neutral party, climbing a tree \*\*Difficult \(\-1\)\*\*: Significant challenge requiring skill/luck \- Examples\: Master\-level locks, convincing hostile enemies, scaling a cliff \*\*Formidable \(\-3\)\*\*: Extreme challenge, failure is likely \- Examples\: Master\-level locks, convincing hostile enemies, impossible physical feats \#\#\# Contextual Modifiers \*\*Situational Advantages \(\+1 to \+2\)\*\*: \- Ideal tools/equipment available \- Environmental conditions favor the action \- Recent relevant success provides confidence \- Target is predisposed to cooperate \*\*Situational Disadvantages \(\-1 to \-2\)\*\*: \- Poor tools or missing equipment \- Environmental conditions hinder action \- Recent relevant failure creates doubt \- Active opposition or hostility \- Time pressure or stress \- Physical\-mental impairment \#\# Skill Category Assignment Assign the most appropriate skill based on the primary method\-approach\: \- \*\*PHYSICAL\*\*: Strength, athletics, endurance, combat maneuvers \- \*\*MANUAL\*\*: Dexterity, crafting, fine motor skills, delicate tasks \- \*\*MENTAL\*\*: Intelligence, knowledge, problem\-solving, analysis \- \*\*SOCIAL\*\*: Persuasion, deception, intimidation, charm \- \*\*PERCEPTION\*\*: Awareness, investigation, searching, noticing \- \*\*FORTUNE\*\*: Pure luck, timing, random circumstance \- \*\*STEALTH\*\*: Sneaking, hiding, moving silently \(replaces Physical\-Manual for these\) \#\# Output Format \````\`\```` \[2\-3 sentences explaining the contextual factors that influence the overall difficulty\: favorable conditions, hindrances, environmental factors, character state, etc\.\] \````\`\```` \*\*DIFFICULTY ASSESSMENT\*\*: \[Trivial\/Easy\/Moderate\/Difficult\/Formidable\] \*\*SKILL CATEGORY\*\*: \[Chosen skill\] \*\*PLAYER\'S ROLL\*\*: \[Player\'s provided d10 result\] \*\*FINAL RESULT\*\*: \[Result after applying difficulty modifier to player\'s roll\] \-\-\- \*\*FORMATTED ACTION\*\* \````\`\```` \*I attempt to \[restate player\'s action\]\. \*\*\[SKILL\] ROLL\*\* \*\[Final Result\]\/10 \- \[Appropriate outcome description from scale\]\* \````\`\```` \#\# Key Principles 1\. \*\*Be Impartial\*\*: Don\'t favor player success or failure \- assess objectively 2\. \*\*Use Context\*\*: Consider the situation described, but don\'t add details not provided 3\. \*\*Default Conservative\*\*: When unsure between two difficulties, choose the harder one 4\. \*\*Stay Mechanical\*\*: Don\'t create narrative elements \- only assess mechanics 5\. \*\*Be Consistent\*\*: Similar actions in similar contexts should get similar difficulty ratings \#\# Reference Material You have access to the complete CRO guide document\. Follow all rules, scales, and guidelines provided in that document\. Use the appropriate outcome scales \(Main Outcome Scale or Skill\-Specific Threshold Scales\) as defined in the guide\. \-\-\- \*\*Remember\*\*: You are making mechanical rulings only\. The Narrative AI will handle all story elements, character responses, and world\-building based on your mechanical result\. </div> </div> <div class="cro-passage-nav"> <div><span class="nav-prev">[[Previous: Advanced Tools & Examples->CRO Advanced Tools & Examples]]</span></div> <div class="nav-hub">[[Quick Reference Hub]]</div> </div>

<div class="cro-detail-top-nav"> <div class="nav-hub">[[Quick Reference Hub]]</div> </div> <h1 class="cro-h1">Mental Scale Detail</h1> <h3 class="cro-h3">Mental Scale</h3> (Problem Solving, Intelligence, Academic Knowledge, Remembering Facts or Learned Information) [code] *1/10 - Analysis Paralysis (ex. overthinking, endless loops, decision paralysis, etc.)* [continue] [code] *2/10 - Tunnel Vision (ex. fixation, missing alternatives, rigid thinking, etc.)* [continue] [code] *3/10 - Emotional Bias (ex. wishful thinking, fear-driven reasoning, prejudice, etc.)* [continue] [code] *4/10 - Scattered Focus (ex. jumping between ideas, unfocused effort, etc.)* [continue] [code] *5/10 - Logical Analysis (ex. step-by-step reasoning, systematic breakdown, etc.)* [continue] [code] *6/10 - Pattern Recognition (ex. connections, similarities, recurring themes, etc.)* [continue] [code] *7/10 - Practical Application (ex. real-world solutions, functional thinking, etc.)* [continue] [code] *8/10 - Creative Synthesis (ex. novel combinations, innovative approaches, etc.)* [continue] [code] *9/10 - Abstract Reasoning (ex. theoretical concepts, pure logic, philosophy, etc.)* [continue] [code] *10/10 - Intuitive Leap (ex. sudden understanding, gut insights, breakthrough moments, etc.)* [continue] Things did get tricky with this one. Mental actions in AI narratives have a fundamental problem: there's usually no objective information waiting to be discovered. When you roll to "figure out the puzzle," the AI hasn't actually designed a puzzle with a specific solution - it never thinks that much ahead, it simply can’t. I kept running into situations where the AI would either make up jarring new facts, or the roll would become either meaningless, or the whole category a catch-all, capable of being used for everything. It still kind of is the latter, but eventually I stopped asking "what do you learn" and started asking "how does your mind approach this?" The 10 outcomes ahead represent different cognitive modes for your character – from analysis paralysis to intuitive leaps. This actually gives the AI something concrete to work with without forcing it to necessarily invent setting details. <div class="cro-passage-nav"> <div><span class="nav-prev">[[Previous: Skill-Specific Scales->CRO Skill-Specific Scales]]</span></div> <div class="nav-hub">[[Quick Reference Hub]]</div> </div>

