Process-First Decision Review
Improve the thinking behind decisions instead of judging only their outcomes
- Difficulty
- Easy
- Time to result
- ~weeks to results
- Steps
- 5
- Confidence
- 97%
A result alone does not reveal whether the decision behind it was good. Chance can reward weak reasoning or punish a sound choice, so reviews should reconstruct the information, assumptions, emotions, and mental checklist present at decision time. The reviewer asks how the conclusion was reached rather than merely identifying the move, investment, or forecast that lost. The process is then compared with what a robust decision procedure should have included. Missing checks, distorted probabilities, and recurring biases become updates to a reusable checklist. Wins receive the same scrutiny because they may conceal luck and reinforce a flawed method. Over repeated decisions, this converts isolated outcomes into systematic improvements in judgment while reducing hindsight bias and resulting.
Origin
Extracted from Masters in Business through Safi Bahcall's explanation of a lesson from Garry Kasparov's How Life Imitates Chess.
Core principles
- 01A good outcome can come from a poor decision
- 02A bad outcome can come from a sound decision under uncertainty
- 03The reusable lesson lies in the decision process
- 04Mental checklists should evolve from reviewed experience
- 05Reviews must include wins as well as losses
How to run it
- 1
Freeze the original context
Reconstruct the information available when the decision was made, excluding facts learned afterward. Capture the goal and constraints at that moment.
Pro tip Use contemporaneous notes whenever possible.
Watch out Hindsight will make later facts feel as though they were always obvious.
- 2
Rebuild the reasoning
List the options considered, assumptions made, probabilities assigned, and checks performed. Include emotional or organizational pressures that shaped the choice.
Pro tip Ask what was in your mind, not merely which action you took.
Watch out A polished retrospective explanation may differ from the actual reasoning.
- 3
Separate process from result
Judge whether the reasoning was appropriate given what was knowable at the time. Evaluate the outcome separately.
Pro tip Use a two-by-two of good or bad process against good or bad outcome.
Watch out Do not automatically praise winners or condemn losers.
- 4
Identify the process defect
Find missing information, poor probability estimates, biased framing, or checklist omissions that could recur in another context.
Pro tip Prioritize defects that appear across multiple decisions.
Watch out Overfitting a rule to one unusual result can make future decisions worse.
- 5
Update and test the checklist
Change the decision checklist and apply it prospectively to comparable choices. Review whether the update improves judgment over time.
Pro tip Keep checklist items concrete enough to answer before committing.
Watch out A checklist that continually expands without prioritization becomes unusable.
In the wild
An investor loses money after buying a company with apparently strong fundamentals. Instead of concluding that the sector is uninvestable, the investor reconstructs the original thesis and discovers that customer concentration was never checked. The checklist is updated to require concentration analysis before future purchases.
→ One loss produces a reusable safeguard across later investments.
After a chess game, the player examines not only questionable moves but the thought process that produced them. Even a win is reviewed for reasoning that an opponent failed to punish, preventing the favorable result from hiding a recurring weakness.
→ The player improves decision quality rather than merely memorizing one position.
Common mistakes
Reviewing only failures
Wins can contain weak reasoning masked by luck. Ignoring them allows dangerous habits to become reinforced.
Using facts unavailable at the time
A fair process review must use the original information set. Otherwise hindsight bias makes every error appear preventable.
Fixing only the individual move
Changing one decision without identifying the recurring reasoning defect produces little transferable improvement.
Is it for you?
Best for
It is best for investors, executives, competitors, and teams making repeated decisions under uncertainty.
Not ideal for
It is not ideal as a substitute for direct technical diagnosis when a deterministic system has a clearly observable defect.
From the transcript
“What was my decision making process by which I arrived at that conclusion, and how should I improve my process.”
“And then you go back and try to change your mental checklists, so you gain much more than just the one understanding of the one…”
“Always process versus results.”
From the episode
Safi Bahcall on Why You Need to Take Crazy Risks
Safi Bahcall