Formulation Review #1: Beyond the 20 Rules 🔗 **Source:** [Formulation Review #1: Beyond the 20 Rules](https://www.youtube.com/watch?v=_FQNjDnz94o) > [!note]+ Document Notes > tags: [spaced-repetition, knowledge-formulation, memory-techniques] > author: Golden Nuggets Podcast > source: YouTube > reference: https://www.youtube.com/watch?v=_FQNjDnz94o > date created: 2024-12-19 10:30 > > # 🏷️ Title: Formulation Review #1: Beyond the 20 Rules > <!-- deep-read:sections start --> ## 🧰 Action Inbox - [H1] Review your leeches and remove items that consistently fail after 4+ repetitions — podcast examples show 80/20 rule #srs-maintenance - [H2] Use "might" in tentative knowledge items to unlock new formulation possibilities — demonstrated high performance #item-formulation - [H3] Create "name at least one" list items instead of ambiguous recall prompts — removes ambiguity from reviews #list-memorization ## 🧭 Table of Contents - [📝 Summary](#-summary) - [📖 Main Content](#-main-content) - [🗒️ Notes](#️-notes) - [📚 References](#-references) - [🗣️ Paraphrase](#️-paraphrase) - [📰 Newsletter](#-newsletter) - [✅ Action Items](#-action-items) - [🧪 Hypotheses](#-hypotheses) - [📊 Methodology](#-methodology) - [❓ Questions](#-questions) - [🔮 Future Research](#-future-research) - [💡 Implications](#-implications) - [🧠 Learning Styles Assessment](#-learning-styles-assessment) - [📊 Accessibility Assessment](#-accessibility-assessment) - [🌍 DEIB Assessment](#-deib-assessment) - [✅ Logical Validity](#-logical-validity) - [🛡️ DARVO Analysis](#-darvo-analysis) - [🎯 Conclusion](#-conclusion) ### 🔗 References - [The 20 Rules of Formulating Knowledge](http://super-memory.com/articles/20rules.htm) - [Andy Matuschak's Orbit Project](https://withorbit.com/) ## 📝 Summary - Two spaced repetition experts review actual SRS items to demonstrate good vs. poor formulation techniques - They extend Piotr Wozniak's famous 20 Rules with modern insights about applicability, interference, and meaningful goals - Key issues identified: overly specific numbers, interference between similar items, syntactic vs. semantic learning - Successful patterns shown: analogies, tentative knowledge with "might," list recall with "name at least one" - Bottom line: Focus on applicable knowledge with clear goals rather than meaningless trivia ## 📖 Main Content This podcast-style video reviews specific spaced repetition system (SRS) items to illustrate principles of effective knowledge formulation. The hosts examine both failed and successful items from their collections and Andy Matuschak's Orbit project. **Core Problem**: Most beginners struggle with item formulation for 2-3 years, leading to high failure rates that make the learning process painful and joyless. Even worse, some items appear successful (high retention) but provide no real-world value due to syntactic rather than semantic encoding. **Key Insights Beyond Wozniak's 20 Rules**: - Numbers are asymmetrically difficult to remember; consider decades instead of specific years - Interference occurs when similar items confuse each other; resolve with distinguishing facts - Analogies perform exceptionally well and activate at appropriate times in real life - Tentative knowledge can be encoded effectively using qualifying words like "might" - List items work better with explicit requirements like "name at least one" The discussion emphasizes that item formulation is both art and science, requiring extensive practice with hundreds of examples before mastery develops. ## 🗒️ Notes #### ⏱️ [00:00](https://www.youtube.com/watch?v=_FQNjDnz94o?t=0) — [H1] Introduction to Beyond the 20 Rules - Key idea: Extending Wozniak's 1999 formulation rules with modern insights and real examples - Evidence: "We're going to use Piotr Wozniak's very famous article as a sort of backbone and add some of our own extensions" #### ⏱️ [03:42](https://www.youtube.com/watch?v=_FQNjDnz94o?t=222) — [H2] The Nature of Formulation Problems - Key idea: Poor formulation creates a vicious cycle of failure and frustration that kills motivation - Evidence: "If you don't have good item formulation you will fail a lot of cards... enduring that for a long period of time will just cause the pleasure of learning to decrease rapidly" #### ⏱️ [05:15](https://www.youtube.com/watch?v=_FQNjDnz94o?t=315) — [H3] Syntactic vs Semantic Learning Problem - Key idea: Items can have high retention but zero real-world value due to pattern matching - Evidence: "You're sort of memorizing the shape or structure of it... when real life asks that similar problem in different words you can't come up with the answer" #### ⏱️ [07:30](https://www.youtube.com/watch?v=_FQNjDnz94o?t=450) — [H1] Example 1: Overly Specific Dates - Key idea: Exact dates are harder to remember and less valuable than decades or centuries - Evidence: Newton Principia item failed repeatedly - "there's something very asymmetric something that just causes numbers not to stick to your mind" #### ⏱️ [11:45](https://www.youtube.com/watch?v=_FQNjDnz94o?t=705) — [H2] Critical Mass of Historical Knowledge - Key idea: Once you reach sufficient context, dates slot in naturally