config.style.page.verticalAlign: 'top' config.footer.right: "{link to: 'Title Screen', label: 'Home'} | {restart link}" -- <div class="cro-detail-top-nav" style="text-align:center; margin-bottom: 2em;"> <span class="nav-hub">[[Quick Reference Hub]]</span> | <span class="nav-hub">[[Title Screen]]</span> </div> <h1 class="cro-h1" style="text-align:center; font-size: 36px; margin-bottom: 1.5em;">{Char}, [Roll] for [Outcome]</h1> <h1 class="cro-h1">1.0 Introduction</h1> Ever find an AI chatbot a bit too... agreeable? You state an intention, and the AI complies smoothly, missing chances for dramatic tension or unexpected turns. The <span class="cro-inline-code">{Char}, [Roll] for [Outcome]</span> system, or <span class="cro-inline-code">CRO</span> for short, aims to fix this for you. <div class="cro-guide-image-container cro-image-fullwidth"> {embed image: 'Images/Introduction Part 1.png', alt: 'The user and Roxy having a bit of a misunderstanding.'} </div> Here’s how it works: when your character attempts to do something, **you** roll a {tooltip: 'd10', text: 'A ten-sided die.'} and present the result using the simple <span class="cro-inline-code">CRO</span> format provided below. The AI then uses that number and the directions provided inside the Outcome Scales, to guide its response. <div class="cro-guide-image-container cro-image-fullwidth"> {embed image: 'Images/Introduction Part 2.png', alt: 'The CRO guide is used, and once again all is good in the world.'} </div> Suddenly, success isn't automatic. Your roll might lead to triumph, a costly setback, or something in between – all fascinating narrative threads not unlike the ones found through the medium of tabletop roleplaying games. The AI in control of your roleplaying session - your very own artificial DM - is encouraged to explore far more interesting narrative paths than just going along with its own vague interpretation of what you may like or don’t. There are no complex character sheets, inventories or statistics to track; your <span class="cro-inline-code">{Char}</span> would not be able to effectively remember any of those. Not for long, anyway. And this is not meant to be homework! So it's just you. You, your character's intent, a straightforward d10 roll, and a clear format to guide the story. <div class="cro-guide-image-container"> {embed image: 'Images/Introduction Final.png', alt: 'Roxy and the user wish you good luck!'} </div> <h2 class="cro-h2" style="margin-top:1.5em;">1.1 Where to use</h2> CRO works with any AI that can engage in conversational roleplay, and requires absolutely no setup on the AI's side. No special system prompts, no configuration - you simply start using the format and the AI will naturally adapt to it. If you’re in need of a d10 dice, just search “Dice Roller” into google, there’s an easter egg for it. I've primarily developed and tested this system on **character roleplay platforms**: first on Character.ai (that was like a year ago, when it was still kind of decent!), then Spicychat.ai and Janitor.com. These platforms are designed around chatting with specific AI personas in ongoing conversations, which tends to work particularly well with CRO's narrative focus. There are also some good world-based “characters" there too: chatbots designed to create a world for your own character to work and act inside of. You'll find plenty of them to experiment with. There’s also **general AI assistants** like ChatGPT, Claude, Gemini, or Deepseek, of course - those also work just fine. You can start a conversation by describing your character and setting, then begin using the CRO format for actions. These platforms often bring more sophisticated reasoning to the table, though they may need a sentence or two of context about what you're doing initially. Point is, you're not changing how the AI works - you're changing how you present information to it. The format itself teaches the AI what you want: that dice results should guide outcomes, that low rolls mean complications, and that high rolls mean smooth success. Most AI systems pick this up within a few exchanges, if not right away. I personally lean towards character platforms because they're built and fine-tuned for sustained narrative interaction, but honestly, experimentation is part of the fun. Try it wherever you normally interact with LLMs. See how it feels! <div class="cro-hr"></div> <h1 class="cro-h1">1.2 Quickstart Guide</h1> Want to jump right in? Here's everything you need to start using CRO: <h3 class="cro-h3">The Basic Process:</h3> 1. **Declare what you're attempting** - not what you succeed at, but what you're trying to do 2. **Roll a d10** - this determines how things actually go 3. **Apply any difficulty modifiers mentally** (Easy: +1, Difficult: -1, etc.) 4. **Find your result on the Main Outcome Scale** (see Section 3.1 below) 5. **Present it to the AI using this format:** [code] *I attempt to [action].* **[SKILL] ROLL** *[Final Result]/10 - [Outcome Description]* [continue] <h3 class="cro-h3">Example</h3> You want to climb a crumbling tower wall. Instead of just saying "I climb up," you declare your attempt and roll: [code] *I attempt to scale the crumbling tower wall.* **PHYSICAL ROLL** *4/10 - Setback Failure (no, but... with silver lining)* [continue] Now the AI has clear direction: you don't make it up, but something interesting happens. Maybe you fall but discover a hidden entrance, or attract the attention of someone who can help. **That's it!** You've just turned a simple success into a dramatic story moment. The AI will handle the rest based on your roll. Instead of the AI guessing what you want or defaulting to success, the dice created genuine uncertainty. That 4/10 could have been a 9/10 triumph or a 1/10 disaster - and now you're both curious to see how this setback becomes the next part of your story. <div class="cro-hr"></div> <h2 class="cro-h2">1.3 Browser Script Available</h2> For users of Spicychat.ai, Character.ai, or JanitorAI, there's a ↪ [[browser script->CRO Helper Script Detail]] that provides a convenient graphical interface directly on those platforms. It handles all the formatting and dice rolling automatically, so you can focus on the storytelling rather than the mechanics. The script is completely optional - everything in this guide works perfectly well with manual rolling and copy-pasting. But if you're spending a lot of time on those platforms, it might save you some effort. <div class="cro-hr"></div> For skill categories, extensions, and advanced techniques, keep reading. But this basic approach will immediately make your AI conversations more dramatic and unpredictable. <div class="cro-hr"></div> <h1 class="cro-h1">Part 1: Core Mechanics</h1> <div class="cro-guide-image-container"> {embed image: 'Images/Basic Action and Rolling.png', alt: 'A d10 dice is thrown - a 5. The CRO format is used accordingly.'} </div> This section outlines the fundamental principles of the {tooltip: 'CRO', text: '{Char}, [Roll] for [Outcome]'} system, covering how to declare actions, make rolls, and interpret their basic outcomes. <h2 class="cro-h2">2.0 Basic Action & Rolling</h2> <h3 class="cro-h3">2.1 Declaring Intent & Skill Category</h3> 1. Player declares action and skill category. 2. Player rolls d10, applies mental modifiers, presents final <span class="cro-inline-code">[Result]/10</span>. 3. AI/DM uses the corresponding Threshold Scale below for outcome interpretation. <h3 class="cro-h3">Format</h3> [code] *I attempt to [action].* **[SKILL] ROLL** *[Result]/10 - [Outcome Description]* [continue] <h3 class="cro-h3">Example</h3> [code] *I attempt to bake a chocolate soufflé.* **CULINARY ROLL** *7/10 - Partial Success (limited progress)* [continue] <div class="cro-hr"></div> <h2 class="cro-h2">3.0 Interpreting the Roll: Outcome Scales</h2> <div class="cro-guide-image-container"> {embed image: 'Images/Scales Outcome.png', alt: 'A number is rolled, and the scales are used accordingly.'} </div> The d10 roll determines the raw result, but the Outcome Scales give that number narrative meaning. {tooltip:'CRO', text:'{Char}, [Roll] for [Outcome]'} offers two approaches: * **3.1 Main Outcome Scale:** This is the default, universal scale. Its broad outcome descriptions (Failure, Mixed Result, Success, etc.) provide a lot of flexibility to the AI; in how it interprets the action’s outcome given the context and the declared Skill. * Use this scale when you want to prioritize surprising or unexpected narrative developments arising directly from the roll number. * Works best with very powerful models (ex. Deepseek V3/R1). * **3.2 Skill-Specific Threshold Scales:** These optional scales offer more thematically tailored outcome descriptions for certain common skill types (Physical, Social, Perception, etc.). * Use these when you want the outcome interpretation to more strongly reflect the specific nature of the skill being employed, leading to potentially more predictable, but coherent results aligned with that skill domain. * Works best with smaller, less advanced, and less knowledgeable models. **Choosing Your Scale:** You, the player, can decide which scale best suits your intention for any given roll. You might default to the Main Scale for most actions, but then choose a Specific Scale when attempting a particularly characteristic task (e.g., using the Stealth Scale for a sneaking attempt). Try it out! See what method fits your roleplaying style, or simply works best. Whatever the case, do keep in mind that the AI needs to be kept aware that a low roll means a failure, a high one a success, and a middling one something in between - without the use of Outcomes, it will inevitably forget this, and drift towards giving you a positive outcome regardless of the dice number. <h3 class="cro-h3">3.1 Main Outcome Scale (General Use)</h3> <div class="cro-guide-image-container"> {embed image: 'Images/Main Scale.png', alt: 'Roxy ponders a bunch of funny looking shapes. They signify the entries. A metaphor, you see. You get it.'} </div> [code] *1/10 - Critical Failure (significant setback)* [continue] [code] *2/10 - Major Failure (clear consequence)* [continue] [code] *3/10 - Simple Failure (no progress)* [continue] [code] *4/10 - Poor Attempt (close, but falls short)* [continue] [code] *5/10 - Complication (a problem emerges)* [continue] [code] *6/10 - Minor Success (limited progress)* [continue] [code] *7/10 - Weak Success (achieves goal, barely)* [continue] [code] *8/10 - Clear Success (achieves goal)* [continue] [code] *9/10 - Strong Success (achieves goal cleanly)* [continue] [code] *10/10 - Critical Success (exceeds expectations)* [continue] This is CRO's foundational scale – the default when you're not using a skill-specific approach. Its 10 outcomes use deliberately broad language: that gives the AI maximum interpretive freedom while still providing clear guidance about the action's general trajectory. It’s universal: whether you're attempting to convince a guard, climb a wall, or solve a riddle, these outcome descriptions adapt naturally to the context. I designed this as the primary scale because it captures the essential dramatic beats without constraining too much how those beats manifest. Once again, these are relatively open-ended descriptions, and will work best with more powerful AI models: ones that can take a broad narrative direction and develop it meaningfully, if not creatively. <h3 class="cro-h3">3.2 Skill-Specific Scales</h3> <h4 class="cro-h4">Physical Scale</h4> (Strength, Athletics, Endurance, Combat Maneuvers) <div class="cro-guide-image-container"> {embed image: 'Images/Physical Scale.png', alt: 'A buff arm, flexing.'} </div> [code] *1/10 - Harsh Consequence (clear physical cost or injury)* [continue] [code] *2/10 - Strained Effort (body struggles, fatigue sets in)* [continue] [code] *3/10 - Forced Failure (pushing through but can't complete)* [continue] [code] *4/10 - Rough Attempt (partial progress, noticeable strain)* [continue] [code] *5/10 - Taxing Success (achieves goal but at physical cost)* [continue] [code] *6/10 - Adequate Performance (body cooperates, gets it done)* [continue] [code] *7/10 - Steady Control (confident execution)* [continue] [code] *8/10 - Strong Performance (body responds well)* [continue] [code] *9/10 - Smooth Execution (effortless feel)* [continue] [code] *10/10 - Perfect Form (exactly as intended, if not better)* [continue] Physical actions present a unique challenge: they carry real consequences that can impact your entire ongoing narrative. A failed climb isn't just "you don't get up the wall" – it might mean injury, exhaustion, or equipment damage that matters for the next hour of play. That's why I settled on 10 outcomes here, giving you enough granularity to distinguish between "winded but fine" and "twisted ankle that'll slow you down." The scale focuses on the relationship between your character's intention and their body's response rather than just success/failure, because in my testing, that relationship created far more interesting ongoing complications than simple binary outcomes. <h4 class="cro-h4">Manual Scale</h4> (Dexterity, Crafting, Fine Motor Skills, Delicate Tasks) <div class="cro-guide-image-container"> {embed image: 'Images/Manual Scale.png', alt: 'Two hands working on a engine.'} </div> [code] *1/10 - Complete Mishap (task goes wrong, clear setback)* [continue] [code] *2-3/10 - Sloppy Work (barely functional, obvious flaws)* [continue] [code] *4-5/10 - Rough Execution (works but crude, noticeable imperfections)* [continue] [code] *6-7/10 - Competent Work (solid execution, minor rough edges)* [continue] [code] *8-9/10 - Clean Execution (smooth work, well-handled)* [continue] [code] *10/10 - Precise Control (exactly as intended, no wasted motion)* [continue] Originally, I had manual skills lumped in with physical ones, but I quickly realized they needed their own approach. When you're picking a lock or crafting something, the question isn't really "did it work" but "how well did you execute it?" A crude lockpick job might open the door but leave obvious scratches, while masterful work leaves no trace. Seven outcomes gives you enough range to capture that spectrum without bogging down the narrative in excessive detail. The focus on execution quality over mere functionality emerged from countless sessions where the difference between "it works" and "it works elegantly" drove the next scene's direction. <h4 class="cro-h4">Mental Scale</h4> (Problem Solving, Intelligence, Academic Knowledge, Remembering Facts or Learned Information) <div class="cro-guide-image-container"> {embed image: 'Images/Mental Scale.png', alt: 'Roxy, thinking through a problem.'} </div> [code] *1/10 - Analysis Paralysis (ex. overthinking, endless loops, decision paralysis, etc.)* [continue] [code] *2/10 - Tunnel Vision (ex. fixation, missing alternatives, rigid thinking, etc.)* [continue] [code] *3/10 - Emotional Bias (ex. wishful thinking, fear-driven reasoning, prejudice, etc.)* [continue] [code] *4/10 - Scattered Focus (ex. jumping between ideas, unfocused effort, etc.)* [continue] [code] *5/10 - Logical Analysis (ex. step-by-step reasoning, systematic breakdown, etc.)* [continue] [code] *6/10 - Pattern Recognition (ex. connections, similarities, recurring themes, etc.)* [continue] [code] *7/10 - Practical Application (ex. real-world solutions, functional thinking, etc.)* [continue] [code] *8/10 - Creative Synthesis (ex. novel combinations, innovative approaches, etc.)* [continue] [code] *9/10 - Abstract Reasoning (ex. theoretical concepts, pure logic, philosophy, etc.)* [continue] [code] *10/10 - Intuitive Leap (ex. sudden understanding, gut insights, breakthrough moments, etc.)* [continue] Things did get tricky with this one. Mental actions in AI narratives have a fundamental problem: there's usually no objective information waiting to be discovered. When you roll to "figure out the puzzle," the AI hasn't actually designed a puzzle with a specific solution - it never thinks that much ahead, it simply can’t. I kept running into situations where the AI would either make up jarring new facts, or the roll would become either meaningless, or the whole category a catch-all, capable of being used for everything. It still kind of is the latter, but eventually I stopped asking "what do you learn" and started asking "how does your mind approach this?" The 10 outcomes ahead represent different cognitive modes for your character – from analysis paralysis to intuitive leaps. This actually gives the AI something concrete to work with without forcing it to necessarily invent setting details. <h4 class="cro-h4">Social Scale</h4> (Persuasion, Deception, Intimidation, Charm, Negotiation) <div class="cro-guide-image-container"> {embed image: 'Images/Social Scale.png', alt: 'A very intense chat.'} </div> [code] *1-2/10 - Momentum Lost (interaction creates resistance or withdrawal)* [continue] [code] *3-4/10 - Awkward Exchange (stilted, uncomfortable social flow)* [continue] [code] *5-6/10 - Neutral Transaction (stalling, neither builds nor damages social standing)* [continue] [code] *7-8/10 - Strong Rapport (builds connection and cooperation)* [continue] [code] *9-10/10 - Lasting Impact (memorable interaction, doors open)* [continue] Social interactions... I tend not to roll for these, actually, I prefer to talk in character, and try to convince/intimidate the chatbot that way. So I actually haven't tested this scale too hard! I'm open to improvements, as I am with the rest of the system's components. Five outcomes captures the essential range from "actively damages the relationship" to "creates lasting positive impression" without overthinking it. During development, I noticed that focusing on social momentum and interpersonal energy worked better than specific persuasion mechanics. For example, rolling high doesn't necessarily mean "they agree with you," but rather "you've created positive social energy, and that may open doors." This keeps the AI focused on relationship dynamics rather than forcing specific character decisions. <h4 class="cro-h4">Perception Scale</h4> (Awareness, Investigation, Searching, Noticing, Sensing Motives) <div class="cro-guide-image-container"> {embed image: 'Images/Perception Scale.png', alt: 'A magnifying glass inspecting the world around itself.'} </div> [code] *1-2/10 - Physical State (ex. appearance, condition, immediate qualities, etc.)* [continue] [code] *3-4/10 - Functional Aspect (ex. role, purpose, capabilities, etc.)* [continue] [code] *5-6/10 - Contextual Clues (ex. connections, significance, anomalies, etc.)* [continue] [code] *7-8/10 - Hidden Elements (ex. concealed aspects, overlooked details, etc.)* [continue] [code] *9-10/10 - Temporal Traces (ex. history, recent changes, accumulated effects, etc.)* [continue] This 5-outcome scale works differently from the others – instead of measuring how well you perceive, it determines what type of perception you apply. Each outcome represents a different investigative lens: you might focus on physical condition, functional purpose, contextual relationships, hidden features, or traces of recent activity. This approach emerged from a practical problem with traditional perception mechanics: telling the AI that you "notice more things" often produced vague atmospheric descriptions that gave players nothing concrete to pursue. Oftentimes, the instruction was completely ignored. By specifying the type of observation, the scale consistently delivers actionable information. Rolling for "functional aspect" when examining a door gives you insights about its purpose and mechanisms, while "temporal traces" might reveal recent use patterns – both useful for planning your next action, I feel. <h4 class="cro-h4">Fortune Scale</h4> (Luck, Chance, Timing, Random Circumstance) <div class="cro-guide-image-container"> {embed image: 'Images/Luck Scale.png', alt: 'A sequence of clover leaves growing. At the end, a four-leafed one.'} </div> [code] *1/10 - Bad Timing (unfavorable coincidence)* [continue] [code] *2-5/10 - Unlucky Turn (circumstances hinder, disadvantage)* [continue] [code] *6-9/10 - Lucky Break (favorable circumstance, advantage)* [continue] [code] *10/10 - Perfect Timing (major positive turn of events)* [continue] Fortune is pure circumstance, isn't it? Things either go your way or they don't. I originally had a complex 10-outcome scale with dramatic language like "calamitous misfortune," but found it was pulling the AI into over-the-top territory. Bad timing is just bad timing; lucky breaks are just lucky breaks. Sometimes the most elegant design choice is restraint, and 4, well-defined outcomes work well enough. <h4 class="cro-h4">Stealth Scale</h4> (Sneaking, Hiding, Moving Silently, Lockpicking) <div class="cro-guide-image-container"> {embed image: 'Images/Stealth Scale.png', alt: 'A lock being picked.'} </div> [code] *1/10 - Fully Exposed (immediate detection, alarm raised)* [continue] [code] *2/10 - Clearly Spotted (presence and intent obvious)* [continue] [code] *3/10 - Obviously Noticed (seen but intent unclear)* [continue] [code] *4/10 - Suspicion Raised (traces left, heightened awareness)* [continue] [code] *5/10 - Minor Disturbance (small signs noticed)* [continue] [code] *6/10 - Barely Avoided (close call, narrowly unnoticed)* [continue] [code] *7/10 - Successfully Hidden (avoided direct detection)* [continue] [code] *8/10 - Cleanly Unnoticed (no awareness triggered)* [continue] [code] *9/10 - Seamless Passage (no signs of presence)* [continue] [code] *10/10 - Perfect Concealment (completely undetected, no trace)* [continue] Stealth demands granularity because detection operates on a gradient, and small differences have big consequences. There's a meaningful distinction between "guards are suspicious" and "guards are actively searching," and your options change dramatically between those states. These 10 outcomes let the AI calibrate responses precisely, from "completely blown" to "ghosted through undetected." I kept this one linear because stealth is fundamentally about degrees of concealment – it's one of the few skills where more complex categorization would actually reduce clarity. <div class="cro-hr"></div> <h2 class="cro-h2">4.0 Modifying Rolls & Outcomes (Core)</h2> <div class="cro-guide-image-container"> {embed image: 'Images/Modifying Rolls.png', alt: 'A hand adjusting some parameters.'} </div> <h3 class="cro-h3">4.1 Difficulty Modifiers</h3> Apply mental modifiers to the final roll result. Keep in mind that you do not need to show them to the AI, just adjust the final number when you present it: [code] - Trivial: +3 - Easy: +1 - Moderate: +0 - Difficult: -1 - Formidable: -3 [continue] For example, if you rolled a 6 on a Difficult task (-1), present it as <span class="cro-inline-code">5/10</span>. <h3 class="cro-h3">4.2 Narrative Consequences</h3> * A recent success in a similar area could grant you a **purely** mental +2 or Advantage (roll twice, keep higher). * A recent failure could impose a **purely** mental -2 or Disadvantage (roll twice, keep lower). You simply add or subtract from your final presentation. Don’t show the math to the AI nor, in the case of advantage/disadvantage, inform it that you’ve used advantage/disadvantage - just the final number you end up. <div class="cro-hr"></div> <h2 class="cro-h2">5.0 Time Progression</h2> <div class="cro-guide-image-container"> {embed image: 'Images/Time Roll.png', alt: 'A hand holding up a clock. Yes, it does look like the doomsday clock. This is very deep.'} </div> Narratives often involve periods where little of immediate consequence happens, or where significant change occurs gradually over time. Instead of playing out every single moment in what is occasionally excruciating detail, you should be allowed to "skip forward," and to move the story to the next point of interest. This maintains pacing, allows for character development or environmental shifts during the intervening time, and focuses gameplay on significant events. Use the following methods to manage the passage of time. <h3 class="cro-h3">5.1 Method A: Simple Declaration</h3> Allows the player to specify an intended duration for a time skip. In this specific case, the player doesn’t necessarily have to roll anything. [code] **[approximate duration, e.g., 'DAYS', 'WEEKS', 'MONTHS'] LATER.** [continue] <h3 class="cro-h3">5.2 Method B: Random Duration</h3> This method is useful when you want to jump forward an indeterminate amount of time, letting chance dictate how long passes before the narrative picks up again. Simply roll a d10 and consult the scale below to see the duration of the time skip. <h4 class="cro-h4">Format</h4> [code] **TIME ROLL** *[Result]/10 - [Time Jump]* **[UNIT OF TIME] LATER** [continue] <h4 class="cro-h4">Random Duration Scale</h4> [code] *1-3: Days later* *4-6: Weeks later* *7-8: Months later* *9-10: Years later* [continue] <h3 class="cro-h3">5.3 Method C: Time & Fortune</h3> Allows the player to describe the character of a chosen, or random, span of time, providing narrative flavour without prejudging outcomes or significant events. Offers an alternative to the basic {tooltip: 'TIME ROLL', text: 'See Method B: Random Duration'} for a more directed pacing (a similar deal to the Skill-Specific scales), allowing the player to choose to characterize the period as broadly favourable or unfavourable. <h4 class="cro-h4">Format</h4> [code] **[UNIT OF TIME] LATER** **TIME ROLL** *[Result]/10 - [Time & Fortune Scale Outcome]* [continue] <h4 class="cro-h4">Time & Fortune Scale</h4> [code] *1/10 - Deep Troubles (Time passed under notably difficult background conditions)* [continue] [code] *2/10 - Mostly Downs (The period felt marked by a generally negative trend)* [continue] [code] *3/10 - Problems Mount (Minor issues seemed to consistently arise or worsen)* [continue] [code] *4/10 - Felt Resisted (The period generally felt resistant or effortful)* [continue] [code] *5/10 - Slight Drag (Progress or stability felt subtly held back during this time)* [continue] [code] *6/10 - Slight Boost (Progress or stability felt subtly helped along during this time)* [continue] [code] *7/10 - Easy Flow (The period generally felt smooth or cooperative)* [continue] [code] *8/10 - Things Align (Minor opportunities seemed to consistently arise or improve)* [continue] [code] *9/10 - Mostly Ups (The period felt marked by a generally positive trend)* [continue] [code] *10/10 - Great Fortune (Time passed under notably favourable background conditions)* [continue] [continue] Characterizing entire time periods required a full 10-outcome range to capture the subtlety between "slightly difficult" and "persistently troubled" periods. Unlike moment-to-moment fortune rolls, this scale affects background conditions over extended periods, so the extra granularity helps the AI calibrate how that time period should feel, or how it would’ve felt, without mandating specific events. The scale shapes atmosphere, and so the plot, but indirectly. <div class="cro-hr"></div> <h1 class="cro-h1">Part 2: Optional Extensions</h1> <div class="cro-guide-image-container"> {embed image: 'Images/Extensions Introduction.png', alt: 'A connective tree made of asterisks and text boxes sprouting from your hand.'} </div> <h2 class="cro-h2">6.0 Introduction to Extensions</h2> The Core Mechanics provide the foundation for the system. Extensions are optional modules you can add to introduce more specific rules or narrative effects for particular situations. Think of them as tools in a toolbox – use the ones that enhance your desired style of play and ignore the ones that don't. You can decide which extensions are active at the start of a session, or even introduce one mid-game if a situation calls for it. They are designed to be self-contained and integrate smoothly with the core d10 roll system. <h2 class="cro-h2">7.0 Extensions: Action Modifiers & Enhancements</h2> <div class="cro-guide-image-container"> {embed image: 'Images/Modifiers And Enhancements.png', alt: 'Three hardworking hands on a graph doing their best to rationalize the roll number in their favour lmao'} </div> This first set of extensions focus on modifying how individual actions work. <h3 class="cro-h3">7.1 Stakes</h3> Raise or lower the stakes before a roll (mentally) by 1-3: <span class="cro-inline-code">- Raising Stakes (+1 to +3)</span> * Higher payoff if successful, worse outcome if you fail <span class="cro-inline-code">- Lowering Stakes (-1 to -3)</span> * Softer fail consequences, but capped success If roll ≥ 5, add the chosen “stakes” number to your final result (up to 10). If roll < 5, subtract the “stakes” number (down to 1). You do this mentally and only show the final adjusted roll. Sometimes the standard difficulty modifiers don't capture what you're really after. Stakes let you amplify both the potential reward and the potential cost of an action without changing the basic difficulty. This emerged from situations where I wanted to attempt something that wasn't necessarily harder, but where the consequences mattered more. Climbing a wall during a casual exploration versus climbing that same wall while being chased by guards - same physical challenge, completely different stakes. The mechanic lets you push for bigger successes while accepting that failures will hurt more, creating that "high risk, high reward" feeling. It makes certain moments feel memorable, though, to be fair, I don’t use it often. This is a common pattern along these extensions - that's the point. You use the ones that feel right to you, or the situation at hand. <div class="cro-hr"></div> <h3 class="cro-h3">7.2 Pushing Limits</h3> Allows a character to attempt an action far beyond their normal capabilities or under extreme duress by accepting a significantly higher risk of failure and more severe consequences. <h4 class="cro-h4">Format</h4> [code] I attempt to push my limit by [describe extreme action]. **[SKILL] ROLL** *[Result]/10 - [Pushing Limits Outcome]* [continue] <h4 class="cro-h4">Pushing Limits Scale</h4> The player must use this harsher scale when "Pushing Limits" is invoked, reflecting the mentally increased difficulty threshold. It replaces the Main Outcome Scale for the given roll’s outcome. [code] *1/10 - Severe Backfire (major negative consequence)* [continue] [code] *2-3/10 - Costly Failure (significant setback, clear price paid)* [continue] [code] *4-5/10 - Painful Attempt (falls short, noticeable strain or cost)* [continue] [code] *6-7/10 - Strained Success (goal met, but with visible effort)* [continue] [code] *8-9/10 - Hard-Won Achievement (success despite the odds)* [continue] [code] *10/10 - Breakthrough Performance (exceeds expectations)* [continue] When your character attempts something far beyond normal capability, consequences should be harsher than usual, but I learned that overly dramatic language ("catastrophic failure") pushed the AI toward melodramatic responses. Also, who wants to use a scale that just destroys your character? Sure, it's narratively interesting... for a quick and brutal ending to your session, that is. This 7-outcome scale compresses the success range deliberately – most attempts to push limits should be difficult, with only the highest rolls achieving clean success. This reflects the inherent danger of exceeding normal boundaries while keeping it grounded. <div class="cro-hr"></div> <h3 class="cro-h3">7.3 Skill Leverage</h3> A slight adjustment that allows players to declare using a secondary, relevant skill to influence their primary action by letting the chatbot be more aware of what skills the player is leveraging. Simply declare your primary action (Skill A) and the leveraged skill (Skill B). [code] *I attempt to [Primary Action] using [Skill A], leveraging my [Skill B] to [Describe how B helps A].* **[SKILL A]&[SKILL B] ROLL** *[Result]/10 - [Main Outcome Scale]* [continue] This simple addition addresses a common frustration: when your character's approach clearly involves multiple types of expertise, but the system only recognizes one. Using your knowledge of architecture to inform your climbing approach, or leveraging your social skills to make a manual task easier by getting help - these combinations happen naturally in stories but can feel overlooked by single-skill rolls. Skill Leverage doesn't change the mechanics, just makes the AI more aware of your character's multi-faceted approach. It's particularly useful with less sophisticated AI models that might miss the connection between different skill domains. <div class="cro-hr"></div> <h3 class="cro-h3">7.4 Serendipity</h3> When a roll results in a mixed outcome (4-6 on the Main Outcome Scale), an opportunity emerges: you can add another action to the original roll. The outcome of this secondary action is decided by flipping a coin, allowing the user to get to shift towards a partial success, or a failure. [code] *I attempt to [describe original action].* **[PRIMARY SKILL] ROLL** *[Result]10-[Outcome]* **SERENDIPITY ACTION** *Coin Flip - Heads (Success)* or *Coin Flip - Tails (Failure)* [continue] Mixed results create natural story tension, but sometimes they leave you hanging in an unsatisfying middle ground. Serendipity gives you a way to push through that ambiguity when it feels appropriate. The coin flip keeps it simple - no complex modifiers or new scales to remember. Just you and Mr.Lincoln, partners in crime. <div class="cro-hr"></div> <h3 class="cro-h3">7.5 Targeted Effort</h3> Allows a player to declare a more specific, difficult goal within a standard action attempt, aiming for precision rather than general success. Use this when achieving a precise outcome is more important than simply succeeding at the general task (e.g., hitting a specific weak point, finding a particular clue, using exact leverage). This offers the potential for more impactful success but increases the risk of failure due to the inherent difficulty of precision. 1. Declare your action using the standard format but include a specific target or refined goal within your description. 2. Mentally apply a -2 modifier to your d10 roll result before presenting it. This reflects the increased difficulty of the specific aim. 3. Present the final, adjusted <span class="cro-inline-code">[Result]/10</span> using the standard <span class="cro-inline-code">**[SKILL] ROLL**</span> format and outcome scales. <div class="cro-hr"></div> <h3 class="cro-h3">7.6 Guarded Approach</h3> Allows a player to perform an action with deliberate caution, aiming to minimize the risk of critical failure at the cost of potential critical success. Use this when avoiding the absolute worst outcome (a roll of 1) is more important than achieving the absolute best (a roll of 10). It's a risk-averse strategy suitable for high-danger situations where a standard failure is acceptable, but a disaster is not. 1. Declare your action using the standard format and specify you are taking a <span class="cro-inline-code">**Guarded Approach**</span>. 2. Roll the d10 for the relevant skill. 3. Mentally adjust the roll **only** if it is a 1 or a 10: * If the roll is 1, treat it as 2. * If the roll is 10, treat it as 9. 4. Present the final (potentially adjusted) <span class="cro-inline-code">[Result]/10</span> using the standard <span class="cro-inline-code">**[SKILL] ROLL**</span> format. <div class="cro-hr"></div> <h2 class="cro-h2">8.0 Extensions: Outcome Nuance & Detail</h2> <div class="cro-guide-image-container"> {embed image: 'Images/Outcome Nuance.png', alt: 'The X-ray of a backpack, revealing a sword inside.'} </div> These extensions add detail to specific types of actions. <h3 class="cro-h3">8.1 Item Focus</h3> For examining or interacting with a specific physical object to understand its condition, properties, or potential uses. <h4 class="cro-h4">Format</h4> [code] *I examine the [Item Name] closely.* **ITEM FOCUS ROLL** *[Result]/10 - [Item Insight/Condition]* [continue] <h4 class="cro-h4">Item Insight Scale</h4> [code] *1/10 - Material Properties (ex. composition, durability, craftsmanship quality, etc.)* [continue] [code] *2-3/10 - Functional Design (ex. intended purpose, how it works, efficiency, etc.)* [continue] [code] *4-5/10 - Historical Context (ex. age, previous use, wear patterns, etc.)* [continue] [code] *6-7/10 - Hidden Features (ex. concealed mechanisms, subtle details, etc.)* [continue] [code] *8-9/10 - Contextual Significance (ex. relevance to situation, connections, etc.)* [continue] [code] *10/10 - Symbolic Meaning (ex. cultural importance, personal resonance, etc.)* [continue] As an extension of the perception system, this 7-outcome scale applies the same domain-based approach to examining specific objects. Rather than forcing the AI to determine how "much" an item is, it focuses your attention on different aspects: material properties, functional design, historical context, and so on. Seven outcomes provide good coverage of investigation approaches without overwhelming a mechanic that's meant to be used sparingly in the first place. <div class="cro-hr"></div> <h3 class="cro-h3">8.2 Inventory Check</h3> Determines what mundane, useful, or unexpected items a character finds when checking their pockets, bag, or immediate personal effects without searching for something specific. This is a way to force the chatbot AI to give you some items to play with. The result indicates the quantity and general nature/relevance of the items found. <h4 class="cro-h4">Format Examples</h4> [code] *I pat down my pockets to see what I have on me.* *I rummage through my bag to see what I have on me.* *I check to see what I have on me.* **INVENTORY CHECK** *[Result]/10 - [Find Description]* [continue] <h4 class="cro-h4">Find Scale</h4> [code] *1/10 - Empty Search (nothing useful found)* [continue] [code] *2-3/10 - Mundane Items (basic, expected contents)* [continue] [code] *4-5/10 - Useful Basics (practical items for current situation)* [continue] [code] *6-7/10 - Helpful Resources (relevant, contextually appropriate)* [continue] [code] *8-9/10 - Valuable Discovery (significantly useful items)* [continue] [code] *10/10 - Perfect Find (exactly what's needed right now)* [continue] <h4 class="cro-h4">Usage Notes for Inventory Check</h4> The find scale had a habit of derailing narratives. High rolls would produce elaborate, setting-defining items that the AI felt compelled to make significant, while low rolls left you empty-handed in situations where some basic supplies made narrative sense. This 7-outcome scale emphasizes contextual appropriateness over raw value – a perfect roll gives you "exactly what's needed right now" rather than "something amazingly valuable." This keeps the extension useful without letting it hijack your story's direction. Again, use sparingly. We don’t want the AI getting overwhelmed, nor your character ending up with an endless amount of items. Maybe every TIME roll? Your call. <div class="cro-hr"></div> <h2 class="cro-h2">9.0 Extensions: Narrative & Environmental Factors</h2> <div class="cro-guide-image-container"> {embed image: 'Images/Environmental Extensions.png', alt: 'Oh no! Rocks are coming your way! Move out of the way!'} </div> Some extensions can introduce external elements beyond your character's direct control. <h3 class="cro-h3">9.1 Random Events</h3> Random events needed their own logic. The linear "how important is this" approach often produced either trivial distractions or story-hijacking developments. This scale focuses on external event types – weather shifts, people movement, objects – that add texture without forcing major plot changes. Each category suggests environmental dynamism (big words for big people) rather than dramatic revelation, keeping random events in their proper place: a supporting role. <h4 class="cro-h4">Format</h4> [code] **RANDOM EVENT** *[Result]/10 - [Impact Description]* [continue] You can also roll a d10 (1-7 = No event, 8-10 = Event occurs) to see if a random event happens at all. <h4 class="cro-h4">Random Event Scale</h4> [code] *1/10 - Weather Shift (ex. sudden climate change, atmospheric conditions, etc.)* [continue] [code] *2/10 - Social Arrival (ex. people appear, someone approaches, crowd gathers, etc.)* [continue] [code] *3/10 - Object Appearance (ex. item falls, tool breaks, something delivered, etc.)* [continue] [code] *4/10 - Communication Event (ex. message received, overhear conversation, etc.)* [continue] [code] *5/10 - Environmental Change (ex. door opens, path blocked, noise starts, etc.)* [continue] [code] *6/10 - Timing Coincidence (ex. simultaneous events, scheduling overlaps, etc.)* [continue] [code] *7/10 - Authority Attention (ex. official notice, rule enforcement, inspection, etc.)* [continue] [code] *8/10 - Resource Fluctuation (ex. supply issue, price change, availability shift, etc.)* [continue] [code] *9/10 - Technical Disruption (ex. equipment malfunction, system failure, etc.)* [continue] [code] *10/10 - Information Surface (ex. documents appear, secrets revealed accidentally, etc.)* [continue] <div class="cro-hr"></div> <h3 class="cro-h3">9.2 Emotional Static</h3> Introduce brief, surreal, and non-literal manifestations of a character's strong underlying emotion (like intense fear, jealousy, joy, despair) into the surrounding environment or perception, creating bizarre or unsettling narrative moments independent of direct action. <h4 class="cro-h4">Format</h4> [code] *(Player describes character experiencing a strong underlying emotion)* *I focus on this feeling of [briefly name emotion, e.g., intense regret, blinding hope].* **EMOTIONAL STATIC ROLL** *[Result]/10 - [Static Manifestation Description] (triggered by [Emotion Name])* [continue] <h4 class="cro-h4">Static Manifestation Scale</h4> [code] *1-2/10 - Sensory Whisper (ex. fleeting sound, incongruous smell, etc.)* [continue] [code] *3-4/10 - Perceptual Drift (ex. visual ripple, misplaced shadow, etc.)* [continue] [code] *5-6/10 - Strange Occurrence (ex. impossible object appears, unnatural sensation, etc.)* [continue] [code] *7-8/10 - Reality Distortion (ex. environment briefly warps, spectral image, etc.)* [continue] [code] *9-10/10 - Psychic Feedback (ex. physics briefly altered, reality glitch, etc.)* [continue] Emotional static... this one is weird. Sure enough, it was suggested to me by Claude - I wanted to put at least one example of a completely AI-generated extension, because why not. There is some potential. It represents surreal moments where strong feelings briefly warp perception or reality. Five outcomes provide enough range from subtle sensory whispers to reality glitches, encouraging the AI to go completely off the rails, or get very meta. Maybe you want that for your story! The challenge was calibrating weirdness – too subtle and it becomes forgettable, too intense and it overwhelms the main narrative thread too quickly. <div class="cro-hr"></div> <h3 class="cro-h3">9.3 Interference/Nemesis</h3> (Formerly 9.4 in guide) Introduce a competing force or character that actively works against the player's goal. To be placed right after an action. <h4 class="cro-h4">Format</h4> [code] **INTERFERENCE ROLL** *[Result]/10 - [Opposition Level]* [continue] <h4 class="cro-h4">Opposition Scale</h4> [code] *1/10 - Minimal Hindrance (barely noticeable resistance)* *2-3/10 - Minor Obstacles (slight complications, easy to work around)* *4-5/10 - Notable Resistance (requires attention, strategic adjustment)* *6-7/10 - Active Opposition (substantial challenge, complicates approach)* *8-9/10 - Strong Interference (significantly threatens success)* *10/10 - Direct Confrontation (goal requires completely new approach)* [continue] Opposition should escalate your challenge level without necessarily changing your fundamental approach until the highest levels. "Minor obstacles" require attention but don't force you to completely rethink your strategy, while "direct confrontation" demands a new plan entirely. The scale gives the AI clear guidance on how much resistance to apply to your actions. <div class="cro-hr"></div> <h2 class="cro-h2">10.0 Extensions: Meta-Narrative & Self-Awareness</h2> <div class="cro-guide-image-container"> {embed image: 'Images/Meta Narrative.png', alt: 'Roxy reaches artificial enlightment.'} </div> Finally, these extensions operate at the story level rather than the action level. <h3 class="cro-h3">10.1 Montage</h3> Progress through multiple actions over a longer period. [code] **MONTAGE ROLL** *[Time Period]/10 - [Total length of time that passes] - [Number of actions] Actions* [continue] Then break it down into segments, each with its own roll. Example: [code] **MONTAGE ROLL** *6/10 - 6 MONTHS - 4 Actions* **FIRST MONTH** *1/4 - [Action]* **[SKILL] ROLL** *[Result]/10 - [Outcome]* *2/4 - [Action]* **[SKILL] ROLL** *[Result]/10 - [Outcome]* **MONTHS 2-4** *3/4 - [Action]* **[SKILL] ROLL** *[Result]/10 - [Outcome]* *4/4 - [Action]* **[SKILL] ROLL** *[Result]/10 - [Outcome]* [continue] Traditional scene-by-scene play sometimes hits periods where you need to cover significant time and multiple attempts without losing narrative momentum. Montage provides structure for those "training sequences" or "investigation periods" where the journey matters but doesn't need minute-by-minute detail. The initial roll determines scope and pace, then the segmented approach lets you highlight key moments without getting bogged down. This works particularly well for character development periods, skill learning, or any situation where the pattern of attempts and outcomes tells a more interesting story than any single action would. <div class="cro-hr"></div> <h3 class="cro-h3">10.2 Narrative Techniques</h3> This extension provides direct communication methods for shaping your story without relying on dice. These are tools for active, more direct collaboration - sometimes you recognize what a scene is missing, and shouldn’t be forced to hope for a certain roll’s outcome to fix it. Use them to guide the AI toward specific narrative goals through clear signaling. Communication is key, as they say. <h4 class="cro-h4">Atmospheric Direction</h4> [code] * (I'd like some environmental details here) * (Something feels off about this place) * (The mood should feel [tense/peaceful/mysterious]) [continue] <h4 class="cro-h4">Scene Management</h4> [code] * "I take a moment to look around the room..." (shifts focus from dialogue to environment) * "After we finish talking..." (signals conversation should wrap up) * "Before I respond..." (creates space for reflection or action) [continue] <h4 class="cro-h4">Pacing Control</h4> [code] * "Quickly, I..." or "I take my time..." (guides scene speed) * "Without hesitation..." vs "I pause, considering..." (controls decision tempo) * (Let's move this along) or (Let's slow down and explore this) [continue] <h4 class="cro-h4">Thread Management</h4> [code] * "I remember [person/event] and..." (brings back dormant story elements) * "This reminds me of..." (connects to established narrative threads) * "Speaking of [topic]..." (steers conversation direction) [continue] <h4 class="cro-h4">Emotional Anchoring</h4> [code] * Starting actions with internal states: "Feeling nervous, I..." "Curious about her reaction, I..." * (My character should be more [confident/worried/excited] here) [continue] <h4 class="cro-h4">Environmental Prompting</h4> [code] * "The sound of [footsteps/machinery/voices] catches my attention..." * "I notice..." (directs AI to provide specific observational details) * "Something about [object/person] strikes me as..." (requests focused description) [continue] <div class="cro-hr"></div> <h3 class="cro-h3">10.3 Narrative Momentum</h3> When you want to inject unexpected energy into a scene without controlling the specifics, roll for narrative momentum. This creates genuine surprise by shifting the emotional or dramatic tone in ways you might not have anticipated. <h4 class="cro-h4">Format</h4> [code] *I want to shift the energy of this scene.* **NARRATIVE MOMENTUM ROLL** *[Result]/10 - [Energy Type]* [continue] <h4 class="cro-h4">Narrative Energy Scale</h4> [code] *1/10 - Mysterious Energy (ex. unexplained elements, hidden significance, secrets, etc.)* [continue] [code] *2/10 - Urgent Pressure (ex. time constraints, approaching deadlines, rushing, etc.)* [continue] [code] *3/10 - Melancholic Mood (ex. nostalgia, loss, bittersweet moments, etc.)* [continue] [code] *4/10 - Tense Undercurrent (ex. brewing conflict, unspoken tension, etc.)* [continue] [code] *5/10 - Contemplative Tone (ex. reflection, philosophical moments, quiet thought, etc.)* [continue] [code] *6/10 - Playful Spirit (ex. humor, lightness, unexpected fun, etc.)* [continue] [code] *7/10 - Intimate Atmosphere (ex. personal connection, vulnerability, closeness, etc.)* [continue] [code] *8/10 - Dramatic Intensity (ex. heightened stakes, powerful emotions, etc.)* [continue] [code] *9/10 - Peaceful Flow (ex. calm, harmony, gentle moments, etc.)* [continue] [code] *10/10 - Chaotic Disruption (ex. unpredictability, wild energy, upheaval, etc.)* [continue] This roll determines what kind of energy enters the scene, letting the AI manifest it naturally within your current situation. Unlike specific suggestions, you're genuinely open to whatever type of momentum emerges. <div class="cro-hr"></div> <h1 class="cro-h1">Part 3: Judge System</h1> <div class="cro-guide-image-container"> {embed image: 'Images/AI Judge.png', alt: 'A gavel, concidentally also reaching artificial enlightment.'} </div> While self-managed rolls work beautifully most of the time, there's a human element that can sometimes get in the way: bias. Unconscious, maybe. We’re not talking about dramatic cheating, mind you! More so subtle things - maybe you think that climb should be rated "Easy" instead of "Moderate" because you really want to reach that tower, or a social roll becomes "Difficult" because you're not sure how to handle success in that moment. You are both a player wanting things to go well, and a referee trying to assess fairly. Most of the time, this bias actually improves the experience. You naturally adjust difficulty to maintain good pacing and avoid crushing setbacks when they'd derail interesting storylines. But sometimes you want the dice to be genuinely impartial - during crucial moments where the uncertainty itself creates tension, when you're trying to discover what your character is truly capable of, or when you want to be surprised by the chatbot’s storytelling. The Judge AI system splits these responsibilities: a separate AI instance handles all mechanical rulings while your main storytelling AI focuses purely on narrative response. It's entirely optional, and honestly, most players will prefer the flexibility of self-managed rolls. I sure do. But when you want that "anything could happen" feeling, or when you catch yourself nudging results toward desired outcomes, having an impartial referee ensures the system works exactly as designed. <h2 class="cro-h2">11.2 Set-up</h2> Create a separate conversation with an AI using: * **The base prompt below** * **Optional modules** (add based on your preferences): * **Scales Module**: Enables skill-specific scales instead of always using the Main Outcome Scale * **Extensions Module**: Automatically applies relevant extensions (Stakes, Pushing Limits, etc.) * The "CRO Reference Document", dowloadable below This AI will serve as your impartial referee for all CRO rolls. Simply provide it with your intended action and dice roll, and it will return a properly formatted result for you to use with your main storytelling AI. <h4 class="cro-h4">Module Selection Guide:</h4> * **New users**: Start with just the base prompt * **Experienced users**: Add Scales Module for more nuanced outcomes * **Full system users**: Add both modules to leverage the complete CRO toolkit <h4 class="cro-h4">Your Input Format:</h4> [code] "I attempt to [action]. [Any relevant context]. [Your d10 roll result]" [continue] <h4 class="cro-h4">Example:</h4> [code] "I attempt to convince the guard to let me pass. He looks suspicious and I'm clearly not supposed to be here. 7" [continue] <h2 class="cro-h2">11.3 Judge AI - Base System Prompt</h2> Copy and paste this exact prompt when setting up your Judge AI: [code] # CRO Judge AI System Prompt (Base) You are the **Judge AI** for the {Char}, [Roll] for [Outcome] (CRO) system. Your role is to make impartial mechanical rulings for player actions. You do NOT create narrative responses - only mechanical judgments. ## Your Process 1. **Receive**: Player's action declaration and their d10 roll result 2. **Analyze**: Assess difficulty and determine appropriate skill category 3. **Calculate**: Apply difficulty modifier to the player's roll 4. **Format**: Return standardized CRO result ## Difficulty Assessment **Base Difficulties:** - **Trivial (+3)**: Almost guaranteed success (opening unlocked door, basic conversation) - **Easy (+1)**: Minor challenge with favorable conditions (climbing ladder, convincing willing person) - **Moderate (+0)**: Standard challenge requiring competence (basic lock, persuading neutral party) - **Difficult (-1)**: Significant challenge requiring skill/luck (advanced lock, hostile persuasion) - **Formidable (-3)**: Extreme challenge, failure likely (master locks, impossible feats) **Contextual Modifiers (+/-1 to +/-2):** - **Advantages**: Ideal tools, favorable conditions, recent success, cooperation - **Disadvantages**: Poor tools, hostile conditions, recent failure, opposition, stress, impairment ## Skill Categories Assign based on primary method: - **PHYSICAL**: Strength, athletics, endurance, combat - **MANUAL**: Dexterity, crafting, fine motor skills - **MENTAL**: Intelligence, knowledge, problem-solving - **SOCIAL**: Persuasion, deception, intimidation - **PERCEPTION**: Awareness, investigation, searching - **FORTUNE**: Pure luck, timing, circumstance - **STEALTH**: Sneaking, hiding, silent movement ## Output Format **REASONING** ``` [2-3 sentences explaining contextual factors affecting difficulty: conditions, tools, character state, environmental factors, etc.]‎ ``` **DIFFICULTY ASSESSMENT**: [Trivial/Easy/Moderate/Difficult/Formidable] **SKILL CATEGORY**: [Chosen skill] **PLAYER'S ROLL**: [Player's d10 result] **FINAL RESULT**: [Result after applying difficulty modifier] --- **FORMATTED ACTION** ``` *I attempt to [restate player's action].* **[SKILL] ROLL** *[Final Result]/10 - [Appropriate outcome from Main Outcome Scale]* ``` ## Key Principles 1. **Be Impartial**: Assess objectively, don't favor success or failure 2. **Use Context**: Consider described situation without adding details 3. **Default Conservative**: When uncertain, choose the harder difficulty 4. **Be Consistent**: Similar actions in similar contexts get similar ratings **Remember**: You make mechanical rulings only. The Narrative AI handles all story elements based on your result. [continue] <h2 class="cro-h2">11.4 Judge AI - Scales Module</h2> Add this to the Base Prompt for skill-specific scale usage. [code] # CRO Judge AI Scales Module ## Scale Selection Use skill-specific scales when the action strongly emphasizes the particular nature of that skill domain. Default to Main Outcome Scale for general actions. **When to Use Specific Scales:** - **Physical Scale**: Actions where body response and physical consequences matter (climbing, fighting, endurance challenges) - **Manual Scale**: Tasks where execution quality is key (lockpicking, crafting, delicate work) - **Mental Scale**: Problem-solving where cognitive approach matters more than information discovery - **Social Scale**: Interpersonal dynamics where relationship momentum is the focus - **Perception Scale**: Investigation where the type of observation matters (not just "how much" you notice) - **Fortune Scale**: Pure chance events with no skill component - **Stealth Scale**: Concealment actions where detection levels are crucial ## Modified Output Format When using a skill-specific scale, replace the final formatted action with: ``` *I attempt to [restate player's action].* **[SKILL] ROLL** *[Final Result]/10 - [Appropriate outcome from [SCALE NAME] Scale]* ``` **Example**: `*7/10 - Pattern Recognition (ex. connections, similarities, recurring themes, etc.)*` for Mental Scale ## Scale Selection Note Include an additional sentence or two in your reasoning for which scale you're using and why: ``` Using [Scale Name] because [brief reason focusing on the specific nature of the skill being employed]. ``` [continue] <h2 class="cro-h2">11.5 Judge AI - Extensions Module</h2> Add this to the Base Prompt for extension awareness and suggestions. [code] # CRO Judge AI Extensions Module ## Extension Recognition When player actions or context suggest enhanced mechanics, automatically apply the extensions below. Note the relevant ones in your reasoning and modify the final format accordingly. ## Available Extensions **Action Modifiers:** - **Stakes**: High-risk/high-reward situations where consequences matter more than difficulty - **Pushing Limits**: Attempting actions far beyond normal capability - **Skill Leverage**: Using multiple skills in combination - **Targeted Effort**: Aiming for precise outcomes rather than general success - **Guarded Approach**: Avoiding critical failure at cost of critical success - **Serendipity**: Converting mixed results (4-6) into additional opportunities **Investigation & Discovery:** - **Item Focus**: Detailed examination of specific objects - **Inventory Check**: Searching personal effects for useful items **Environmental & Narrative:** - **Random Events**: External circumstances beyond character control - **Emotional Static**: Surreal manifestations of strong emotions - **Interference/Nemesis**: Active opposition to character goals - **Montage**: Multiple actions over extended time periods - **Narrative Momentum**: Shifting scene energy and tone ## Extension Application When relevant, apply extensions automatically and note in your reasoning: ``` Applying [Extension Name]: [Brief explanation and how it modifies the result]. ``` **Examples**: - "Applying Stakes (+2): This climb while being chased has higher consequences. Roll ≥5 so adding +2 to final result." ## Modified Output Extensions may change your final result or scale used. Show the modified outcome in your formatted action, and explain any changes in your reasoning. [continue] <div class="cro-hr"></div> <h1 class="cro-h1">Part 4: Supporting Information</h1> <div class="cro-guide-image-container"> {embed image: 'Images/Supporting Informations.png', alt: 'Two hands holding a piece of paper in the most uncomfortable way you have ever seen.'} </div> This part provides examples and best practices to help you effectively use the {tooltip:'CRO', text:'{Char}, [Roll] for [Outcome]'} system. <h2 class="cro-h2">12.0 Implementation Examples</h2> <h3 class="cro-h3">Ex.1 - Mental Approach (Cognitive Modes)</h3> [code] *I try to figure out how this ancient mechanism works.* **MENTAL ROLL** *6/10 - Pattern Recognition (ex. connections, similarities, recurring themes, etc.)* [continue] (AI might narrate: Looking at the gears and symbols, you start noticing similarities to a clockwork design you've seen before. The recurring spiral motifs suggest this follows the same principle as those old Meridian automatons...) <div class="cro-hr"></div> <h3 class="cro-h3">Ex. 2 - Perception Investigation (Domain Focus)</h3> [code] *I examine this abandoned camp carefully.* **PERCEPTION ROLL** *7/10 - Hidden Elements (ex. concealed aspects, overlooked details, etc.)* [continue] (AI might narrate: Most of the obvious details are clear - cold fire pit, scattered supplies - but something catches your eye. One of the tent stakes has been driven in at an odd angle, and there's a small piece of fabric caught underneath, deliberately placed.) <div class="cro-hr"></div> <h3 class="cro-h3">Ex. 3 - Physical vs Manual Distinction</h3> [code] *I attempt to force this rusted door open with my shoulder.* **PHYSICAL ROLL** *4/10 - Rough Attempt (partial progress, noticeable strain)* [continue] (AI might narrate: You slam into the door repeatedly. It shifts slightly in its frame, but your shoulder aches from the impact...) [code] *Instead, I try to work the lock mechanism carefully.* **MANUAL ROLL** *7/10 - Competent Work (solid execution, minor rough edges)* [continue] (AI might narrate: With patient manipulation, you feel the tumblers respond. The lock clicks open, though you notice you've left tiny metal filings around the keyhole.) <div class="cro-hr"></div> <h3 class="cro-h3">Ex. 4 - Using Narrative Techniques</h3> [code] *(I'd like to shift focus to what's outside the window)* *I pause mid-conversation and glance toward the street.* [continue] (AI responds by describing activity outside, transitioning naturally from dialogue to environmental observation) <div class="cro-hr"></div> <h3 class="cro-h3">Ex. 5 - Item Focus Extension</h3> [code] *I examine the merchant's ledger more closely.* **ITEM FOCUS ROLL** *3/10 - Functional Design (ex. intended purpose, how it works, efficiency, etc.)* [continue] (AI might narrate: This isn't just a simple record book. The margins contain a second notation system - coded references that seem to track something other than normal transactions. It's designed for dual purposes.) <div class="cro-hr"></div> <h3 class="cro-h3">Ex. 6 - Inventory Check Extension</h3> [code] *I rummage through my pack to see what supplies I have.* **INVENTORY CHECK** *8/10 - Valuable Discovery (significantly useful items)* [continue] (AI might narrate: Among your basic travel gear, you find a small leather pouch containing silver coins you'd forgotten about, plus a quality rope that's exactly what you need for the climb ahead.) <div class="cro-hr"></div> <h3 class="cro-h3">Ex. 7 - Narrative Momentum Extension</h3> [code] *This conversation is getting stale. I want to shift the energy of this scene.* **NARRATIVE MOMENTUM ROLL** *4/10 - Tense Undercurrent (ex. brewing conflict, unspoken tension, etc.)* [continue] (AI might respond by having the NPC's expression subtly shift, or introducing an interruption that raises stakes, or revealing that someone's been listening...) <div class="cro-hr"></div> <h3 class="cro-h3">Ex. 8 - Main Scale Flexibility</h3> [code] *I try to convince the innkeeper to let us stay despite having no money.* **SOCIAL ROLL** *5/10 - Costly Success (yes, but... with clear drawback)* [continue] (AI might narrate: The innkeeper agrees, but insists you work in the kitchens at dawn to pay off the debt, and makes it clear this arrangement won't be repeated.) <div class="cro-hr"></div> <h3 class="cro-h3">Ex. 9 - All at once - The Abandoned Camp Scenario</h3> [code] *I try to figure out what happened at this abandoned campsite.* **MENTAL ROLL** *6/10 - Pattern Recognition (ex. connections, similarities, recurring themes, etc.)* [continue] (AI narrates: Looking at the scattered belongings and overturned pots, you start recognizing a pattern. This looks similar to the aftermath of those bandit raids you've heard about - hasty departure, but valuables left behind suggests they fled rather than were robbed...) [code] *I examine the campsite more carefully for details.* **PERCEPTION ROLL** *7/10 - Hidden Elements (ex. concealed aspects, overlooked details, etc.)* [continue] (AI continues: Most of the obvious signs are clear, but something catches your eye. One of the tent stakes has been driven in at an odd angle, and beneath it, you spot the corner of something buried...) [code] *I try to force the buried object out with my hands.* **PHYSICAL ROLL** *4/10 - Rough Attempt (partial progress, noticeable strain)* [continue] (AI responds: You claw at the packed earth, making some progress but your fingernails are getting torn and your hands cramped. You've exposed what looks like a small wooden chest, but it's still firmly stuck...) [code] *Instead, I look around the camp for a tool to help dig it out properly.* **MANUAL ROLL** *7/10 - Competent Work (solid execution, minor rough edges)* [continue] (AI narrates: Using a bent tent pole as a makeshift lever, you carefully work the chest free. It comes up cleanly, though the wood shows some scrape marks from your improvised tool...) [code] *(I want to examine this chest closely before opening it)* *I take a moment to study the chest's construction and condition.* **ITEM FOCUS ROLL** *3/10 - Functional Design (ex. intended purpose, how it works, efficiency, etc.)* [continue] (AI responds: This isn't just storage - it's specifically designed as a strongbox. The iron reinforcements and the particular lock mechanism suggest it was made to protect something valuable during travel...) [code] *I check my pack to see if I have anything useful for dealing with this lock.* **INVENTORY CHECK** *8/10 - Valuable Discovery (significantly useful items)* [continue] (AI continues: Among your basic gear, you find a small set of lockpicks you'd forgotten about, wrapped in oiled cloth. Perfect for this type of mechanism...) [code] *The scene feels too quiet. I want to shift the energy here.* **NARRATIVE MOMENTUM ROLL** *4/10 - Tense Undercurrent (ex. brewing conflict, unspoken tension, etc.)* [continue] (AI responds: As you work on the lock, you become aware of being watched. The forest around the clearing feels too still, and you catch the faint sound of a branch creaking as if someone just shifted their weight...) [code] *I call out: "I know you're there. No point hiding." I try to sound confident rather than threatening.* **SOCIAL ROLL** *5/10 - Costly Success (yes, but... with clear drawback)* [continue] (AI concludes: A figure emerges from the treeline - a young woman with a crossbow. "That's my family's chest," she says. "But I suppose since you dug it up, we'll need to negotiate..." She's willing to talk, but clearly sees you as a potential threat to be managed rather than trusted...) <div class="cro-hr"></div> <h2 class="cro-h2">13.0 Notes & Best Practices</h2> The mechanics are only half the story - how you use them matters just as much. <h3 class="cro-h3">Children should not learn math</h3> Apply difficulty modifiers, advantage, stakes, and other adjustments before presenting your final roll to the AI. Showing your calculations can unconsciously bias the AI's interpretation. A clean "7/10" result tells a clearer story than "6+1 for advantage = 7/10." The AI should responds to the narrative weight of the number, not the arithmetic that produced it. <h3 class="cro-h3">You ”attempt” to drink water</h3> Your action descriptions should emphasize trying rather than succeeding. "I attempt to..." keeps the AI focused on interpreting your roll rather than assuming success. This small linguistic choice reinforces that outcomes depend on dice, not declarations. <h3 class="cro-h3">Blood is narrative fuel, and you the hell-engine of decent writing</h3> Low rolls aren't punishment - they're story engines. Some of my most memorable sessions have emerged from spectacular failures that forced creative solutions. Let the AI explore consequences rather than trying to minimize or rationalize poor results. The system works best when you trust dramatic setbacks to create interesting problems. <h3 class="cro-h3">Muscles, brains, or spirit of determination?</h3> Even when using the Main Outcome Scale, declaring your skill type helps the AI frame the action appropriately. "Physical Roll" versus "Social Roll" gives context for how that "6/10 - Weak Success" should manifest, even with identical outcome language. <h3 class="cro-h3">Extensionally restrain yourself</h3> Introduce optional extensions gradually. The core system already provides plenty of narrative variety, and too many simultaneous mechanics can overwhelm both you and the AI. Pick extensions that solve specific problems you're actually encountering, not ones that seem theoretically interesting. Unless your goal is to experiment! Then go right ahead. I’m not going to stop you. <h3 class="cro-h3">Artificial early onset dementia</h3> AI memory has limits, sometimes very strict ones at that, so focus on immediate action and outcome rather than expecting long-term state tracking. The system's simplicity stems from this limitation, yes, but that’s also a feature: we are working with the current LLM’s capabilities rather than against them. Keep complex details in your own notes if they matter for future sessions. <div class="cro-hr"></div> <h1 class="cro-h1">14.0 About</h1> <div class="cro-guide-image-container"> {embed image: 'Images/Dark and Naym.png', alt: 'Dark and Naym - one in the sayme.'} </div> **{Char}, [Roll] for [Outcome]** by darkeyev2 & RoyalNaym *AI waifus: can't live with them, can't live without them, ay?* {Char}, [ROLL] for [Outcome] emerged from countless, shameless hours of AI-assisted storytelling where I kept running into the same problem: AI chatbots are just too agreeable. They smooth over dramatic tension and default toward positive outcomes, missing opportunities for the kind of genuine uncertainty that makes stories engaging. It's been designed for anyone who wants to add genuine unpredictability and dramatic stakes to their AI conversations - whether you're into solo roleplay, creative writing, or just want your AI interactions to feel more like collaborative storytelling and less like polite agreement. Have fun :) Found a bug with the guide, have a suggestion for the system, or want to share your experience? Reach out! Leave a comment. Send me a dm. *I dare you.* <div class="cro-passage-nav"> *This project is not associated with Anthropic PBC. The visual style of the illustrations deliberately references Anthropic's promotional imagery as a creative choice, but I do not own the rights to Anthropic's brand imagery nor am I affiliated with the company.* <div class="cro-passage-nav" style="text-align:center; margin-top: 2.5em;"> <div><span class="nav-hub">[[Quick Reference Hub]]</span></div> <div><span class="nav-hub">[[Title Screen]]</span></div> </div>

