Keeping a Decision Journal and Reviewing Outcomes

The fundamental problem with experiential learning is that the learner controls the record. Every person who has survived to adulthood has accumulated years of decisions and outcomes, and almost none of them have studied that history with any rigor. Instead, they carry a curated highlight reel — the decisions that worked remembered as wisdom, the ones that failed reattributed to bad luck or external interference, the predictions that missed quietly dropped from the story.

This is not pathology. It is cognitive architecture. The brain encodes outcomes preferentially over process, updates predictions retroactively to match what happened, and constructs a coherent narrative of past events that makes the rememberer look more competent than they were. Daniel Kahneman called this "the illusion of understanding." Philip Tetlock, who spent decades studying expert forecasters, found that most experts performed only marginally better than chance, and that the ones who were most confident in their predictions were often the least accurate. The commonality among the best forecasters was not brilliance — it was the habit of tracking their predictions systematically and updating their models when they were wrong.

A decision journal operationalizes that habit.

The mechanics matter. The entry must be written before the outcome is known — this is non-negotiable, because post-hoc entries are contaminated by outcome knowledge. The entry should capture: the decision in neutral language (not self-flattering framing), the alternatives you considered, the information you had and its quality, your confidence level, the specific outcome you expect and by when, and the single most important assumption underlying your reasoning. That last item is the lever. Most decisions rest on one key assumption. Naming it explicitly makes the review concrete: was the assumption correct?

The review structure should mirror the entry. The questions to answer at review time: What actually happened? Was the key assumption correct? If the outcome differed from prediction, at what point did the divergence occur? Was the reasoning sound given what you knew then, even if the outcome was bad? What information would have changed the decision?

That last question prevents a common error: updating on outcomes rather than on information. If a decision failed because of a genuinely unforeseeable black swan, you should not update your decision-making process much — just your probability estimates for rare events. If it failed because you ignored available information, or because your reasoning had a structural flaw, the update should be significant. The journal forces you to distinguish these cases.

Several patterns tend to emerge from sustained journaling that are difficult to see any other way:

Overconfidence calibration curves. By tracking confidence levels and outcomes, you can plot whether your 80% confident predictions are right 80% of the time. Most people find their high-confidence predictions win far less often than they expect. This is actionable: you can apply a systematic discount to your future confidence estimates.

Domain-specific accuracy. You may find that your judgment is reliable in some domains and poor in others. Most people have domains where their intuition is actually well-calibrated — usually areas of deep experience — and domains where they systematically over- or underperform. The journal reveals this map.

Emotional state signatures. Many journalers start noting their emotional state at decision time as a matter of course. Over time, you may find that decisions made under time pressure, social approval-seeking, or anxiety have characteristic failure modes. This is genuinely useful knowledge: it tells you when to impose delays or additional review on your own decision process.

Recurring rationalizations. The same justifications for bad decisions tend to appear repeatedly. "This time is different." "The upside is worth it." "I'll compensate later." Seeing these phrases in your own handwriting, across multiple entries, over multiple years, makes them much harder to deploy uncritically.

Scope insensitivity. The journal often reveals that small decisions receive the same quality of reasoning as large ones, and vice versa. Time invested in analysis correlates poorly with decision stakes. Tracking both helps calibrate your attention.

The question of what counts as a "significant decision" worth journaling is worth addressing directly. The threshold should be lower than you think. Decisions about how to spend a week, whether to raise an issue with someone, what to read, how to structure a project — these seem minor, but they are the texture of a life. Their patterns matter as much as the occasional major decision. The major decisions often matter less than people believe: most studies of life satisfaction suggest that major life events (job changes, moves, relationship status) adapt to relatively quickly, while the quality of daily habits and small choices proves more durable.

On the practice of separating process from outcome: this is a core discipline from the poker literature, particularly Annie Duke's work. A poker player who gets their chips in with a 70% edge and loses has made a good decision that produced a bad outcome. Evaluating that decision as a failure because of the outcome is how you learn to play scared, to refuse good bets because they sometimes lose. The decision journal enforces the discipline of evaluating reasoning at the time of the decision, before you know the result. This is genuinely difficult — it requires holding the timeline of knowledge carefully — but it is the mechanism that separates learning from superstition.

There is also a subtler benefit that accumulates over years: the journal makes you take your own future seriously. When you commit to reviewing a decision in ninety days, you are treating your future self as a real person who will actually check your work. This changes how you decide now. The unreviewed decision is essentially a promise you know you will never be called on. The journaled decision is accountable.

The Stoics had a related practice: the evening review, in which they examined the day's decisions and actions for conformity with virtue and reason. Marcus Aurelius's Meditations reads, in many passages, like exactly this kind of review — a man holding himself accountable to his own stated principles, noting where he fell short, not to punish himself, but to sharpen his aim. The decision journal is that practice, made systematic and extended across longer time horizons.

Start with high-stakes decisions and build from there. The first six months will feel awkward — the predictions will be embarrassing, the reviews uncomfortable. That discomfort is confirmation that the practice is working. You are meeting the part of yourself that operates on unexamined assumptions, and it is rarely flattering. The alternative is to let that self continue operating unexamined for another decade.