Judging a Decision by Its Process

A process judgment is a dated note of why the choice made sense given what was known, written ahead of the result. A sloppy pick can still get lucky, and a sound one can still fail. Those two cases are why the note, not the later result, is the grade.

Why the outcome is not the grade

Use this after the result is known. Decisions are made with incomplete information. Outcomes are affected by timing, chance, changing conditions, other people, and variables that were never under control. Outcome-only judgment produces two errors: it rewards a bad process that got lucky, and it abandons a good process that met bad conditions.

The better question: given what was known at the time, was the process reasonable?

Different situations have different amounts of control. Some outcomes are mostly the process. Some are mostly luck. The more of the result that was actually owned, the more the result may grade the process. A well-specified exam is higher-control than a career bet. That is a direction, not a percentage.

The instrument that makes the question answerable is a note written before the outcome exists: why the choice made sense then. Afterward, process quality and outcome quality stay separate. The result is one piece of feedback, not the entire verdict. The next repetition is where the process improves.

The reviews

A process review asks six things. What was known at the time? What was uncertain? What downside was protected? What option improved the position — the question Positional Decisions and Expected Value owns? What would change my mind — the handoff to Changing Decisions? What did the outcome teach?

Four lightweight checks sit beside that list. They replace a separate toolkit. What assumption would break this if false? What is the strongest case against the preferred option? If this fails, what is the most likely reason? Is this being chosen because it is good, or because it protects comfort, identity, or self-image?

Learning outcomes are noisy, but academic assessments are the high-control case. With a sound process, results are more predictive; luck shrinks toward being sick on the day. A poor test score does not automatically mean the method was bad. A good score does not automatically mean it was strong. Inspect the process, and do not treat a single score as nothing.

Was encoding deep enough? Was retrieval spaced and interleaved? Were gaps found early? Was the exam different from expected? Was the result caused by knowledge, execution, stress, or timing?

That inspection turns performance into Self-Regulation instead of a verdict on the person. Kolbs Experiential Cycle is the after-action shape: what happened, how it felt and why, what rule that suggests, what changes next time.

Resulting is the name for the first row: treating the result as the grade.

In a high-control repeat the result may amend the dated note. In a low-control bet it is one noisy input. Either way the next move is to improve the process for the next repetition — regulation, not a verdict on the person.

Links into the knowledge base

• Decision Making — hub; the split lives here in full.
• Positional Decisions and Expected Value — position / EV questions in the review.
• Changing Decisions — the change review this page’s last questions hand off to.
• Kolbs Experiential Cycle — after-action: what happened, what rule, what changes.
• Self-Regulation — performance as regulation, not a verdict on the person.
• Suicidal Empathy — empathy replacing judgment.

Open questions

Is the pre-outcome note enough, or does a high-consequence choice need a repeated journal?

Sources

• Duke, Thinking in Bets (2018) — resulting: grading a decision by how it turned out. Names the two errors (lucky-bad-process, unlucky-good-process).
• Fischhoff 1975 — hindsight rewrites what was knowable. Support for writing the reasons before the result exists.
• Howard / decision-analysis — decision quality and outcome quality are separate objects.