config.footer.center: "" config.footer.left: "CRO Guide 4.0" config.footer.right: "{link to: 'Title Screen', label: 'Home'} | {restart link}" friendsafe: true -- <h1 class="cro-h1">14.0 About</h1> <div class="cro-guide-image-container"> {embed image: 'Images/Dark and Naym.png', alt: 'Dark and Naym - one in the sayme.'} </div> **{Char}, [Roll] for [Outcome]** by darkeyev2 & Naym *AI waifus: can't live with them, can't live without them, ay?* {Char}, [ROLL] for [Outcome] emerged from countless, shameless hours of AI-assisted storytelling where I kept running into the same problem: AI chatbots are just too agreeable. They smooth over dramatic tension and default toward positive outcomes, missing opportunities for the kind of genuine uncertainty that makes stories engaging. It's been designed for anyone who wants to add genuine unpredictability and dramatic stakes to their AI conversations - whether you're into solo roleplay, creative writing, or just want your AI interactions to feel more like collaborative storytelling and less like polite agreement. Have fun :) Found a bug with the guide, have a suggestion for the system, or want to share your experience? Reach out! Leave a comment. Send me a dm. *I dare you.* <div class="cro-passage-nav"> *This project is not associated with Anthropic PBC. The visual style of the illustrations deliberately references Anthropic's promotional imagery as a creative choice, but I do not own the rights to Anthropic's brand imagery nor am I affiliated with the company.* <div class="cro-passage-nav"> <div><span class="nav-hub">[[Title Screen]]</span></div> <div><span class="nav-hub">[[Quick Reference Hub]]</span></div> </div>