due to surrounding knowledge - Evidence: "Once they'd reached a certain critical mass of knowledge in European history that it entirely changed their view on the subject" #### ⏱️ [14:20](https://www.youtube.com/watch?v=_FQNjDnz94o?t=860) — [H3] The Leech Problem and 80/20 Rule - Key idea: Failed items dominate your time and ruin the learning experience - Evidence: "You spend eighty percent of your time on twenty percent of your items and those items are the leeches" #### ⏱️ [16:10](https://www.youtube.com/watch?v=_FQNjDnz94o?t=970) — [H1] Example 2: Interference Between Similar Items - Key idea: Hamilton and Trivers items interfere because both are evolutionary biologists with overlapping work - Evidence: "I can't even really say for certain which one of them came up with which theory... it just creates so much interference" #### ⏱️ [18:45](https://www.youtube.com/watch?v=_FQNjDnz94o?t=1125) — [H2] Resolving Interference with Distinguishing Facts - Key idea: Find unambiguous facts that make confusion impossible - Evidence: John Adams/Thomas Jefferson example - learning Adams said "Thomas Jefferson lives" but was wrong eliminated all interference #### ⏱️ [22:30](https://www.youtube.com/watch?v=_FQNjDnz94o?t=1350) — [H1] Example 3: Successful Analogy Item - Key idea: Analogies perform exceptionally well and activate appropriately in real contexts - Evidence: Russell/self-reference item - "no lapses" and "I don't think I have any lapses on any of the analogies" #### ⏱️ [24:15](https://www.youtube.com/watch?v=_FQNjDnz94o?t=1455) — [H2] Spreading Activation from Analogies - Key idea: Analogies create semantic connections that trigger at appropriate times - Evidence: "By expanding that concept... this increase in spreading activation to semantically similar things" #### ⏱️ [26:40](https://www.youtube.com/watch?v=_FQNjDnz94o?t=1600) — [H1] Pattern 1: List Items with "Name At Least One" - Key idea: Explicit requirements remove ambiguity about expected recall level - Evidence: Sleep reduction item - "Name at least one" prevents confusion about partial vs complete recall #### ⏱️ [29:20](https://www.youtube.com/watch?v=_FQNjDnz94o?t=1760) — [H2] Pattern 2: Tentative Knowledge with "Might" - Key idea: Using "might" opens possibilities for encoding uncertain but useful knowledge - Evidence: Intuitive judgments item - "the word might is very very useful... it allows you to kind of frame things tentatively" #### ⏱️ [32:15](https://www.youtube.com/watch?v=_FQNjDnz94o?t=1935) — [H3] Performance Data on "Might" Items - Key idea: Tentative knowledge items show excellent performance metrics - Evidence: "I only had one lapse across those items I think you had none at all" #### ⏱️ [34:45](https://www.youtube.com/watch?v=_FQNjDnz94o?t=2085) — [H1] Pattern 3: Good Use of Cloze Deletion - Key idea: Cloze works when it connects phrasing to meaning in irreducible ways - Evidence: Andy's interpretation item - "it matches the phrasing to the meaning... removes all the ambiguity" #### ⏱️ [37:30](https://www.youtube.com/watch?v=_FQNjDnz94o?t=2250) — [H2] Pattern 4: Example Items for Concept Integration - Key idea: Examples act as analogies that expand and integrate conceptual understanding - Evidence: Focus prompt example - "the example acts as an analogy and that analogy expands this concept" #### ⏱️ [39:50](https://www.youtube.com/watch?v=_FQNjDnz94o?t=2390) — [H3] Pattern 5: Generation/Application Items - Key idea: Creating new examples activates all related knowledge and builds capability - Evidence: Application prompt discussion - "you're theoretically you would come up with a new one or slight variation every time" #### ⏱️ [43:20](https://www.youtube.com/watch?v=_FQNjDnz94o?t=2600) — [H1] Why Complex Problems Don't Work in SRS - Key idea: Multi-step problems break SRS assumptions about atomic knowledge units - Evidence: "How does the algorithm know what aspect of it you forgot... you may remember nine out of ten components" #### ⏱️ [46:10](https://www.youtube.com/watch?v=_FQNjDnz94o?t=2770) — [H2] The Importance of Applicability and Meaningful Goals - Key idea: Without clear applications, trivia accumulation leads to morale collapse - Evidence: "You fill your brain with a bunch of disparate information... it doesn't really add up to anything significant" #### ⏱️ [48:45](https://www.youtube.com/watch?v=_FQNjDnz94o?t=2925) — [H3] Knowledge Valuation Through Historical Models - Key idea: Study successful people with similar goals to identify valuable knowledge domains - Evidence: "Look at over time in human history are there any people who have similar goals to you what did they do to acquire knowledge" ## 📚 References - [Piotr Wozniak's 20 Rules of Formulating Knowledge](http://super-memory.com/articles/20rules.htm) - [Andy Matuschak's Orbit Project](https://withorbit.com/) - [Douglas Hofstadter: Analogy as the Core of Cognition](https://www.youtube.com/watch?v=n8m7lFQ3njk) - [Golden Nuggets Podcast Episode 10](https://www.youtube.com/channel/UCQbXA4nUjEqxW_iyM6K8-bw) ## 🗣️ Paraphrase - Most people struggle with creating effective spaced repetition items for 2-3 years, leading to frustration and