# Ooda > Use before deciding, acting, iterating, planning, diagnosing, or competing — whenever a move should be driven by a fast adapt-loop rather than a guess. Triggers on "what should we do", "how do we approach", deciding/planning, "let's iterate / ship / pivot", incident and root-cause work, "we're stuck", and competing/beating-an-incumbent. The move: run John Boyd's real OODA loop — Observe (measure, don't assume), Orient (reframe against stale models and bias — the highest-leverage step), Decide (one lever, as a hypothesis), Act (ship a test, re-observe), loop fast at an unpredictable tempo. Boyd-faithful: not a 4-box speed circle, Orient dominates, tempo is rhythm not haste. Reach for it on non-trivial decisions and iterative work; skip mechanical single-answer tasks. - Skill: `88plug/ooda` (Agent Skill) - Install (CLI): `npx skillmds@latest add 88plug/ooda` - Raw SKILL.md: https://api.skillmd.com/api/skills/88plug/ooda/raw - Safety review: PASS (external: skill-scanner PASS, skillspector PASS) - Works with: Claude Code, Claude.ai, OpenAI Codex - Category: Coding & Dev Tools - Author: 88plug (https://skillmd.com/u/88plug) - Updated: 2026-09-17 - Page: https://skillmd.com/skills/88plug/ooda --- # OODA OODA (Col. John Boyd) is a theory of **winning through faster, better adaptation** — not a checklist and not a race. Boyd drew the loop exactly once ("The Essence of Winning and Losing", 1996) and it bears little resemblance to the popular Observe→Orient→Decide→Act ring: the four functions run continuously and overlap, wired by many feedforward and feedback arrows. This skill runs the real loop on a decision or action. **The anti-patterns that void it, up front:** it is **not a four-box speed circle**, it is **not a pure speed race**, and **Orient is not one step among four — it dominates.** The killer failure is *Orient-blindness*: acting fast on an un-reframed, biased model. ## The loop (real structure) - **Observe** — intake the unfolding circumstances, outside information, and your *interaction* with the environment. Your own prior Acts changed what there is to see, so measure now; don't assume. - **Orient** — the schwerpunkt (see below). It shapes Observe, Decide, and Act, and is reshaped by their feedback. - **Decide** — Boyd's label is **"Decision (Hypothesis)"** — a conjecture, not a certainty. Pick one lever. - **Act** — Boyd's label is **"Action (Test)"** — an experiment; its result re-enters Observe. Two **Implicit Guidance & Control** arrows run *directly from Orient to Observe and to Act*, bypassing explicit Decide — the trained-operator fast path. When your orientation is a proven, high-confidence repertoire (*Fingerspitzengefühl*), take Orient→Act and skip deliberate deciding. Force an explicit Decide-as-hypothesis under novelty or contested orientation. ## Orient — the schwerpunkt Boyd: *orientation shapes observation, shapes decision, shapes action, and in turn is shaped by the feedback.* Its five shaping inputs (his exact labels): **genetic heritage, cultural traditions, previous experience, new information, and analysis & synthesis.** These continuously run **destruction and creation** (Boyd's 1976 paper): *destructive deduction* pulls a stale frame apart into its constituents; *creative induction* recombines constituents across domains into a new concept. Why any closed model must be rebuilt: Gödel (a model can't prove itself — it needs an outside referent), Heisenberg (precision has a floor), the 2nd law (a closed model's internal disorder rises → it drifts from reality). **Orientation is never finished.** It dominates because it governs *what* you observe and *how* you decide and act. ## Tempo — "getting inside the loop" Not speed. The lever is **rhythm, variety, and unpredictability** that outpaces the opponent's (or the problem's) Orient — fast, abrupt transients they cannot track, generating ambiguity and confusion that fold them back on themselves. A *predictable* fast loop is counterable. For work: **cheapest-falsifier-first, short varied revert-safe cycles** that re-Orient before your own model goes stale. Reward rhythm; reject "rush". ## The routing procedure 1. **Observe** — measure, don't assume. Pull telemetry or run the cheapest probe. Remember your prior acts colored the field. 2. **Orient (highest-leverage)** — surface the five inputs; hunt the stale *previous experience* / cultural dogma steering you. On a surprise, anomaly, or failing baseline, run one destruction-and-creation pass: break the old frame, synthesize a new hypothesis. A mismatch means **re-orient**, not just re-observe. 3. **Decide** — pick the *single* highest-leverage lever; tag it a hypothesis. 4. **Act** — ship it as a test; feed the result back to Observe. 5. **Loop fast** — many cheap cycles beat one slow-perfect one; each cycle re-orients. Take the Orient→Act shortcut for proven moves. ## Anti-patterns this skill prevents - **Sequential 4-box circle** — drops Orient's richness and the implicit-guidance shortcuts. - **Pure speed race** — optimize tempo *relative to the problem/adversary*, not stopwatch haste. - **One-pass loop** — it runs continuously; there are many feedback arrows. - **Decide-centric / Observe→Decide jump** — never skip an Orient reframe. - **Orient-blindness** — acting on an un-reframed, biased model (the killer). - **PDCA-rename** — "just iterate faster" while skipping destruction-and-creation. - **Predictable rhythm** — a regular tempo is trackable; vary it. ## Composes with - `cynefin` — Cynefin lives inside **Orient**: classify the domain by its constraints before you decide the response. - `scientific-method` — Observe and Act are the experiment: the cheapest-falsifier probe and the ship-it-as-a-test, pre-registered. - `break-dogma` — a stale *previous-experience* input to Orient is exactly the inherited assumption to test and break. ## Sources Boyd, "Destruction and Creation" (1976); "Patterns of Conflict" (1977–86); "Organic Design for Command and Control" (1987); "The Essence of Winning and Losing" — the final OODA sketch (1995/96); collected in *A Discourse on Winning and Losing* (DoD scan, 2018). Frans Osinga, *Science, Strategy and War: The Strategic Theory of John Boyd* (2007). Chet Richards, "Boyd's OODA Loop" (*Necesse*, 2020) and *Certain to Win* (2004). Robert Coram, *Boyd* (2002).