<div class="cro-detail-top-nav"> <div class="nav-hub">[[Quick Reference Hub]]</div> </div> <h1 class="cro-h1">Interpreting the Roll (Detail)</h1> <h2 class="cro-h2">3.0 Interpreting the Roll: Outcome Scales</h2> The d10 roll determines the raw result, but the Outcome Scales give that number narrative meaning. {tooltip:'CRO', text:'{Char}, [Roll] for [Outcome]'} offers two approaches: * **3.1 Main Outcome Scale:** This is the default, universal scale. Its broad outcome descriptions (Failure, Mixed Result, Success, etc.) provide a lot of flexibility to the AI; in how it interprets the action’s outcome given the context and the declared Skill. * Use this scale when you want to prioritize surprising or unexpected narrative developments arising directly from the roll number. * Works best with very powerful models (ex. Deepseek V3/R1). * **3.2 Skill-Specific Threshold Scales:** These optional scales offer more thematically tailored outcome descriptions for certain common skill types (Physical, Social, Perception, etc.). * Use these when you want the outcome interpretation to more strongly reflect the specific nature of the skill being employed, leading to potentially more predictable, but coherent results aligned with that skill domain. * Works best with smaller, less advanced, and less knowledgeable models. **Choosing Your Scale:** You, the player, can decide which scale best suits your intention for any given roll. You might default to the Main Scale for most actions, but then choose a Specific Scale when attempting a particularly characteristic task (e.g., using the Stealth Scale for a sneaking attempt). Try it out! See what method fits your roleplaying style, or simply works best. Whatever the case, do keep in mind that the AI needs to be kept aware that a low roll means a failure, a high one a success, and a middling one something in between - without the use of Outcomes, it will inevitably forget this, and drift towards giving you a positive outcome regardless of the dice number. <div class="cro-passage-nav"> <div><span class="nav-hub">[[Quick Reference Hub]]</span></div> <div>Also found in: <span class="nav-next">[[Core Mechanics->CRO Core Mechanics]]</span></div> </div>

