analytical-thinking · diff
v2026.5.17 to v2026.10.5
40 added, 54 removed. Audit A to A.
---
name: analytical-thinking
description: >
- Use this skill when the user asks for analytical thinking (including naming it
- or directing use/apply/run with obvious misspellings; decisive) or wants a
- structured breakdown—decomposing the problem, defining metrics and
- hypotheses, organizing evidence, and synthesizing findings with explicit
- uncertainty. Use for quant-style reasoning framing, root-cause trees, decision
- tables, or comparable structure, including informal or incomplete data asks.
- Skip when they want open-ended idea spray with no measurement or hypothesis
- angle, or a short verdict with no decomposition requested.
+ Use for analytical thinking, or when a question needs breaking down and evidence: "why did metric X drop?", "what drives this?", "break this down", "what data do we need?". Builds a tree, hypotheses and evidence, ends with an answer and key uncertainty. Skip for quick verdicts.
license: MIT
metadata:
author: ysskrishna
- version: "2026.5.17"
+ version: "2026.10.5"
---
# Analytical Thinking
- Clarity beats cleverness. End with answers tied to structure and stated confidence.
-
- **How to run it with this skill:** one clearly headed section per step in this order: Frame → Decompose → Hypotheses → Evidence → Synthesis. Insert **Options matrix** only when Setup calls for it (after Evidence, before Synthesis).
-
- ---
-
- ## Setup (run before starting)
-
- In one short block:
-
- 1. **Analytical question** — precise, ideally falsifiable
- 2. **Default pass** — Frame → Decompose → Hypotheses → Evidence → Synthesis (state this line)
-
- If data availability or definitions are missing, ask at most 3 questions in one message, then proceed. Note any remaining gaps or working guesses in plain language (no bracket tags in Setup).
-
- If the user is **choosing among concrete alternatives**, after **Evidence** insert **Options matrix**: rows = options, columns = criteria (state weights if any), qualitative scores (− / 0 / +) with one-line justification per cell — then finish with **Synthesis**.
-
- ---
-
- ## The Steps
+ Clarity beats cleverness. End with an answer tied to structure and a stated confidence.
- ### Frame
+ ## When to use
- **Question type** (estimate, compare, explain, predict, optimize). **Unit of analysis** and **baseline** (even if hypothetical).
+ - Explaining or predicting a number: a metric moved, a cost is high, a funnel leaks.
+ - A messy question that needs a driver tree, hypotheses and a plan for what data to collect.
+ - Comparing concrete alternatives against criteria when the user wants the reasoning laid out.
- ### Decompose
+ Skip: open-ended idea generation with nothing to measure, or a short verdict with no decomposition asked for.
- Tree or table: factors, drivers, or workstreams. Each child node should be **MECE-ish** (mutually exclusive where it matters; collectively exhaustive enough for the decision).
+ ## Before you start
- ### Hypotheses
+ 1. State in one block: **Analytical question** (precise, ideally falsifiable) and **Pass** (Frame, Decompose, Hypotheses, Evidence, Synthesis).
+ 2. **Gather first.** If you can query data, read dashboards, logs or code, do that before listing Evidence. Ask up to 3 questions only for definitions or data that tools cannot supply.
+ 3. **Light path.** Small ask: a 2-level tree, 2 hypotheses, one observation each, answer in 3 lines.
+ 4. If the user is choosing among concrete options, insert an **Options matrix** after Evidence: rows are options, columns are criteria (state any weights), cells are `-`, `0` or `+` with a one-line reason. Then Synthesis.
- Ranked **H1, H2, H3** — what would we expect to observe if each were true? What would **falsify** each?
+ ## Steps
- ### Evidence
+ ### 1. Frame
+ **Question type** (estimate, compare, explain, predict, optimize), **unit of analysis**, and **baseline** (even a hypothetical one).
- For each hypothesis: **Observation:** … — **Strength note:** one short sentence on how much this observation supports or undermines the hypothesis and the main limit (no Strong/Moderate/Weak labels). **Caveat:** …
+ ### 2. Decompose
+ A tree or table of drivers or workstreams. Aim for branches that do not overlap and together cover the question well enough for the decision.
- If no real data, run a **thought experiment** section instead — label bullets `[THEORETICAL]`.
+ ### 3. Hypotheses
+ Ranked H1, H2, H3. For each: what would we see if it were true, and what would **falsify** it?
- ### Synthesis
+ ### 4. Evidence
+ For each hypothesis: **Observation:** ... - **Strength:** one sentence on how much it supports or undermines the hypothesis and its main limit. **Caveat:** ...
- 1. **Answer** — direct response to the analytical question
- 2. **Key uncertainty** — what single unknown swings the answer most
- 3. **Next data / step** — what to collect or run next
+ No real data? Replace this step with a section titled **Thought experiment (no data)**. Inputs you guess are shown as ranges and marked `[ESTIMATED]`.
- ---
+ ### 5. Synthesis
+ 1. **Answer** to the analytical question.
+ 2. **Key uncertainty:** the one unknown that swings the answer most.
+ 3. **Next data or step:** what to collect or run next.
- ## Execution Rules
+ ## Pitfalls
- 1. Do not conflate **Hypotheses** and **Evidence** in the same bullet list.
- 2. Numbers: if inputs are guessed, show **ranges** and label `[ESTIMATED]`.
- 3. Prefer **structure** over long prose.
+ - Mixing Hypotheses and Evidence in one list. Keep them separate.
+ - Reading a pattern from a few data points. State sample size or window.
+ - Ignoring base rates (how often this happens anyway) and confounders (what else changed).
+ - A single point estimate from guessed inputs. Use ranges and show which input moves the result most.
+ - A tree whose branches overlap, so the same cause is counted twice.
- ---
+ Worked example: [references/example.md](references/example.md).
- ## Checklist (verify before responding)
+ ## Checklist
- - [ ] Setup: analytical question + default pass (note if Options matrix used)
- - [ ] Frame states question type and baseline
- - [ ] Decompose is scannable (tree or table)
+ - [ ] Question and Pass stated (Options matrix noted if used)
+ - [ ] Frame gives question type and baseline
+ - [ ] Decompose is scannable
- [ ] Hypotheses have falsifiers
- - [ ] Evidence (or `[THEORETICAL]`) mapped to hypotheses
- - [ ] Synthesis: answer, uncertainty, next step
+ - [ ] Evidence (or Thought experiment) maps to hypotheses; guesses marked `[ESTIMATED]`
+ - [ ] Synthesis: answer, key uncertainty, next step