You Don’t Remember What You Decided

Ask a man about the best call he’s made in the last three years and he’ll give you a clinic. The reasoning. The signals he spotted that nobody else did. The moment he knew. It’ll be detailed, it’ll be structured, and he’ll tell it well, because he’s told it before.

Ask the same man about his worst call and watch what happens to the shape of the story.

Suddenly there’s context. The market shifted. The partner didn’t hold up his end. The timing was unlucky. There were things he couldn’t have known. It’s not that he’s lying to you, and it’s not even that he’s lying to himself exactly. It’s that his brain did the work for him, quietly, over months, and handed him a version he can live with.

This is hindsight bias, and it is not a small effect. It’s one of the most reliably demonstrated things in cognitive psychology, it runs harder in confident people than in uncertain ones, and it has one specific consequence for you:

You are not learning from experience. You are learning from a story about experience, written by the guy with the most to lose from the truth.

Twenty years of that and you don’t have twenty years of judgment. You have one year of judgment and nineteen years of increasingly polished narration.

You’d think the smart ones would be immune. It’s the reverse.

The better you are at reasoning, the better you are at constructing a coherent story after the fact. A man with weak reasoning skills makes a bad call, feels bad, and remembers feeling bad. A man with excellent reasoning skills makes a bad call and builds a genuinely airtight case for why it was the right decision given what he knew at the time.

Sometimes that case is even true. That’s what makes it so dangerous. You cannot tell from the inside which of your explanations are accurate reconstructions and which are sophisticated cover, because they feel identical.

And there’s a second layer. Your track record is load-bearing for your confidence, and your confidence is load-bearing for your ability to lead. So there’s a real, functional reason your brain protects the record. It’s not vanity. It’s structural. Which means willpower and good intentions will not fix it, any more than squinting harder fixes bad eyesight.

You need an instrument. Something that captured the truth before you knew how it turned out, because after you know, it’s gone.

Ten minutes. Five fields. Written before the outcome is known.

Field One: The call. One sentence, plain language. “Hiring Sarah as VP of Ops at 165.” Not a paragraph of context. One sentence you’ll understand in a year.

Field Two: The state. How you feel right now. How you slept. What pressure you’re under. Whether you’re coming off a fight, a win, a bad quarter, a compliment. Whether anyone is waiting on this.

This is the field everyone skips and it is by a wide margin the most predictive one in the whole log. Nobody wants to write “I’m exhausted and I want this off my plate.” That’s exactly why you write it.

Field Three: The expectation. What you expect to happen, with a number and a date. “By March, this cuts my ops involvement from fifteen hours a week to under five.” Vague expectations can’t be graded, and ungradeable predictions are how you stay comfortable forever.

Field Four: The confidence. A percentage. Force yourself. Not “pretty confident.” Seventy percent. The number feels arbitrary the first ten times and then it starts becoming the single most useful column you have, because you’ll discover your eighty percents come in at about fifty five, consistently, for your whole life.

Field Five: The kill criteria. What would tell you this was wrong, and by when. Written now, while you’re still capable of being objective about it. “If she’s still asking me to approve routine spend in month four, I misjudged the fit.”

That’s the whole thing. Five fields, ten minutes, and you only do it for decisions that matter, which is maybe four to six a month.

Abstract templates are easy to nod at and hard to use, so here’s a real one, lightly scrubbed. This is an entry from my own log from about two years ago.

Now look at what that entry did for me eight months later, when I was deciding whether to keep him.

The state field told me I’d hired partly out of embarrassment, which is worth knowing when you’re evaluating your own satisfaction. The expectation gave me something specific to grade instead of a vague sense of “he’s been helpful.” And the kill criteria, which I’d written while calm, had actually tripped in month three. I’d noticed and rationalized past it, the way everyone does, because by then I liked the guy.

Without the log I’d have kept him another year on vibes. With it, I had a note from a more objective version of myself telling me exactly what I’d said would matter.

That’s the entire value proposition. You’re leaving instructions for a future version of you who will be compromised.

The log is only half of it. A record you never read is a diary.