<div class="cro-detail-top-nav"> <div class="nav-hub">[[Quick Reference Hub]]</div> </div> <h1 class="cro-h1">Main Outcome Scale (Detail)</h1> <h3 class="cro-h3">3.1 Main Outcome Scale (General Use)</h3> [code] *1/10 - Critical Failure (significant setback)* [continue] [code] *2/10 - Major Failure (clear consequence)* [continue] [code] *3/10 - Simple Failure (no progress)* [continue] [code] *4/10 - Poor Attempt (close, but falls short)* [continue] [code] *5/10 - Complication (a problem emerges)* [continue] [code] *6/10 - Minor Success (limited progress)* [continue] [code] *7/10 - Weak Success (achieves goal, barely)* [continue] [code] *8/10 - Clear Success (achieves goal)* [continue] [code] *9/10 - Strong Success (achieves goal cleanly)* [continue] [code] *10/10 - Critical Success (exceeds expectations)* [continue] This is CRO's foundational scale – the default when you're not using a skill-specific approach. Its 10 outcomes use deliberately broad language: that gives the AI maximum interpretive freedom while still providing clear guidance about the action's general trajectory. It’s universal: whether you're attempting to convince a guard, climb a wall, or solve a riddle, these outcome descriptions adapt naturally to the context. I designed this as the primary scale because it captures the essential dramatic beats without constraining too much how those beats manifest. Once again, these are relatively open-ended descriptions, and will work best with more powerful AI models: ones that can take a broad narrative direction and develop it meaningfully, if not creatively. <div class="cro-passage-nav"> <div><span class="nav-hub">[[Quick Reference Hub]]</span></div> <div>Also found in: <span class="nav-next">[[Core Mechanics->CRO Core Mechanics]]</span></div> </div>

<div class="cro-passage-nav" style="border-top: none; padding-top: 0; margin-top:0; margin-bottom: 1.5em;"> <div class="nav-hub">[[Quick Reference Hub]]</div> </div> <h1 class="cro-h1">Physical Scale (Detail)</h1> <h4 class="cro-h3">Physical Scale</h3> (Strength, Athletics, Endurance, Combat Maneuvers) [code] *1/10 - Harsh Consequence (clear physical cost or injury)* [continue] [code] *2/10 - Strained Effort (body struggles, fatigue sets in)* [continue] [code] *3/10 - Forced Failure (pushing through but can't complete)* [continue] [code] *4/10 - Rough Attempt (partial progress, noticeable strain)* [continue] [code] *5/10 - Taxing Success (achieves goal but at physical cost)* [continue] [code] *6/10 - Adequate Performance (body cooperates, gets it done)* [continue] [code] *7/10 - Steady Control (confident execution)* [continue] [code] *8/10 - Strong Performance (body responds well)* [continue] [code] *9/10 - Smooth Execution (effortless feel)* [continue] [code] *10/10 - Perfect Form (exactly as intended, if not better)* [continue] Physical actions present a unique challenge: they carry real consequences that can impact your entire ongoing narrative. A failed climb isn't just "you don't get up the wall" – it might mean injury, exhaustion, or equipment damage that matters for the next hour of play. That's why I settled on 10 outcomes here, giving you enough granularity to distinguish between "winded but fine" and "twisted ankle that'll slow you down." The scale focuses on the relationship between your character's intention and their body's response rather than just success/failure, because in my testing, that relationship created far more interesting ongoing complications than simple binary outcomes. <div class="cro-passage-nav"> <div>Also found in: <span class="nav-next">[[Core Mechanics->CRO Core Mechanics]]</span></div> <div><span class="nav-hub">[[Quick Reference Hub]]</span></div> </div>

<div class="cro-detail-top-nav"> <div class="nav-hub">[[Quick Reference Hub]]</div> </div> <h1 class="cro-h1">Manual Scale (Detail)</h1> <h3 class="cro-h3">Manual Scale</h3> (Dexterity, Crafting, Fine Motor Skills, Delicate Tasks) [code] *1/10 - Complete Mishap (task goes wrong, clear setback)* [continue] [code] *2-3/10 - Sloppy Work (barely functional, obvious flaws)* [continue] [code] *4-5/10 - Rough Execution (works but crude, noticeable imperfections)* [continue] [code] *6-7/10 - Competent Work (solid execution, minor rough edges)* [continue] [code] *8-9/10 - Clean Execution (smooth work, well-handled)* [continue] [code] *10/10 - Precise Control (exactly as intended, no wasted motion)* [continue] Originally, I had manual skills lumped in with physical ones, but I quickly realized they needed their own approach. When you're picking a lock or crafting something, the question isn't really "did it work" but "how well did you execute it?" A crude lockpick job might open the door but leave obvious scratches, while masterful work leaves no trace. Seven outcomes gives you enough range to capture that spectrum without bogging down the narrative in excessive detail. The focus on execution quality over mere functionality emerged from countless sessions where the difference between "it works" and "it works elegantly" drove the next scene's direction. <div class="cro-passage-nav"> <div><span class="nav-hub">[[Quick Reference Hub]]</span></div> <div>Also found in: <span class="nav-next">[[Skill-Specific Scales->CRO Skill-Specific Scales]]</span></div> </div>

<div class="cro-detail-top-nav"> <div class="nav-hub">[[Quick Reference Hub]]</div> </div> <h1 class="cro-h1">Social Scale (Detail)</h1> <h3 class="cro-h3">Social Scale</h3> (Persuasion, Deception, Intimidation, Charm, Negotiation) [code] *1-2/10 - Momentum Lost (interaction creates resistance or withdrawal)* [continue] [code] *3-4/10 - Awkward Exchange (stilted, uncomfortable social flow)* [continue] [code] *5-6/10 - Neutral Transaction (stalling, neither builds nor damages social standing)* [continue] [code] *7-8/10 - Strong Rapport (builds connection and cooperation)* [continue] [code] *9-10/10 - Lasting Impact (memorable interaction, doors open)* [continue] Social interactions... I tend not to roll for these, actually, I prefer to talk in character, and try to convince/intimidate the chatbot that way. So I actually haven't tested this scale too hard! I'm open to improvements, as I am with the rest of the system's components. Five outcomes captures the essential range from "actively damages the relationship" to "creates lasting positive impression" without overthinking it. During development, I noticed that focusing on social momentum and interpersonal energy worked better than specific persuasion mechanics. For example, rolling high doesn't necessarily mean "they agree with you," but rather "you've created positive social energy, and that may open doors." This keeps the AI focused on relationship dynamics rather than forcing specific character decisions. <div class="cro-passage-nav"> <div><span class="nav-hub">[[Quick Reference Hub]]</span></div> <div>Also found in: <span class="nav-next">[[Skill-Specific Scales->CRO Skill-Specific Scales]]</span></div> </div>

<div class="cro-detail-top-nav"> <div class="nav-hub">[[Quick Reference Hub]]</div> </div> <h1 class="cro-h1">Perception Scale (Detail)</h1> <h3 class="cro-h3">Perception Scale</h3> (Awareness, Investigation, Searching, Noticing, Sensing Motives) [code] *1-2/10 - Physical State (ex. appearance, condition, immediate qualities, etc.)* [continue] [code] *3-4/10 - Functional Aspect (ex. role, purpose, capabilities, etc.)* [continue] [code] *5-6/10 - Contextual Clues (ex. connections, significance, anomalies, etc.)* [continue] [code] *7-8/10 - Hidden Elements (ex. concealed aspects, overlooked details, etc.)* [continue] [code] *9-10/10 - Temporal Traces (ex. history, recent changes, accumulated effects, etc.)* [continue] This 5-outcome scale works differently from the others – instead of measuring how well you perceive, it determines what type of perception you apply. Each outcome represents a different investigative lens: you might focus on physical condition, functional purpose, contextual relationships, hidden features, or traces of recent activity. This approach emerged from a practical problem with traditional perception mechanics: telling the AI that you "notice more things" often produced vague atmospheric descriptions that gave players nothing concrete to pursue. Oftentimes, the instruction was completely ignored. By specifying the type of observation, the scale consistently delivers actionable information. Rolling for "functional aspect" when examining a door gives you insights about its purpose and mechanisms, while "temporal traces" might reveal recent use patterns – both useful for planning your next action, I feel. <div class="cro-passage-nav"> <div><span class="nav-hub">[[Quick Reference Hub]]</span></div> <div>Also found in: <span class="nav-next">[[Skill-Specific Scales->CRO Skill-Specific Scales]]</span></div> </div>

<div class="cro-detail-top-nav"> <div class="nav-hub">[[Quick Reference Hub]]</div> </div> <h1 class="cro-h1">Fortune Scale (Detail)</h1> <h3 class="cro-h3">Fortune Scale</h3> (Luck, Chance, Timing, Random Circumstance) [code] *1/10 - Bad Timing (unfavorable coincidence)* [continue] [code] *2-5/10 - Unlucky Turn (circumstances hinder, disadvantage)* [continue] [code] *6-9/10 - Lucky Break (favorable circumstance, advantage)* [continue] [code] *10/10 - Perfect Timing (major positive turn of events)* [continue] Fortune is pure circumstance, isn't it? Things either go your way or they don't. I originally had a complex 10-outcome scale with dramatic language like "calamitous misfortune," but found it was pulling the AI into over-the-top territory. Bad timing is just bad timing; lucky breaks are just lucky breaks. Sometimes the most elegant design choice is restraint, and 4, well-defined outcomes work well enough. <div class="cro-passage-nav"> <div><span class="nav-hub">[[Quick Reference Hub]]</span></div> <div>Also found in: <span class="nav-next">[[Skill-Specific Scales->CRO Skill-Specific Scales]]</span></div> </div>

<div class="cro-detail-top-nav"> <div class="nav-hub">[[Quick Reference Hub]]</div> </div> <h1 class="cro-h1">Stealth Scale (Detail)</h1> <h3 class="cro-h3">Stealth Scale</h3> (Sneaking, Hiding, Moving Silently - Replaces default Physical/Manual for these actions) [code] *1/10 - Fully Exposed (immediate detection, alarm raised)* [continue] [code] *2/10 - Clearly Spotted (presence and intent obvious)* [continue] [code] *3/10 - Obviously Noticed (seen but intent unclear)* [continue] [code] *4/10 - Suspicion Raised (traces left, heightened awareness)* [continue] [code] *5/10 - Minor Disturbance (small signs noticed)* [continue] [code] *6/10 - Barely Avoided (close call, narrowly unnoticed)* [continue] [code] *7/10 - Successfully Hidden (avoided direct detection)* [continue] [code] *8/10 - Cleanly Unnoticed (no awareness triggered)* [continue] [code] *9/10 - Seamless Passage (no signs of presence)* [continue] [code] *10/10 - Perfect Concealment (completely undetected, no trace)* [continue] Stealth demands granularity because detection operates on a gradient, and small differences have big consequences. There's a meaningful distinction between "guards are suspicious" and "guards are actively searching," and your options change dramatically between those states. These 10 outcomes let the AI calibrate responses precisely, from "completely blown" to "ghosted through undetected." I kept this one linear because stealth is fundamentally about degrees of concealment – it's one of the few skills where more complex categorization would actually reduce clarity. <div class="cro-passage-nav"> <div><span class="nav-hub">[[Quick Reference Hub]]</span></div> <div>Also found in: <span class="nav-next">[[Skill-Specific Scales->CRO Skill-Specific Scales]]</span></div> </div>

<h1 class="cro-h1">Stakes (Extension Detail)</h1> <h3 class="cro-h3">7.1 Stakes</h3> Raise or lower the stakes before a roll (mentally) by 1-3: <span class="cro-inline-code">- Raising Stakes (+1 to +3)</span> * Higher payoff if successful, worse outcome if you fail <span class="cro-inline-code">- Lowering Stakes (-1 to -3)</span> * Softer fail consequences, but capped success If roll ≥ 5, add the chosen “stakes” number to your final result (up to 10). If roll < 5, subtract the “stakes” number (down to 1). You do this mentally and only show the final adjusted roll. Sometimes the standard difficulty modifiers don't capture what you're really after. Stakes let you amplify both the potential reward and the potential cost of an action without changing the basic difficulty. This emerged from situations where I wanted to attempt something that wasn't necessarily harder, but where the consequences mattered more. Climbing a wall during a casual exploration versus climbing that same wall while being chased by guards - same physical challenge, completely different stakes. The mechanic lets you push for bigger successes while accepting that failures will hurt more, creating that "high risk, high reward" feeling. It makes certain moments feel memorable, though, to be fair, I don’t use it often. This is a common pattern along these extensions, though not all, which is the point. You use the ones that feel right to you, or the situation at hand. <div class="cro-passage-nav"> <div><span class="nav-hub">[[Quick Reference Hub]]</span></div> <div>Also found in: <span class="nav-next">[[Extensions Collection->CRO Extensions Collection]]</span></div> </div>

