Lesson 6 of 7
Keeping a decision journal
Record each decision with its reasons, expectations and confidence at the time, so hindsight cannot rewrite what you knew.
By Javier Sanz
Founder of ValueScreener. Built Ninety Nine, a securities services provider, and sold it to a tier-one global bank; ran operations and internationalization at Alpaca, the brokerage API.
Who writes thisEditorial processLinkedIn (opens in a new tab)X (opens in a new tab)
In one minute
- A decision journal entry records the reasons, the expectation and the confidence you held at the time, written down at the time.
- It exists to defeat hindsight bias, which quietly rewrites what you thought you knew once an outcome is known.
- An entry is short: what you decided, why, what you expect to happen, and how sure you are.
Memory is not a reliable record of a past decision. Once an outcome is known, memory edits the confidence you had and the reasons you gave. A decision journal is a dated entry that hindsight cannot touch.
One holding, four journal entries over two years
- Month 0BoughtMargin expansion thesis. Confidence written as 60 percent. Watching gross margin and the capital expenditure behind the automation program.
- Month 6First checkGross margin up 40 basis points, matching the expectation written at purchase. Confidence unchanged.
- Month 14Trigger reachedThe program's sponsor left, the trigger written at purchase. Confidence cut to 35 percent. New condition: gross margin below 40 percent in a fiscal year.
- Month 24ClosedGross margin fell below 40 percent, the condition named at month 14. Closed at a loss, as that entry anticipated.
What an entry records
An entry needs four things: the decision, the reason behind it, what you expect to happen and by when, and a number for how confident you are. In Tetlock's forecasting tournaments, reported with Gardner in Superforecasting, every forecast was written down, dated and scored against what happened. The best forecasters updated often, in small steps. A journal gives one investor that same record.
Why memory alone is not enough
Kahneman describes the mind's poor ability to reconstruct what it believed before it learned the outcome. A year after a loss, memory tends to report that you always doubted the thesis and were ready to sell the moment anything went wrong. The written entry, made before the outcome existed, is the only record that cannot be edited this way.
Keep it short and dated
An entry that takes ten minutes gets written. One that takes an hour gets skipped after the third holding. Four lines, dated, are enough: the decision, the reason, the expectation, the confidence.
Read it back later, not just at the time
The value of the journal is mostly realized on the second reading, when you compare what you actually wrote against what you now believe you thought. The gap between the two is hindsight, made visible. Across many entries, compare your stated confidence with how often you were right: 60 percent calls should come true about 60 percent of the time.
Check yourself
What does a decision journal entry mainly need to record?
- A summary written after the position is closed
- The reasons, the expectation and the confidence held at the time
- Only the final outcome of the decision
Show the answer
Answer: B.
An entry is only useful if it is written at the time of the decision and captures what you actually believed then, not a later summary.Why is a dated entry more reliable than memory a year later?
- Because a written entry takes longer to produce
- Because memory is always wrong about numbers
- Hindsight quietly edits memory once an outcome is known
Show the answer
Answer: C.
Kahneman's point about hindsight is that the mind rewrites the past to fit what it now knows. A written entry made before the outcome is the only thing that resists that.When should an entry be written?
- At the time of each significant decision, including the purchase itself
- Whenever the price moves enough to be worth noting
- Only once, when the position is finally closed
Show the answer
Answer: A.
The entry has to be written at the time of the decision to be a record hindsight cannot rewrite. A price move on its own is not a decision.Sources
- Tetlock and Gardner, Superforecasting: The Art and Science of Prediction, 2015.
- Kahneman, Thinking, Fast and Slow, 2011, on hindsight bias and the construction of a coherent story after the fact.
What this leaves out
- A journal is only as useful as its honesty at the time it is written. An entry written to look good in hindsight defeats the purpose.
- The example entries and dates here are illustrative. A real journal is built entry by entry, over years, not reconstructed after the fact.
- Nothing here is advice about any security.