In 1975, a graduate student named Baruch Fischhoff ran an experiment that should worry every manager who trusts their own judgment. He gave people a paragraph about an obscure 1814 conflict between British forces and the Gurkhas of Nepal, then told different groups different outcomes: a British victory, a Gurkha victory, a stalemate. Each group was asked how likely that outcome had been, given only the background they’d read. The people who were told the British won rated a British victory as far more probable than the people who were told something else. Same facts. Different answer. Everyone thought the ending they’d been handed was the obvious one all along.
Fischhoff called it creeping determinism. Most people know it as the “knew-it-all-along” effect, and it has since been replicated across more than a hundred studies. It is one of the most durable findings in cognitive psychology, and it has a specific, expensive consequence for how you manage: your memory is a bad witness to the decisions you make.
The reason you’re not getting better at judgment
Managers are supposed to improve with experience. You make calls, you see how they turn out, you learn. That’s the theory. The problem is the feedback loop that theory depends on is broken in two places at once.
The first break is hindsight bias. Once you know how something turned out, your brain quietly rewrites what you believed beforehand. The vendor who missed every deadline becomes the vendor you “always had a bad feeling about,” even though you signed the contract with real enthusiasm. The hire who didn’t work out becomes the candidate whose red flags were “right there,” even though you fought to get them approved. The Decision Lab describes this cleanly: current knowledge of an outcome contaminates your memory of the judgment you actually made, which makes you overestimate how well you saw it coming.
The second break is what poker champion turned decision scientist Annie Duke calls “resulting.” In her book Thinking in Bets, she names the error of grading a decision by how it turned out instead of by how sound it was at the time. Her go-to example is Pete Carroll’s goal-line pass call at the end of the 2015 Super Bowl. It got intercepted, and the whole world called it the dumbest play in football history. Duke’s point: if the pass had been caught, the identical call would have been hailed as genius. The decision didn’t change. Only the outcome did, and the outcome came partly down to luck.
Put those two together and you get a manager who cannot actually learn from experience. Good outcomes teach you that whatever you did was smart, even when you got lucky. Bad outcomes teach you that you should have known, even when the information genuinely wasn’t there. You accumulate years of “experience” that mostly confirms you were right to trust yourself. That is not the same as getting better.
The call that looked stupid until it didn’t
Years ago, running IT operations, I pushed hard to delay a system migration everyone else wanted to ship on schedule. The business case for waiting was thin on paper. We’d already sunk months into the project, the executive team wanted it done, and I was the lone voice saying the data-integrity testing wasn’t finished. We slipped the date by six weeks. During those six weeks, nothing dramatic happened, which made me look like the guy who cost the company momentum over nothing.
Then the testing surfaced a mapping error that would have corrupted a chunk of customer billing records at cutover. The delay went from a black mark to a save.
Here is what I want you to notice. If you asked me to reconstruct my reasoning from memory today, I’d tell you a clean story about how I saw the risk clearly and held the line. That story is mostly a lie my memory built after the fact. What I actually remember, if I’m honest, is that I was about 60 percent sure the testing gap mattered and 40 percent sure I was being paranoid and career-damaging. The good outcome erased the doubt. And if that mapping error had never existed, I’d remember the whole episode as the time I let caution cost us six weeks. Same decision. My memory would have filed it under “win” or “mistake” depending entirely on a fact I didn’t have when I decided.
I only know the real numbers because, on that project, I’d started writing decisions down before I knew how they’d end.
What a decision journal actually is
A decision journal is a record of your reasoning captured at the moment of the decision, before the outcome exists to distort it. Shane Parrish at Farnam Street has popularized a practical version that works well for managers. When you face a decision that matters, you write down a few things in plain language:
- The situation and how you’re framing the problem. Not the polished version. The messy one you’d actually say out loud.
- The alternatives you considered and why you rejected them. This is the field people skip, and it’s the most valuable one. Rejected options are where your reasoning lives.
- What you expect to happen, with a rough probability. “I think there’s about a 70 percent chance this vendor delivers on time.” Numbers force honesty. “It’ll probably be fine” tells your future self nothing.
- The range of ways it could go, including the bad ones. A sentence or two on what failure would look like.
- Your physical and mental state. Time of day, stress level, whether you’re deciding this tired at 6 p.m. or fresh at 9 a.m. You’ll be surprised how often your worst calls cluster.
That’s it. Five short entries, five minutes. You’re not writing an essay. You’re taking a photograph of your thinking before your brain gets a chance to retouch it.
The probability estimate is the part that separates a decision journal from ordinary reflection. When you commit to “60 percent confident,” you’ve created something your memory cannot renegotiate later. Six months on, you either hit that band or you didn’t. There’s no room to tell yourself you “always knew.” The number is right there in your own handwriting.
Why this is the single highest-leverage self-management habit
The best evidence that this works comes from the largest forecasting study ever run. Philip Tetlock and colleagues Barbara Mellers and Don Moore at the University of Pennsylvania ran the Good Judgment Project, a multi-year tournament where thousands of ordinary people forecast world events. A small group, the top two percent Tetlock dubbed superforecasters, consistently and dramatically outperformed everyone else, including intelligence analysts with access to classified information.
What made them better was not raw intelligence. It was method. They made specific, probability-tagged predictions, they kept score, and they went back and reviewed where they’d been wrong. A randomized controlled trial inside the project showed that even a short course in probabilistic reasoning measurably improved accuracy. The superforecasters weren’t born calibrated. They got calibrated by keeping records and confronting them.
That is exactly what a decision journal does for a manager. It turns your job into a forecasting tournament where you finally keep score. Without it, you’re running the tournament with no scoreboard and a memory that reports you’re winning.
How to actually run it
Start narrow. Do not journal every decision, or you’ll quit in a week. Reserve it for calls that are consequential and uncertain: a hire, a reorg, a big vendor commitment, a bet on a project timeline, a decision to let someone go. Roughly the decisions you’d lie awake thinking about anyway.
Then, once a quarter, block ninety minutes and read the entries whose outcomes are now known. You’re looking for patterns, not verdicts on individual calls:
- Where is your confidence miscalibrated? If everything you rated “80 percent likely” happened only half the time, you’re systematically overconfident, and now you know by how much.
- Which conditions produce your bad calls? Time pressure, a specific person in the room, decisions made at the end of a draining week. My own journal taught me that my judgment on people decisions degraded noticeably when I was avoiding a different hard conversation. I was displacing.
- What did you get right for the wrong reasons? These are the dangerous wins, the ones that teach you bad lessons because they happened to work.
This quarterly review is where the compounding happens, and it pairs naturally with the after-action review discipline you might already run on projects. The difference is that a decision journal reviews your internal process, not the team’s execution. One looks outward at what the group did. This one looks inward at how you actually think, which is the harder mirror to hold up.
The uncomfortable part
Most managers resist this for a reason they won’t say out loud: a decision journal removes your ability to feel like a better decision-maker than you are. It replaces a flattering story with a scorecard. The first honest quarterly review is humbling, because you discover you were overconfident on the calls you were proudest of and that a couple of your “instinct saves” were coin flips that landed your way.
That discomfort is the entire value. The manager who can look at their own miscalibration without flinching is the one who stops making the same misjudgment for the fifth time. Everyone else is stuck running on autopilot, collecting years of experience that just keeps confirming what they already believe. Twenty years into this work, the habit I’d protect above almost any other is the boring one: write down what you think will happen, and why, before you find out. Your memory will not do it for you. It’s too busy making you look good.