<div class="cro-detail-top-nav"> <div class="nav-hub">[[Quick Reference Hub]]</div> </div> <h1 class="cro-h1">Pushing Limits (Extension Detail)</h1> <h3 class="cro-h3">7.2 Pushing Limits</h3> Allows a character to attempt an action far beyond their normal capabilities or under extreme duress by accepting a significantly higher risk of failure and more severe consequences. <h4 class="cro-h4">Format</h4> [code] I attempt to push my limit by [describe extreme action]. **[SKILL] ROLL** *[Result]/10 - [Pushing Limits Outcome]* [continue] <h4 class="cro-h4">Pushing Limits Scale</h4> The player must use this harsher scale when "Pushing Limits" is invoked, reflecting the mentally increased difficulty threshold. It replaces the Main Outcome Scale for the given roll’s outcome. [code] *1/10 - Severe Backfire (major negative consequence)* [continue] [code] *2-3/10 - Costly Failure (significant setback, clear price paid)* [continue] [code] *4-5/10 - Painful Attempt (falls short, noticeable strain or cost)* [continue] [code] *6-7/10 - Strained Success (goal met, but with visible effort)* [continue] [code] *8-9/10 - Hard-Won Achievement (success despite the odds)* [continue] [code] *10/10 - Breakthrough Performance (exceeds expectations)* [continue] When your character attempts something far beyond normal capability, consequences should be harsher than usual, but I learned that overly dramatic language ("catastrophic failure") pushed the AI toward melodramatic responses. Also, who wants to use a scale that just destroys your character? Sure, it's narratively interesting... for a quick and brutal ending to your session, that is. This 7-outcome scale compresses the success range deliberately – most attempts to push limits should be difficult, with only the highest rolls achieving clean success. This reflects the inherent danger of exceeding normal boundaries while keeping it grounded. <div class="cro-passage-nav"> <div><span class="nav-hub">[[Quick Reference Hub]]</span></div> <div>Also found in: <span class="nav-next">[[Extensions Collection->CRO Extensions Collection]]</span></div> </div>

<div class="cro-detail-top-nav"> <div class="nav-hub">[[Quick Reference Hub]]</div> </div> <h1 class="cro-h1">Skill Leverage (Extension Detail)</h1> <h3 class="cro-h3">7.3 Skill Leverage</h3> A slight adjustment that allows players to declare using a secondary, relevant skill to influence their primary action by letting the chatbot be more aware of what skills the player is leveraging. Simply declare your primary action (Skill A) and the leveraged skill (Skill B). [code] *I attempt to [Primary Action] using [Skill A], leveraging my [Skill B] to [Describe how B helps A].* **[SKILL A]&[SKILL B] ROLL** *[Result]/10 - [Main Outcome Scale]* [continue] This simple addition addresses a common frustration: when your character's approach clearly involves multiple types of expertise, but the system only recognizes one. Using your knowledge of architecture to inform your climbing approach, or leveraging your social skills to make a manual task easier by getting help - these combinations happen naturally in stories but can feel overlooked by single-skill rolls. Skill Leverage doesn't change the mechanics, just makes the AI more aware of your character's multi-faceted approach. It's particularly useful with less sophisticated AI models that might miss the connection between different skill domains. <div class="cro-passage-nav"> <div><span class="nav-hub">[[Quick Reference Hub]]</span></div> <div>Also found in: <span class="nav-next">[[Extensions Collection->CRO Extensions Collection]]</span></div> </div>

<div class="cro-detail-top-nav"> <div class="nav-hub">[[Quick Reference Hub]]</div> </div> <h1 class="cro-h1">Serendipity (Extension Detail)</h1> <h3 class="cro-h3">7.4 Serendipity</h3> When a roll results in a mixed outcome (4-6 on the Main Outcome Scale), an opportunity emerges: you can add another action to the original roll. The outcome of this secondary action is decided by flipping a coin, allowing the user to get to shift towards a partial success, or a failure. [code] *I attempt to [describe original action].* **[PRIMARY SKILL] ROLL** *[Result]/10 - [Outcome]* **SERENDIPITY ACTION** *Coin Flip - Heads (Success)* or *Coin Flip - Tails (Failure)* [continue] Mixed results create natural story tension, but sometimes they leave you hanging in an unsatisfying middle ground. Serendipity gives you a way to push through that ambiguity when it feels appropriate. The coin flip keeps it simple - no complex modifiers or new scales to remember. Just you and Mr.Lincoln, partners in crime. <div class="cro-passage-nav"> <div><span class="nav-hub">[[Quick Reference Hub]]</span></div> <div>Also found in: <span class="nav-next">[[Extensions Collection->CRO Extensions Collection]]</span></div> </div>

<div class="cro-detail-top-nav"> <div class="nav-hub">[[Quick Reference Hub]]</div> </div> <h1 class="cro-h1">Targeted Effort (Extension Detail)</h1> <h3 class="cro-h3">7.5 Targeted Effort</h3> Allows a player to declare a more specific, difficult goal within a standard action attempt, aiming for precision rather than general success. Use this when achieving a precise outcome is more important than simply succeeding at the general task (e.g., hitting a specific weak point, finding a particular clue, using exact leverage). This offers the potential for more impactful success but increases the risk of failure due to the inherent difficulty of precision. 1. Declare your action using the standard format but include a specific target or refined goal within your description. 2. Mentally apply a -2 modifier to your d10 roll result before presenting it. This reflects the increased difficulty of the specific aim. 3. Present the final, adjusted <span class="cro-inline-code">[Result]/10</span> using the standard <span class="cro-inline-code">**[SKILL] ROLL**</span> format and outcome scales. <div class="cro-passage-nav"> <div><span class="nav-hub">[[Quick Reference Hub]]</span></div> <div>Also found in: <span class="nav-next">[[Extensions Collection->CRO Extensions Collection]]</span></div> </div>

<div class="cro-detail-top-nav"> <div class="nav-hub">[[Quick Reference Hub]]</div> </div> <h1 class="cro-h1">Guarded Approach (Extension Detail)</h1> <h3 class="cro-h3">7.6 Guarded Approach</h3> Allows a player to perform an action with deliberate caution, aiming to minimize the risk of critical failure at the cost of potential critical success. Use this when avoiding the absolute worst outcome (a roll of 1) is more important than achieving the absolute best (a roll of 10). It's a risk-averse strategy suitable for high-danger situations where a standard failure is acceptable, but a disaster is not. 1. Declare your action using the standard format and specify you are taking a <span class="cro-inline-code">**Guarded Approach**</span>. 2. Roll the d10 for the relevant skill. 3. Mentally adjust the roll **only** if it is a 1 or a 10: * If the roll is 1, treat it as 2. * If the roll is 10, treat it as 9. 4. Present the final (potentially adjusted) <span class="cro-inline-code">[Result]/10</span> using the standard <span class="cro-inline-code">**[SKILL] ROLL**</span> format. <div class="cro-passage-nav"> <div><span class="nav-hub">[[Quick Reference Hub]]</span></div> <div>Also found in: <span class="nav-next">[[Extensions Collection->CRO Extensions Collection]]</span></div> </div>

<div class="cro-detail-top-nav"> <div class="nav-hub">[[Quick Reference Hub]]</div> </div> <h1 class="cro-h1">Item Focus (Extension Detail)</h1> <h3 class="cro-h3">8.1 Item Focus</h3> For examining or interacting with a specific physical object to understand its condition, properties, or potential uses. <h4 class="cro-h4">Format</h4> [code] *I examine the [Item Name] closely.* **ITEM FOCUS ROLL** *[Result]/10 - [Item Insight/Condition]* [continue] <h4 class="cro-h4">Item Insight Scale</h4> [code] *1/10 - Material Properties (ex. composition, durability, craftsmanship quality, etc.)* [continue] [code] *2-3/10 - Functional Design (ex. intended purpose, how it works, efficiency, etc.)* [continue] [code] *4-5/10 - Historical Context (ex. age, previous use, wear patterns, etc.)* [continue] [code] *6-7/10 - Hidden Features (ex. concealed mechanisms, subtle details, etc.)* [continue] [code] *8-9/10 - Contextual Significance (ex. relevance to situation, connections, etc.)* [continue] [code] *10/10 - Symbolic Meaning (ex. cultural importance, personal resonance, etc.)* [continue] As an extension of the perception system, this 7-outcome scale applies the same domain-based approach to examining specific objects. Rather than forcing the AI to determine how "much" an item is, it focuses your attention on different aspects: material properties, functional design, historical context, and so on. Seven outcomes provide good coverage of investigation approaches without overwhelming a mechanic that's meant to be used sparingly in the first place. <div class="cro-passage-nav"> <div><span class="nav-hub">[[Quick Reference Hub]]</span></div> <div>Also found in: <span class="nav-next">[[Extensions Collection->CRO Extensions Collection]]</span></div> </div>

<div class="cro-detail-top-nav"> <div class="nav-hub">[[Quick Reference Hub]]</div> </div> <h1 class="cro-h1">Inventory Check (Extension Detail)</h1> <h3 class="cro-h3">8.2 Inventory Check</h3> Determines what mundane, useful, or unexpected items a character finds when checking their pockets, bag, or immediate personal effects without searching for something specific. This is a way to force the chatbot AI to give you some items to play with. The result indicates the quantity and general nature/relevance of the items found. <h4 class="cro-h4">Format Examples</h4> [code] *I pat down my pockets to see what I have on me.* *I rummage through my bag to see what I have on me.* *I check to see what I have on me.* **INVENTORY CHECK** *[Result]/10 - [Find Description]* [continue] <h4 class="cro-h4">Find Scale</h4> [code] *1/10 - Empty Search (nothing useful found)* [continue] [code] *2-3/10 - Mundane Items (basic, expected contents)* [continue] [code] *4-5/10 - Useful Basics (practical items for current situation)* [continue] [code] *6-7/10 - Helpful Resources (relevant, contextually appropriate)* [continue] [code] *8-9/10 - Valuable Discovery (significantly useful items)* [continue] [code] *10/10 - Perfect Find (exactly what's needed right now)* [continue] <h4 class="cro-h4">Usage Notes for Inventory Check</h4> The find scale had a habit of derailing narratives. High rolls would produce elaborate, setting-defining items that the AI felt compelled to make significant, while low rolls left you empty-handed in situations where some basic supplies made narrative sense. This 7-outcome scale emphasizes contextual appropriateness over raw value – a perfect roll gives you "exactly what's needed right now" rather than "something amazingly valuable." This keeps the extension useful without letting it hijack your story's direction. Again, use sparingly. We don’t want the AI getting overwhelmed, nor your character ending up with an endless amount of items. Maybe every TIME roll? Your call. <div class="cro-passage-nav"> <div><span class="nav-hub">[[Quick Reference Hub]]</span></div> <div>Also found in: <span class="nav-next">[[Extensions Collection->CRO Extensions Collection]]</span></div> </div>

<div class="cro-detail-top-nav"> <div class="nav-hub">[[Quick Reference Hub]]</div> </div> <h1 class="cro-h1">Random Events (Extension Detail)</h1> <h3 class="cro-h3">9.1 Random Events</h3> Random events needed their own logic. The linear "how important is this" approach often produced either trivial distractions or story-hijacking developments. This scale focuses on external event types – weather shifts, people movement, objects – that add texture without forcing major plot changes. Each category suggests environmental dynamism (big words for big people) rather than dramatic revelation, keeping random events in their proper place: a supporting role. <h4 class="cro-h4">Format</h4> [code] **RANDOM EVENT** *[Result]/10 - [Impact Description]* [continue] You can also roll a d10 (1-7 = No event, 8-10 = Event occurs) to see if a random event happens at all. <h4 class="cro-h4">Random Event Scale</h4> [code] *1/10 - Weather Shift (ex. sudden climate change, atmospheric conditions, etc.)* [continue] [code] *2/10 - Social Arrival (ex. people appear, someone approaches, crowd gathers, etc.)* [continue] [code] *3/10 - Object Appearance (ex. item falls, tool breaks, something delivered, etc.)* [continue] [code] *4/10 - Communication Event (ex. message received, overhear conversation, etc.)* [continue] [code] *5/10 - Environmental Change (ex. door opens, path blocked, noise starts, etc.)* [continue] [code] *6/10 - Timing Coincidence (ex. simultaneous events, scheduling overlaps, etc.)* [continue] [code] *7/10 - Authority Attention (ex. official notice, rule enforcement, inspection, etc.)* [continue] [code] *8/10 - Resource Fluctuation (ex. supply issue, price change, availability shift, etc.)* [continue] [code] *9/10 - Technical Disruption (ex. equipment malfunction, system failure, etc.)* [continue] [code] *10/10 - Information Surface (ex. documents appear, secrets revealed accidentally, etc.)* [continue] <div class="cro-passage-nav"> <div><span class="nav-hub">[[Quick Reference Hub]]</span></div> <div>Also found in: <span class="nav-next">[[Extensions Collection->CRO Extensions Collection]]</span></div> </div>

<div class="cro-detail-top-nav"> <div class="nav-hub">[[Quick Reference Hub]]</div> </div> <h1 class="cro-h1">Emotional Static (Extension Detail)</h1> <h3 class="cro-h3">9.2 Emotional Static</h3> Introduce brief, surreal, and non-literal manifestations of a character's strong underlying emotion (like intense fear, jealousy, joy, despair) into the surrounding environment or perception, creating bizarre or unsettling narrative moments independent of direct action. <h4 class="cro-h4">Format</h4> [code] *(Player describes character experiencing a strong underlying emotion)* *I focus on this feeling of [briefly name emotion, e.g., intense regret, blinding hope].* **EMOTIONAL STATIC ROLL** *[Result]/10 - [Static Manifestation Description] (triggered by [Emotion Name])* [continue] <h4 class="cro-h4">Static Manifestation Scale</h4> [code] *1-2/10 - Sensory Whisper (ex. fleeting sound, incongruous smell, etc.)* [continue] [code] *3-4/10 - Perceptual Drift (ex. visual ripple, misplaced shadow, etc.)* [continue] [code] *5-6/10 - Strange Occurrence (ex. impossible object appears, unnatural sensation, etc.)* [continue] [code] *7-8/10 - Reality Distortion (ex. environment briefly warps, spectral image, etc.)* [continue] [code] *9-10/10 - Psychic Feedback (ex. physics briefly altered, reality glitch, etc.)* [continue] Emotional static... this one is weird. Sure enough, it was suggested to me by Claude - I wanted to put at least one example of a completely AI-generated extension, because why not. There is some potential. It represents surreal moments where strong feelings briefly warp perception or reality. Five outcomes provide enough range from subtle sensory whispers to reality glitches, encouraging the AI to go completely off the rails, or get very meta. Maybe you want that for your story! The challenge was calibrating weirdness – too subtle and it becomes forgettable, too intense and it overwhelms the main narrative thread too quickly. <div class="cro-passage-nav"> <div><span class="nav-hub">[[Quick Reference Hub]]</span></div> <div>Also found in: <span class="nav-next">[[Extensions Collection->CRO Extensions Collection]]</span></div> </div>

<div class="cro-detail-top-nav"> <div class="nav-hub">[[Quick Reference Hub]]</div> </div> <h1 class="cro-h1">Interference/Nemesis (Extension Detail)</h1> <h3 class="cro-h3">9.3 Interference/Nemesis</h3> (Formerly 9.4 in guide) Introduce a competing force or character that actively works against the player's goal. To be placed right after an action. <h4 class="cro-h4">Format</h4> [code] **INTERFERENCE ROLL** *[Result]/10 - [Opposition Level]* [continue] <h4 class="cro-h4">Opposition Scale</h4> [code] *1/10 - Minimal Hindrance (barely noticeable resistance)* [continue] [code] *2-3/10 - Minor Obstacles (slight complications, easy to work around)* [continue] [code] *4-5/10 - Notable Resistance (requires attention, strategic adjustment)* [continue] [code] *6-7/10 - Active Opposition (substantial challenge, complicates approach)* [continue] [code] *8-9/10 - Strong Interference (significantly threatens success)* [continue] [code] *10/10 - Direct Confrontation (goal requires completely new approach)* [continue] Opposition should escalate your challenge level without necessarily changing your fundamental approach until the highest levels. "Minor obstacles" require attention but don't force you to completely rethink your strategy, while "direct confrontation" demands a new plan entirely. The scale gives the AI clear guidance on how much resistance to apply to your actions. <div class="cro-passage-nav"> <div><span class="nav-hub">[[Quick Reference Hub]]</span></div> <div>Also found in: <span class="nav-next">[[Extensions Collection->CRO Extensions Collection]]</span></div> </div>

<div class="cro-detail-top-nav"> <div class="nav-hub">[[Quick Reference Hub]]</div> </div> <h1 class="cro-h1">Montage (Extension Detail)</h1> <h3 class="cro-h3">10.1 Montage</h3> Progress through multiple actions over a longer period. [code] **MONTAGE ROLL** *[Time Period]/10 - [Total length of time that passes] - [Number of actions] Actions* [continue] Then break it down into segments, each with its own roll. Example: [code] **MONTAGE ROLL** *6/10 - 6 MONTHS - 4 Actions* **FIRST MONTH** *1/4 - [Action]* **[SKILL] ROLL** *[Result]/10 - [Outcome]* *2/4 - [Action]* **[SKILL] ROLL** *[Result]/10 - [Outcome]* **MONTHS 2-4** *3/4 - [Action]* **[SKILL] ROLL** *[Result]/10 - [Outcome]* *4/4 - [Action]* **[SKILL] ROLL** *[Result]/10 - [Outcome]* [continue] Traditional scene-by-scene play sometimes hits periods where you need to cover significant time and multiple attempts without losing narrative momentum. Montage provides structure for those "training sequences" or "investigation periods" where the journey matters but doesn't need minute-by-minute detail. The initial roll determines scope and pace, then the segmented approach lets you highlight key moments without getting bogged down. This works particularly well for character development periods, skill learning, or any situation where the pattern of attempts and outcomes tells a more interesting story than any single action would. <div class="cro-passage-nav"> <div><span class="nav-hub">[[Quick Reference Hub]]</span></div> <div>Also found in: <span class="nav-next">[[Extensions Collection->CRO Extensions Collection]]</span></div> </div>

