RunsheetKnowledge Modeling

Feynman Technique

ComplexityLow

Time20-60 min

Participants1

FormatAsync

MaturityEstablished

Preparation

What needs to be ready before start

Materials

Notebook or empty file; sources for the concept (book, paper, video); pen; timer; optional sparring partner to test the explanation (non-expert person).

People / roles

One person who wants to understand the concept; optionally a sparring person who listens and asks questions without background knowledge; no facilitator needed.

Pre-read

Concept as a single sentence; learning goal (what should I be able to explain afterwards, for whom, at what depth); available sources; time budget.

Time needed

20-60 min

Setup

Write concept at top of empty page. Keep sources available but collapsed. Attempt first explanation without looking at sources. Schedule a sparring partner if possible.

Core question

The one question this method answers

Do I understand the concept well enough to explain it clearly to someone without background knowledge, and which gaps become visible while explaining?

Flow

Marker: Phase

Step	Duration	Action	Hint
1Phase 1: Choose concept	5 min	Write concept precisely at top of page. Define audience (for example, "new colleague without background knowledge," "technically curious beginner," "Sabine, solo tax advisor").	Choosing concept too broadly is a pitfall. "Machine learning" is too broad; "How backpropagation works in a two-layer network" is manageable.
2Phase 2: Explain simply first time	10-15 min	Write explanation in plain language as if speaking to the target audience. Avoid jargon without explanation. Use everyday analogies.	Using jargon without clarification hides understanding gaps. Either explain each technical term or replace it with accessible wording.
3Phase 3: Mark gaps	5-10 min	Review explanation critically. Where is it vague, evasive, or textbook-like? Mark those spots. For each spot create a question that remains open.	Vague wording such as "in some way" or "something happens somehow" are gap signals. Mark these as red and form a question.
4Phase 4: Verify with sources and close gaps	10-20 min	Open sources and answer marked questions directly. Translate answers into own wording, do not copy. Add clarification to explanation at marked spots.	If sources do not close the gap, that is a valuable result. Either find more sources or narrow the concept.
5Phase 5: Simplify again and test	10-15 min	Rewrite explanation again, simpler and shorter. Ideally ask sparring partner for feedback from someone without prior knowledge. Record remaining gaps in Study Notes.	Sparring test is the strongest filter. Without sparring, the user risks self-confirmation.

Artifact

What comes out at the end

Form

Markdown note with concept, target audience, final plain-language explanation, explicit gap list, and list of next learning goals or sources.

Versioning / ownership

One note per concept with date. On re-read, add update date at top. Reapplying the technique in sequence builds depth; retain old explanations.

Tool alternatives

Notion page per concept
Obsidian with backlinks to related concepts
Markdown in learning repo
Handwritten notebook

feynman-technique-working-template.md

Compact working template for Feynman Technique with context, input, output artifacts, and next step.

Feynman Technique Working Template

Goal

Checks understanding by explaining a concept in simple terms and closing knowledge gaps deliberately.

Context

When and for what do we use this method?

Input

Which data, observations, decisions, or materials are available?

Execution

Short notes along the runsheet.

Output artifacts

Plain-language Explanation:
Knowledge Gaps:
Study Notes:

Assumptions and open questions

Decision / Next step

Owner, date, and success signal.

Example output

Concrete filled scenario, fictional example

feynman-technique-beispiel.md

Concrete filled scenario, fictional example

Feynman — How connection pooling works in PostgreSQL (2026-05-15)

Target audience: New backend engineer with web fundamentals, no PostgreSQL experience.

Second explanation: If your app wants to talk to the database, it first needs to establish a connection. Think of it like making a phone call: dial, ring, answer, conversation, hang up. With PostgreSQL, establishing connections takes comparatively long (50-100 ms), and the database can only handle a limited number of simultaneous calls (typically 100).

If each request creates a new call, it becomes slow and quickly overloads the database. Connection pooling means your app keeps 20-40 calls open and reassigns them to requests. Each request gets a short line, uses it, and returns it.

PostgreSQL has two modes: Session pooling (one request may hold a line for an entire session) and Transaction pooling (line is returned per transaction). Transaction pooling saves more lines but does not work with session features such as PREPARE.

Symptom

Learner appears clear because own knowledge fills gaps.

What to do

If possible, use a sparring person without prior knowledge. Online feedback via forum answer or blog draft is also possible.

Stop criteria

Done signals checkable in under a minute

Concept is too broad, explanation stays superficial or endless.

No sources are available, gaps cannot be closed.

Time budget under 15 min and depth is not achievable.

Method is used for fact lookup (date, number) and no learning goal exists.

No willingness to rewrite explanation, old notes are recycled.

Concept has no meaningful everyday analogy and math remains unexplained; use another learning method.

Finished the runsheet?

Go to the profile for purpose, similar methods, and sources or continue to the next method in the catalog.

Go to profile More methods