Skip to content
How to ThinkIn the Age of AI
Rulev2RL-470

Score process and outcome independently on a 2x2…

Evaluate decision quality separately from outcome quality by scoring process and results independently, placing decisions in a 2x2 matrix to distinguish deserved success, bad luck, dumb luck, and deserved failure.

Why This Is a Rule

This is the operational companion to Evaluate decision process separately from outcome — 'was… (evaluate process separately from outcome). While Evaluate decision process separately from outcome — 'was… establishes the principle, this rule provides the specific tool: a 2×2 matrix that creates four distinct categories, each requiring a different response.

Deserved success (good process, good outcome): your analysis was sound, your execution was strong, and the result confirmed it. Reinforce this process — it works. Bad luck (good process, bad outcome): you did everything right, but unpredictable factors produced a poor result. Don't change the process — the variance was external. Dumb luck (bad process, good outcome): you got lucky despite flawed analysis or execution. This is the most dangerous category because the good outcome masks the flawed process. If you reinforce this process, future outcomes will regress to what the process quality predicts. Deserved failure (bad process, bad outcome): the process was flawed and the outcome reflected it. Fix the process.

The independent scoring is crucial. If you evaluate holistically ("how did this decision go?"), you collapse two dimensions into one — and good outcomes always make processes look better in retrospect (resulting — see Evaluate decision process separately from outcome — 'was…). Scoring each dimension independently preserves the distinction.

When This Fires

  • During any formal post-decision review or retrospective
  • When reviewing decision journal entries (Review decisions in three steps)
  • When someone claims a decision was "good" or "bad" — ask which quadrant
  • When building a culture of process accountability rather than outcome accountability

Common Failure Mode

Conflating the quadrants: treating dumb luck as deserved success ("we chose well!") and bad luck as deserved failure ("we screwed up!"). Both errors produce wrong lessons. The team that got lucky on a bad process learns to repeat the bad process. The team that got unlucky on a good process abandons a process that would produce good outcomes over time.

The Protocol

(1) After a decision outcome is known, score two dimensions independently (1-5 each): Process quality: given what was knowable, was the reasoning sound? Were appropriate alternatives considered? Were risks assessed? Was the right decision framework used? Outcome quality: did the result meet expectations? Was the outcome good, neutral, or bad? (2) Place the decision on the 2×2: high process + high outcome = deserved success. High process + low outcome = bad luck. Low process + high outcome = dumb luck. Low process + low outcome = deserved failure. (3) Take the appropriate action for each quadrant: deserved success → reinforce process. Bad luck → keep process, increase resilience to variance. Dumb luck → fix process before luck runs out. Deserved failure → diagnose and fix process. (4) Track quadrant distribution over time. A healthy decision practice clusters in "deserved success" with occasional "bad luck."