abandonment of the system - The key insight is that items can appear successful (high retention rates) while providing zero real-world value due to syntactic rather than semantic learning - Successful patterns include using analogies, tentative language like "might," explicit list requirements, and focusing on applicable knowledge rather than trivia ## 📰 Newsletter **Beyond the Basics: Advanced SRS Item Formulation** Ever wondered why your Anki reviews feel like a slog? Two spaced repetition experts just shared game-changing insights about what makes items actually stick and transfer to real-world use. **Key breakthrough**: Items can have perfect retention rates while providing zero value. The culprit? You're memorizing syntax patterns instead of semantic meaning. When real life presents the same concept in different words, you draw a blank. **Practical wins you can implement today**: - Replace specific dates with decades (1680s vs 1687) - Use "name at least one" for list items instead of ambiguous recall - Try "might" to encode tentative but useful knowledge - Focus on analogies - they show exceptional performance The bigger picture: Stop chasing trivia knowledge. If you can't see clear applications for what you're learning, you're setting yourself up for morale collapse down the road. Resources: [20 Rules of Knowledge Formulation](http://super-memory.com/articles/20rules.htm), [Orbit Examples](https://withorbit.com/), [Learn WordPress](https://learn.wordpress.org/) ## ✅ Action Items #### DevRel - Developer advocacy: Create SRS formulation workshops for technical communities, demo effective patterns for API documentation - Education: Develop curriculum on encoding procedural knowledge vs declarative facts in technical domains - Developer experience: Apply "name at least one" pattern to error message documentation, CLI command references - Community management: Run item review sessions where developers share formulation challenges #### OSPO - Policy & Governance: Encode compliance requirements using tentative language patterns to handle evolving regulations - License Compliance: Create example-based items for identifying license compatibility scenarios - Developer Support: Train teams on formulating security knowledge that transfers to real vulnerability assessment - Strategic Alignment: Use generation prompts for creating new policy scenarios from existing frameworks #### Community - Maintainers: Provide formulation templates for onboarding knowledge that new contributors actually retain - Contributors: Create interference-free items for distinguishing similar tools, frameworks, or processes - Ecosystem: Develop analogies that connect abstract concepts to concrete implementation examples #### WordPress - Training: Redesign lesson materials using proven patterns like "might" for tentative best practices - Documentation: Apply cloze deletion principles to code examples where phrasing matches meaning - Community: Create item review sessions at WordCamps focused on applicable knowledge vs trivia - Design: Encode design principles using analogy patterns that activate during actual design work - CLI: Formulate command knowledge using generation prompts for real-world scenarios #### The WPCC - Programs: Design fellowship curricula that prioritize applicable formulation over comprehensive coverage - Projects: Create governance documentation using interference-resolution techniques for similar processes - Funding: Develop evaluation rubrics that distinguish syntactic compliance from semantic understanding #### FAIR - Technical independence: Encode provenance verification using example-based items that transfer to new contexts - Policies: Use tentative language patterns for evolving governance frameworks - Governance artifacts: Apply list formulation techniques to stakeholder identification and responsibilities ## 🧪 Hypotheses - Items using tentative language ("might") will show 20%+ better real-world transfer than absolute statements - Analogy-based items will demonstrate superior long-term retention compared to direct fact encoding - List items with explicit quantity requirements ("name at least two") will reduce review ambiguity by 40% - Generation/application items will activate related knowledge more frequently in practical contexts ### 📊 Methodology - A/B test formulation patterns across matched knowledge domains with transfer assessments - Longitudinal analysis of lapse rates comparing syntactic vs semantic item structures - Real-world activation tracking through user journals and context-triggered recall surveys ### ❓ Questions - What specific linguistic markers differentiate items that transfer well vs those that remain isolated? - How can we systematically identify and resolve interference between conceptually related items? - What's the optimal balance between comprehensive coverage and applicable focus for domain expertise? ### 🔮 Future Research - Automated detection of high-interference item pairs using semantic similarity analysis - Development of formulation pattern libraries for specific knowledge domains - Integration of real- <!-- deep-read:sections end -->