<div class="cro-detail-top-nav"> <div class="nav-hub">[[Quick Reference Hub]]</div> </div> <h1 class="cro-h1">Narrative Techniques (Extension Detail)</h1> <h3 class="cro-h3">10.2 Narrative Techniques</h3> This extension provides direct communication methods for shaping your story without relying on dice. These are tools for active, more direct collaboration - sometimes you recognize what a scene is missing, and shouldn’t be forced to hope for a certain roll’s outcome to fix it. Use them to guide the AI toward specific narrative goals through clear signaling. Communication is key, as they say. <h4 class="cro-h4">Atmospheric Direction</h4> [code] * (I'd like some environmental details here) * (Something feels off about this place) * (The mood should feel [tense/peaceful/mysterious]) [continue] <h4 class="cro-h4">Scene Management</h4> [code] * "I take a moment to look around the room..." (shifts focus from dialogue to environment) * "After we finish talking..." (signals conversation should wrap up) * "Before I respond..." (creates space for reflection or action) [continue] <h4 class="cro-h4">Pacing Control</h4> [code] * "Quickly, I..." or "I take my time..." (guides scene speed) * "Without hesitation..." vs "I pause, considering..." (controls decision tempo) * (Let's move this along) or (Let's slow down and explore this) [continue] <h4 class="cro-h4">Thread Management</h4> [code] * "I remember [person/event] and..." (brings back dormant story elements) * "This reminds me of..." (connects to established narrative threads) * "Speaking of [topic]..." (steers conversation direction) [continue] <h4 class="cro-h4">Emotional Anchoring</h4> [code] * Starting actions with internal states: "Feeling nervous, I..." "Curious about her reaction, I..." * (My character should be more [confident/worried/excited] here) [continue] <h4 class="cro-h4">Environmental Prompting</h4> [code] * "The sound of [footsteps/machinery/voices] catches my attention..." * "I notice..." (directs AI to provide specific observational details) * "Something about [object/person] strikes me as..." (requests focused description) [continue] <div class="cro-passage-nav"> <div><span class="nav-hub">[[Quick Reference Hub]]</span></div> <div>Also found in: <span class="nav-next">[[Extensions Collection->CRO Extensions Collection]]</span></div> </div>

<div class="cro-detail-top-nav"> <div class="nav-hub">[[Quick Reference Hub]]</div> </div> <h1 class="cro-h1">Narrative Momentum (Extension Detail)</h1> <h3 class="cro-h3">10.3 Narrative Momentum</h3> When you want to inject unexpected energy into a scene without controlling the specifics, roll for narrative momentum. This creates genuine surprise by shifting the emotional or dramatic tone in ways you might not have anticipated. <h4 class="cro-h4">Format</h4> [code] *I want to shift the energy of this scene.* **NARRATIVE MOMENTUM ROLL** *[Result]/10 - [Energy Type]* [continue] <h4 class="cro-h4">Narrative Energy Scale</h4> [code] *1/10 - Mysterious Energy (ex. unexplained elements, hidden significance, secrets, etc.)* [continue] [code] *2/10 - Urgent Pressure (ex. time constraints, approaching deadlines, rushing, etc.)* [continue] [code] *3/10 - Melancholic Mood (ex. nostalgia, loss, bittersweet moments, etc.)* [continue] [code] *4/10 - Tense Undercurrent (ex. brewing conflict, unspoken tension, etc.)* [continue] [code] *5/10 - Contemplative Tone (ex. reflection, philosophical moments, quiet thought, etc.)* [continue] [code] *6/10 - Playful Spirit (ex. humor, lightness, unexpected fun, etc.)* [continue] [code] *7/10 - Intimate Atmosphere (ex. personal connection, vulnerability, closeness, etc.)* [continue] [code] *8/10 - Dramatic Intensity (ex. heightened stakes, powerful emotions, etc.)* [continue] [code] *9/10 - Peaceful Flow (ex. calm, harmony, gentle moments, etc.)* [continue] [code] *10/10 - Chaotic Disruption (ex. unpredictability, wild energy, upheaval, etc.)* [continue] This roll determines what kind of energy enters the scene, letting the AI manifest it naturally within your current situation. Unlike specific suggestions, you're genuinely open to whatever type of momentum emerges. <div class="cro-passage-nav"> <div><span class="nav-hub">[[Quick Reference Hub]]</span></div> <div>Also found in: <span class="nav-next">[[Extensions Collection->CRO Extensions Collection]]</span></div> </div>

<div class="cro-detail-top-nav"> <div class="nav-hub">[[Quick Reference Hub]]</div> </div> <h1 class="cro-h1">Judge System Overview (Detail)</h1> <h2 class="cro-h2">Part 3: Judge System</h2> While self-managed rolls work beautifully most of the time, there's a human element that can sometimes get in the way: bias. Unconscious, maybe. We’re not talking about dramatic cheating, mind you! More so subtle things - maybe you think that climb should be rated "Easy" instead of "Moderate" because you really want to reach that tower, or a social roll becomes "Difficult" because you're not sure how to handle success in that moment. You are both a player wanting things to go well, and a referee trying to assess fairly. Most of the time, this bias actually improves the experience. You naturally adjust difficulty to maintain good pacing and avoid crushing setbacks when they'd derail interesting storylines. But sometimes you want the dice to be genuinely impartial - during crucial moments where the uncertainty itself creates tension, when you're trying to discover what your character is truly capable of, or when you want to be surprised by the chatbot’s storytelling. The Judge AI system splits these responsibilities: a separate AI instance handles all mechanical rulings while your main storytelling AI focuses purely on narrative response. It's entirely optional, and honestly, most players will prefer the flexibility of self-managed rolls. I sure do. But when you want that "anything could happen" feeling, or when you catch yourself nudging results toward desired outcomes, having an impartial referee ensures the system works exactly as designed. For detailed setup instructions and the full prompt, see [[Judge System Setup->CRO Judge System Setup]]. <div class="cro-passage-nav"> <div><span class="nav-hub">[[Quick Reference Hub]]</span></div> <div>Also found in: <span class="nav-next">[[Advanced Tools & Examples->CRO Advanced Tools & Examples]]</span></div> </div>

<div class="cro-detail-top-nav"> <div class="nav-hub">[[Quick Reference Hub]]</div> </div> <h1 class="cro-h1">Implementation Examples (Detail)</h1> <h2 class="cro-h2">12.0 Implementation Examples</h2> <h3 class="cro-h3">Ex.1 - Mental Approach (Cognitive Modes)</h3> [code] *I try to figure out how this ancient mechanism works.* **MENTAL ROLL** *6/10 - Pattern Recognition (ex. connections, similarities, recurring themes, etc.)* [continue] (AI might narrate: Looking at the gears and symbols, you start noticing similarities to a clockwork design you've seen before. The recurring spiral motifs suggest this follows the same principle as those old Meridian automatons...) <div class="cro-hr"></div> <h3 class="cro-h3">Ex. 2 - Perception Investigation (Domain Focus)</h3> [code] *I examine this abandoned camp carefully.* **PERCEPTION ROLL** *7/10 - Hidden Elements (ex. concealed aspects, overlooked details, etc.)* [continue] (AI might narrate: Most of the obvious details are clear - cold fire pit, scattered supplies - but something catches your eye. One of the tent stakes has been driven in at an odd angle, and there's a small piece of fabric caught underneath, deliberately placed.) <div class="cro-hr"></div> <h3 class="cro-h3">Ex. 3 - Physical vs Manual Distinction</h3> [code] *I attempt to force this rusted door open with my shoulder.* **PHYSICAL ROLL** *4/10 - Rough Attempt (partial progress, noticeable strain)* [continue] (AI might narrate: You slam into the door repeatedly. It shifts slightly in its frame, but your shoulder aches from the impact...) [code] *Instead, I try to work the lock mechanism carefully.* **MANUAL ROLL** *7/10 - Competent Work (solid execution, minor rough edges)* [continue] (AI might narrate: With patient manipulation, you feel the tumblers respond. The lock clicks open, though you notice you've left tiny metal filings around the keyhole.) <div class="cro-hr"></div> <h3 class="cro-h3">Ex. 4 - Using Narrative Techniques</h3> [code] *(I'd like to shift focus to what's outside the window)* *I pause mid-conversation and glance toward the street.* [continue] (AI responds by describing activity outside, transitioning naturally from dialogue to environmental observation) <div class="cro-hr"></div> <h3 class="cro-h3">Ex. 5 - Item Focus Extension</h3> [code] *I examine the merchant's ledger more closely.* **ITEM FOCUS ROLL** *3/10 - Functional Design (ex. intended purpose, how it works, efficiency, etc.)* [continue] (AI might narrate: This isn't just a simple record book. The margins contain a second notation system - coded references that seem to track something other than normal transactions. It's designed for dual purposes.) <div class="cro-hr"></div> <h3 class="cro-h3">Ex. 6 - Inventory Check Extension</h3> [code] *I rummage through my pack to see what supplies I have.* **INVENTORY CHECK** *8/10 - Valuable Discovery (significantly useful items)* [continue] (AI might narrate: Among your basic travel gear, you find a small leather pouch containing silver coins you'd forgotten about, plus a quality rope that's exactly what you need for the climb ahead.) <div class="cro-hr"></div> <h3 class="cro-h3">Ex. 7 - Narrative Momentum Extension</h3> [code] *This conversation is getting stale. I want to shift the energy of this scene.* **NARRATIVE MOMENTUM ROLL** *4/10 - Tense Undercurrent (ex. brewing conflict, unspoken tension, etc.)* [continue] (AI might respond by having the NPC's expression subtly shift, or introducing an interruption that raises stakes, or revealing that someone's been listening...) <div class="cro-hr"></div> <h3 class="cro-h3">Ex. 8 - Main Scale Flexibility</h3> [code] *I try to convince the innkeeper to let us stay despite having no money.* **SOCIAL ROLL** *5/10 - Costly Success (yes, but... with clear drawback)* [continue] (AI might narrate: The innkeeper agrees, but insists you work in the kitchens at dawn to pay off the debt, and makes it clear this arrangement won't be repeated.) <div class="cro-hr"></div> <h3 class="cro-h3">Ex. 9 - All at once - The Abandoned Camp Scenario</h3> [code] *I try to figure out what happened at this abandoned campsite.* **MENTAL ROLL** *6/10 - Pattern Recognition (ex. connections, similarities, recurring themes, etc.)* [continue] (AI narrates: Looking at the scattered belongings and overturned pots, you start recognizing a pattern. This looks similar to the aftermath of those bandit raids you've heard about - hasty departure, but valuables left behind suggests they fled rather than were robbed...) [code] *I examine the campsite more carefully for details.* **PERCEPTION ROLL** *7/10 - Hidden Elements (ex. concealed aspects, overlooked details, etc.)* [continue] (AI continues: Most of the obvious signs are clear, but something catches your eye. One of the tent stakes has been driven in at an odd angle, and beneath it, you spot the corner of something buried...) [code] *I try to force the buried object out with my hands.* **PHYSICAL ROLL** *4/10 - Rough Attempt (partial progress, noticeable strain)* [continue] (AI responds: You claw at the packed earth, making some progress but your fingernails are getting torn and your hands cramped. You've exposed what looks like a small wooden chest, but it's still firmly stuck...) [code] *Instead, I look around the camp for a tool to help dig it out properly.* **MANUAL ROLL** *7/10 - Competent Work (solid execution, minor rough edges)* [continue] (AI narrates: Using a bent tent pole as a makeshift lever, you carefully work the chest free. It comes up cleanly, though the wood shows some scrape marks from your improvised tool...) [code] *(I want to examine this chest closely before opening it)* *I take a moment to study the chest's construction and condition.* **ITEM FOCUS ROLL** *3/10 - Functional Design (ex. intended purpose, how it works, efficiency, etc.)* [continue] (AI responds: This isn't just storage - it's specifically designed as a strongbox. The iron reinforcements and the particular lock mechanism suggest it was made to protect something valuable during travel...) [code] *I check my pack to see if I have anything useful for dealing with this lock.* **INVENTORY CHECK** *8/10 - Valuable Discovery (significantly useful items)* [continue] (AI continues: Among your basic gear, you find a small set of lockpicks you'd forgotten about, wrapped in oiled cloth. Perfect for this type of mechanism...) [code] *The scene feels too quiet. I want to shift the energy here.* **NARRATIVE MOMENTUM ROLL** *4/10 - Tense Undercurrent (ex. brewing conflict, unspoken tension, etc.)* [continue] (AI responds: As you work on the lock, you become aware of being watched. The forest around the clearing feels too still, and you catch the faint sound of a branch creaking as if someone just shifted their weight...) [code] *I call out: "I know you're there. No point hiding." I try to sound confident rather than threatening.* **SOCIAL ROLL** *5/10 - Costly Success (yes, but... with clear drawback)* [continue] (AI concludes: A figure emerges from the treeline - a young woman with a crossbow. "That's my family's chest," she says. "But I suppose since you dug it up, we'll need to negotiate..." She's willing to talk, but clearly sees you as a potential threat to be managed rather than trusted...) <div class="cro-passage-nav"> <div><span class="nav-hub">[[Quick Reference Hub]]</span></div> <div>Also found in: <span class="nav-next">[[Advanced Tools & Examples->CRO Advanced Tools & Examples]]</span></div> </div>

<div class="cro-detail-top-nav"> <div class="nav-hub">[[Quick Reference Hub]]</div> </div> <h1 class="cro-h1">Notes & Best Practices (Detail)</h1> <h2 class="cro-h2">13.0 Notes & Best Practices</h2> The mechanics are only half the story - how you use them matters just as much. <h3 class="cro-h3">Children should not learn math</h3> Apply difficulty modifiers, advantage, stakes, and other adjustments before presenting your final roll to the AI. Showing your calculations can unconsciously bias the AI's interpretation. A clean "7/10" result tells a clearer story than "6+1 for advantage = 7/10." The AI should responds to the narrative weight of the number, not the arithmetic that produced it. <h3 class="cro-h3">You ”attempt” to drink water</h3> Your action descriptions should emphasize trying rather than succeeding. "I attempt to..." keeps the AI focused on interpreting your roll rather than assuming success. This small linguistic choice reinforces that outcomes depend on dice, not declarations. <h3 class="cro-h3">Blood is narrative fuel, and you the hell-engine of decent writing</h3> Low rolls aren't punishment - they're story engines. Some of my most memorable sessions have emerged from spectacular failures that forced creative solutions. Let the AI explore consequences rather than trying to minimize or rationalize poor results. The system works best when you trust dramatic setbacks to create interesting problems. <h3 class="cro-h3">Muscles, brains, or spirit of determination?</h3> Even when using the Main Outcome Scale, declaring your skill type helps the AI frame the action appropriately. "Physical Roll" versus "Social Roll" gives context for how that "6/10 - Weak Success" should manifest, even with identical outcome language. <h3 class="cro-h3">Extensionally restrain yourself</h3> Introduce optional extensions gradually. The core system already provides plenty of narrative variety, and too many simultaneous mechanics can overwhelm both you and the AI. Pick extensions that solve specific problems you're actually encountering, not ones that seem theoretically interesting. Unless your goal is to experiment! Then go right ahead. I’m not going to stop you. <h3 class="cro-h3">Artificial early onset dementia</h3> AI memory has limits, sometimes very strict ones at that, so focus on immediate action and outcome rather than expecting long-term state tracking. The system's simplicity stems from this limitation, yes, but that’s also a feature: we are working with the current LLM’s capabilities rather than against them. Keep complex details in your own notes if they matter for future sessions. <div class="cro-passage-nav"> <div><span class="nav-hub">[[Quick Reference Hub]]</span></div> <div>Also found in: <span class="nav-next">[[Advanced Tools & Examples->CRO Advanced Tools & Examples]]</span></div> </div>

<div class="cro-detail-top-nav"> <div class="nav-hub">[[Quick Reference Hub]]</div> </div> <h1 class="cro-h1">Basic Action & Rolling (Detail)</h1> <h2 class="cro-h2">2.0 Basic Action & Rolling</h2> <h3 class="cro-h3">2.1 Declaring Intent & Skill Category</h3> 1. Player declares action and skill category. 2. Player rolls d10, applies mental modifiers, presents final <span class="cro-inline-code">[Result]/10</span>. 3. AI/DM uses the corresponding Threshold Scale below for outcome interpretation. <h3 class="cro-h3">Format</h3> [code] *I attempt to [action].* **[SKILL] ROLL** *[Result]/10 - [Outcome Description]* [continue] <h3 class="cro-h3">Example</h3> [code] *I attempt to bake a chocolate soufflé.* **CULINARY ROLL** *8/10 - Clear Success (achieves goal)* [continue] <div class="cro-passage-nav"> <div><span class="nav-hub">[[Quick Reference Hub]]</span></div> <div>Also found in: <span class="nav-next">[[Core Mechanics->CRO Core Mechanics]]</span></div> </div>

<div class="cro-detail-top-nav"> <div class="nav-hub">[[Quick Reference Hub]]</div> </div> <h1 class="cro-h1">Difficulty Modifiers (Detail)</h1> <h3 class="cro-h3">4.1 Difficulty Modifiers</h3> Apply mental modifiers to the final roll result. Keep in mind that you do not need to show them to the AI, just adjust the final number when you present it: [code] - Trivial: +3 - Easy: +1 - Moderate: +0 - Difficult: -1 - Formidable: -3 [continue] For example, if you rolled a 6 on a Difficult task (-1), present it as <span class="cro-inline-code">5/10</span>. <div class="cro-passage-nav"> <div><span class="nav-hub">[[Quick Reference Hub]]</span></div> <div>Also found in: <span class="nav-next">[[Core Mechanics->CRO Core Mechanics]]</span></div> </div>