One. Log only Tier 3 and Tier 4 decisions. If you read Monday’s edition, you already have the sorting system. Heavy Doors and One Way Streets get logged. Coin flips and cheap bets don’t. This keeps the practice sustainable, and sustainable beats thorough.

Two. Weekly, five minutes. Scan for anything that’s hit its review date or tripped its kill criteria. That’s it. Most weeks, nothing has, and you close the doc.

Three. Quarterly, forty five minutes. This is the one that matters. Pull every closed decision and sort it into three columns:

Notice what’s missing. There’s no “bad decision, good outcome” column, because that belongs in the third bucket. Getting away with something is not a win, it’s a warning, and treating it as a win is exactly how men build confidence in processes that are quietly going to destroy them later.

Only the third column contains lessons. Column two is variance and you have to make peace with it. If you go looking for lessons in column two you’ll learn superstition, and half the “wisdom” in business is exactly that: a guy who got unlucky once and built a rule around it.

Four. Hunt for state patterns. This is where the log pays for itself. Line up every entry in the third column and read the state field only. Ignore everything else.

You will find a pattern. Almost everyone does, and it’s rarely what they’d have guessed. Common ones I’ve seen in my own log and in other men’s:

Decisions made after 9pm. Decisions made within a day of a conflict with someone close. Decisions made when someone impressive had just flattered them. Decisions made on the back end of a long stretch without a real day off. Decisions made when the deciding itself had become the burden and any answer felt like relief.

That last one is the big one. A huge portion of bad calls aren’t wrong analysis. They’re a man buying his way out of the discomfort of an open question.

Once you can see your pattern, you get the most powerful move available: you learn to postpone. Not decide better in that state. Just not decide in that state.

“I’ll never keep it up.” Ten minutes, five or six times a month. That’s about an hour a month against decisions worth six and seven figures. If that’s the thing that doesn’t fit, the problem isn’t the log.

“This feels like homework.” So did the first budget you ever built, and the first time you tracked what you actually ate, and every other instrument that made an invisible thing visible. The homework feeling is the tell that it’s working. Comfortable practices don’t change anything.

“I’ll be too honest and it’ll be uncomfortable.” Correct. Field two in particular is going to make you cringe about four months in, when you reread “I’m tired of this conversation and I want it done” attached to a decision that cost you eighty grand. That cringe is the entire product. You’re paying in discomfort now to avoid paying in dollars later, which is a trade you make in every other part of your life.

“What if I write it down and someone sees it?” Then keep it somewhere private. This is not a document for your team.

Here’s the practical gap in all of this.

A good chunk of your real decisions don’t get made at a desk with ten quiet minutes available. They get made in a meeting, out loud, in the middle of a conversation, and then you move on to the next thing. By the time you sit down to log it that evening, you’re already reconstructing. You remember the conclusion. You’ve lost the reasoning, the objection somebody raised that you waved off, and the exact commitment you made.

Fathom closed this gap for me. It sits in the meeting, records, and hands you a transcript and summary afterward. So when I log a decision that got made on a call, I’m starting from what was actually said, not from my recollection of what was said, which is the very thing this entire article says you can’t trust.

It’s free for the core use, it takes about four minutes to set up, and it means the record starts honest.

Experience does not automatically become judgment. That conversion requires an honest record, and you do not have one, because the organ keeping the record has a vested interest in the outcome.

Most men will read this, nod, and keep running on the flattering version. That’s fine. It’s comfortable, and it’s most of the field.

The ones who write it down find out something specific within about two quarters: their judgment is better than they feared in some areas and much worse in one or two very particular conditions. And knowing which conditions is worth more than any amount of general self improvement, because it’s actionable tonight.

You don’t need to become a better decision maker across the board. You need to stop making decisions in the two states where you’re demonstrably terrible.

You just have to be able to see them.

Open a blank doc. Write down the last three real decisions you made this month, from memory, using the five fields.

Then go find the emails, the texts, the calendar entries from when you actually made them.

Compare. The gap between what you just wrote and what the record shows is the exact size of the problem this fixes.

Refined. Relentless. Unapologetic.

Marcus