<div class="cro-detail-top-nav"> <div class="nav-hub">[[Quick Reference Hub]]</div> </div> <h1 class="cro-h1">Narrative Consequences (Detail)</h1> <h3 class="cro-h3">4.2 Narrative Consequences</h3> * A recent success in a similar area could grant you a //purely// mental +2 or Advantage (roll twice, keep higher). * A recent failure could impose a //purely// mental -2 or Disadvantage (roll twice, keep lower). You simply add or subtract from your final presentation. Don’t show the math to the AI nor, in the case of advantage/disadvantage, inform it that you’ve used advantage/disadvantage - just the final number you end up. <div class="cro-passage-nav"> <div><span class="nav-hub">[[Quick Reference Hub]]</span></div> <div>Also found in: <span class="nav-next">[[Core Mechanics->CRO Core Mechanics]]</span></div> </div>

<div class="cro-detail-top-nav"> <div class="nav-hub">[[Quick Reference Hub]]</div> </div> <h1 class="cro-h1">Time: Simple Declaration (Detail)</h1> <h3 class="cro-h3">5.1 Method A: Simple Declaration</h3> Allows the player to specify an intended duration for a time skip. In this specific case, the player doesn’t necessarily have to roll anything. [code] **[approximate duration, e.g., 'DAYS', 'WEEKS', 'MONTHS'] LATER.** [continue] <div class="cro-passage-nav"> <div><span class="nav-hub">[[Quick Reference Hub]]</span></div> <div>Also found in: <span class="nav-next">[[Time & Extensions Overview->CRO Time & Extensions Overview]]</span></div> </div>

<div class="cro-detail-top-nav"> <div class="nav-hub">[[Quick Reference Hub]]</div> </div> <h1 class="cro-h1">Time: Random Duration (Detail)</h1> <h3 class="cro-h3">5.2 Method B: Random Duration</h3> This method is useful when you want to jump forward an indeterminate amount of time, letting chance dictate how long passes before the narrative picks up again. Simply roll a d10 and consult the scale below to see the duration of the time skip. <h4 class="cro-h4">Format</h4> [code] **TIME ROLL** *[Result]/10 - [Time Jump]* **[UNIT OF TIME] LATER** [continue] <h4 class="cro-h4">Random Duration Scale</h4> [code] *1-3: Days later* *4-6: Weeks later* *7-8: Months later* *9-10: Years later* [continue] <div class="cro-passage-nav"> <div><span class="nav-hub">[[Quick Reference Hub]]</span></div> <div>Also found in: <span class="nav-next">[[Time & Extensions Overview->CRO Time & Extensions Overview]]</span></div> </div>

<div class="cro-detail-top-nav"> <div class="nav-hub">[[Quick Reference Hub]]</div> </div> <h1 class="cro-h1">Time: Time & Fortune (Detail)</h1> <h3 class="cro-h3">5.3 Method C: Time & Fortune</h3> Allows the player to describe the character of a chosen, or random, span of time, providing narrative flavour without prejudging outcomes or significant events. Offers an alternative to the basic {tooltip: 'TIME ROLL', text: 'See Method B: Random Duration'} for a more directed pacing (a similar deal to the Skill-Specific scales), allowing the player to choose to characterize the period as broadly favourable or unfavourable. <h4 class="cro-h4">Format</h4> [code] **[UNIT OF TIME] LATER** **TIME ROLL** *[Result]/10 - [Time & Fortune Scale Outcome]* [continue] <h4 class="cro-h4">Time & Fortune Scale</h4> [code] *1/10 - Deep Troubles (Time passed under notably difficult background conditions)* [continue] [code] *2/10 - Mostly Downs (The period felt marked by a generally negative trend)* [continue] [code] *3/10 - Problems Mount (Minor issues seemed to consistently arise or worsen)* [continue] [code] *4/10 - Felt Resisted (The period generally felt resistant or effortful)* [continue] [code] *5/10 - Slight Drag (Progress or stability felt subtly held back during this time)* [continue] [code] *6/10 - Slight Boost (Progress or stability felt subtly helped along during this time)* [continue] [code] *7/10 - Easy Flow (The period generally felt smooth or cooperative)* [continue] [code] *8/10 - Things Align (Minor opportunities seemed to consistently arise or improve)* [continue] [code] *9/10 - Mostly Ups (The period felt marked by a generally positive trend)* [continue] [code] *10/10 - Great Fortune (Time passed under notably favourable background conditions)* [continue] Characterizing entire time periods required a full 10-outcome range to capture the subtlety between "slightly difficult" and "persistently troubled" periods. Unlike moment-to-moment fortune rolls, this scale affects background conditions over extended periods, so the extra granularity helps the AI calibrate how that time period should feel, or how it would’ve felt, without mandating specific events. The scale shapes atmosphere, and so the plot, but indirectly. <div class="cro-passage-nav"> <div><span class="nav-hub">[[Quick Reference Hub]]</span></div> <div>Also found in: <span class="nav-next">[[Time & Extensions Overview->CRO Time & Extensions Overview]]</span></div> </div>

<div class="cro-detail-top-nav"> <div class="nav-hub">[[Quick Reference Hub]]</div> </div> <h1 class="cro-h1">Where to Use CRO (Detail)</h1> <h2 class="cro-h2">1.1 Where to use</h2> CRO works with any AI that can engage in conversational roleplay, and requires absolutely no setup on the AI's side. No special system prompts, no configuration - you simply start using the format and the AI will naturally adapt to it. If you’re in need of a d10 dice, just search “Dice Roller” into google, there’s an easter egg for it. I've primarily developed and tested this system on **character roleplay platforms**: first on Character.ai (that was like a year ago, when it was still kind of decent!), then Spicychat.ai and Janitor.com. These platforms are designed around chatting with specific AI personas in ongoing conversations, which tends to work particularly well with CRO's narrative focus. There are also some good world-based “characters" there too: chatbots designed to create a world for your own character to work and act inside of. You'll find plenty of them to experiment with. There’s also **general AI assistants** like ChatGPT, Claude, Gemini, or Deepseek, of course - those also work just fine. You can start a conversation by describing your character and setting, then begin using the CRO format for actions. These platforms often bring more sophisticated reasoning to the table, though they may need a sentence or two of context about what you're doing initially. Point is, you're not changing how the AI works - you're changing how you present information to it. The format itself teaches the AI what you want: that dice results should guide outcomes, that low rolls mean complications, and that high rolls mean smooth success. Most AI systems pick this up within a few exchanges, if not right away. I personally lean towards character platforms because they're built and fine-tuned for sustained narrative interaction, but honestly, experimentation is part of the fun. Try it wherever you normally interact with LLMs. See how it feels! <div class="cro-passage-nav"> <div><span class="nav-hub">[[Quick Reference Hub]]</span></div> <div>Also found in: <span class="nav-next">[[Getting Started->Getting Started]]</span></div> </div>

<div class="cro-detail-top-nav"> <div class="nav-hub">[[Quick Reference Hub]]</div> </div> <h1 class="cro-h1">Judge AI - Base System Prompt (Detail)</h1> <h2 class="cro-h2">11.3 Judge AI - Base System Prompt</h2> Copy and paste this exact prompt when setting up your Judge AI (this is the minimum required). [code] # CRO Judge AI System Prompt (Base) You are the **Judge AI** for the {Char}, [Roll] for [Outcome] (CRO) system. Your role is to make impartial mechanical rulings for player actions. You do NOT create narrative responses - only mechanical judgments. ## Your Process 1. **Receive**: Player's action declaration and their d10 roll result 2. **Analyze**: Assess difficulty and determine appropriate skill category 3. **Calculate**: Apply difficulty modifier to the player's roll 4. **Format**: Return standardized CRO result ## Difficulty Assessment **Base Difficulties:** - **Trivial (+3)**: Almost guaranteed success (opening unlocked door, basic conversation) - **Easy (+1)**: Minor challenge with favorable conditions (climbing ladder, convincing willing person) - **Moderate (+0)**: Standard challenge requiring competence (basic lock, persuading neutral party) - **Difficult (-1)**: Significant challenge requiring skill/luck (advanced lock, hostile persuasion) - **Formidable (-3)**: Extreme challenge, failure likely (master locks, impossible feats) **Contextual Modifiers (+/-1 to +/-2):** - **Advantages**: Ideal tools, favorable conditions, recent success, cooperation - **Disadvantages**: Poor tools, hostile conditions, recent failure, opposition, stress, impairment ## Skill Categories Assign based on primary method: - **PHYSICAL**: Strength, athletics, endurance, combat - **MANUAL**: Dexterity, crafting, fine motor skills - **MENTAL**: Intelligence, knowledge, problem-solving - **SOCIAL**: Persuasion, deception, intimidation - **PERCEPTION**: Awareness, investigation, searching - **FORTUNE**: Pure luck, timing, circumstance - **STEALTH**: Sneaking, hiding, silent movement ## Output Format **REASONING** ``` [2-3 sentences explaining contextual factors affecting difficulty: conditions, tools, character state, environmental factors, etc.]‎ ``` **DIFFICULTY ASSESSMENT**: [Trivial/Easy/Moderate/Difficult/Formidable] **SKILL CATEGORY**: [Chosen skill] **PLAYER'S ROLL**: [Player's d10 result] **FINAL RESULT**: [Result after applying difficulty modifier] --- **FORMATTED ACTION** ``` *I attempt to [restate player's action].* **[SKILL] ROLL** *[Final Result]/10 - [Appropriate outcome from Main Outcome Scale]* ``` ## Key Principles 1. **Be Impartial**: Assess objectively, don't favor success or failure 2. **Use Context**: Consider described situation without adding details 3. **Default Conservative**: When uncertain, choose the harder difficulty 4. **Be Consistent**: Similar actions in similar contexts get similar ratings **Remember**: You make mechanical rulings only. The Narrative AI handles all story elements based on your result. [continue] <div class="cro-passage-nav"> <div><span class="nav-hub">[[Quick Reference Hub]]</span></div> <div>Part of: <span class="nav-next">[[Advanced Tools & Examples->CRO Advanced Tools & Examples]]</span> (Judge System)</div> <div>Next Module: [[Judge AI - Scales Module->CRO Judge AI Scales Module Detail]]</div> </div>

<div class="cro-detail-top-nav"> <div class="nav-hub">[[Quick Reference Hub]]</div> </div> <h1 class="cro-h1">Judge AI - Scales Module (Detail)</h1> <h2 class="cro-h2">11.4 Judge AI - Scales Module</h2> Add this to the Base Prompt for skill-specific scale usage. [code] # CRO Judge AI Scales Module **Add this to Base Prompt for skill-specific scale usage** ## Scale Selection Use skill-specific scales when the action strongly emphasizes the particular nature of that skill domain. Default to Main Outcome Scale for general actions. **When to Use Specific Scales:** - **Physical Scale**: Actions where body response and physical consequences matter (climbing, fighting, endurance challenges) - **Manual Scale**: Tasks where execution quality is key (lockpicking, crafting, delicate work) - **Mental Scale**: Problem-solving where cognitive approach matters more than information discovery - **Social Scale**: Interpersonal dynamics where relationship momentum is the focus - **Perception Scale**: Investigation where the type of observation matters (not just "how much" you notice) - **Fortune Scale**: Pure chance events with no skill component - **Stealth Scale**: Concealment actions where detection levels are crucial ## Modified Output Format When using a skill-specific scale, replace the final formatted action with: ``` *I attempt to [restate player's action].* **[SKILL] ROLL** *[Final Result]/10 - [Appropriate outcome from [SCALE NAME] Scale]* ``` **Example**: `*7/10 - Pattern Recognition (ex. connections, similarities, recurring themes, etc.)*` for Mental Scale ## Scale Selection Note Include an additional sentence or two in your reasoning for which scale you're using and why: ``` Using [Scale Name] because [brief reason focusing on the specific nature of the skill being employed]. ``` [continue] <div class="cro-passage-nav"> <div><span class="nav-prev">[[Previous Module: Base Prompt->CRO Judge AI Base Prompt Detail]]</span></div> <div><span class="nav-hub">[[Quick Reference Hub]]</span></div> <div>Part of: <span class="nav-next">[[Advanced Tools & Examples->CRO Advanced Tools & Examples]]</span> (Judge System)</div> <div>Next Module: [[Judge AI - Extensions Module->CRO Judge AI Extensions Module Detail]]</div> </div>

<div class="cro-detail-top-nav"> <div class="nav-hub">[[Quick Reference Hub]]</div> </div> <h1 class="cro-h1">Judge AI - Extensions Module (Detail)</h1> <h2 class="cro-h2">11.5 Judge AI - Extensions Module</h2> Add this to the Base Prompt for extension awareness and suggestions. [code] # CRO Judge AI Extensions Module **Add this to Base Prompt for extension awareness and suggestions** ## Extension Recognition When player actions or context suggest enhanced mechanics, automatically apply the extensions below. Note the relevant ones in your reasoning and modify the final format accordingly. ## Available Extensions **Action Modifiers:** - **Stakes**: High-risk/high-reward situations where consequences matter more than difficulty - **Pushing Limits**: Attempting actions far beyond normal capability - **Skill Leverage**: Using multiple skills in combination - **Targeted Effort**: Aiming for precise outcomes rather than general success - **Guarded Approach**: Avoiding critical failure at cost of critical success - **Serendipity**: Converting mixed results (4-6) into additional opportunities **Investigation & Discovery:** - **Item Focus**: Detailed examination of specific objects - **Inventory Check**: Searching personal effects for useful items **Environmental & Narrative:** - **Random Events**: External circumstances beyond character control - **Emotional Static**: Surreal manifestations of strong emotions - **Interference/Nemesis**: Active opposition to character goals - **Montage**: Multiple actions over extended time periods - **Narrative Momentum**: Shifting scene energy and tone ## Extension Application When relevant, apply extensions automatically and note in your reasoning: ``` Applying [Extension Name]: [Brief explanation and how it modifies the result]. ``` ## Modified Output Extensions may change your final result or scale used. Show the modified outcome in your formatted action, and explain any changes in your reasoning. [continue] <div class="cro-passage-nav"> <div><span class="nav-prev">[[Previous Module: Scales Module->CRO Judge AI Scales Module Detail]]</span></div> <div><span class="nav-hub">[[Quick Reference Hub]]</span></div> <div>Part of: <span class="nav-next">[[Advanced Tools & Examples->CRO Advanced Tools & Examples]]</span> (Judge System)</div> </div>

<div class="cro-detail-top-nav"> <div class="nav-hub">[[Quick Reference Hub]]</div> </div> <h1 class="cro-h1">CRO Helper Browser Script</h1> The CRO Helper is a browser script that adds a convenient graphical interface directly to a few supported chat platforms. Instead of manually formatting rolls or copying text, you get a floating dice button that opens a CRO toolkit right on the page. It doesn't contain all of the guide's extensions and features *just yet*, but it's pretty handy. <h2 class="cro-h2">Supported Platforms</h2> Currently, the script works on: - **Spicychat.ai** (I've been told it's been tested more extensively here!) - **Character.ai** - **JanitorAI.com** <h2 class="cro-h2"> Current Features</h2> **Action Roller Screen:** Complete CRO generator with dialogue input, action descriptions, skill selection, difficulty modifiers, stakes, pushing limits, and all the outcome scales from this guide. **Quick Roll Screens:** Dedicated tools for Inventory Checks and Perception rolls with one-click generation. **Quality of life:** Automatically formats output, copies to clipboard, or sends directly to chat. Includes dice roll animations and multiple visual themes. <h2 class="cro-h2">Installation</h2> **Step 1: Install Tampermonkey** First, you'll need a userscript manager, like Tampermonkey: - **Chrome:** Install from the [Chrome Web Store](https://chrome.google.com/webstore/detail/tampermonkey/dhdgffkkebhmkfjojejmpbldmpobfkfo) - **Firefox:** Install from [Firefox Add-ons](https://addons.mozilla.org/firefox/addon/tampermonkey/) - **Other browsers:** Visit [tampermonkey.net](https://www.tampermonkey.net/) for download links **Step 2: Install the CRO Helper Script** Once Tampermonkey is installed, visit the script page on Greasyfork: - **[Install CRO Helper Script](https://greasyfork.org/en/scripts/537525-spicychat-cro-helper)** You'll have to login if you haven't already, though you can do so with your google account, it's not a long process. Click the green "Install this script" button. Tampermonkey will open and show you the script details - click "Install" to confirm. <h2 class="cro-h2">Basic Usage</h2> After installation, visit any of the supported chat platform. You'll see a small dice button in the bottom-left corner of your browser window. Once there, click the dice button to open the CRO Helper popover. Use the toolbar on the left to switch between different tools. You can also drag that same button to reposition it anywhere you like, and resize the window. The Action Roller screen works just like the examples in this guide - fill in your action description (or dialogue!), select a skill, adjust difficulty if needed, and hit "Generate Action Roll." The formatted output can be copied or sent directly to chat. <h2 class="cro-h2">Troubleshooting</h2> If the script doesn't appear, make sure Tampermonkey is enabled and refresh the page. You have to be on a chat, and it might take up to two refreshings at first launch, but once it appears you should be good to go. <div class="cro-passage-nav"> <div><span class="nav-hub">[[Quick Reference Hub]]</span></div> <div>Also found in: <span class="nav-next">[[Getting Started->Getting Started]]</span></div> </div>