Reviewing poker hands is what separates the player who grinds from the player who improves. And most people do it backwards: open the history, find the biggest pot you lost, confirm the other guy got lucky, close the program feeling fine. That isn't review, that's consolation.
The method that works takes 20 minutes and starts by choosing badly on purpose: the hands worth studying are almost never the ones that hurt.
Why result is the worst filter
If you pick hands by how much you lost, you're studying variance, not your game. The big pot you dropped with a set against a flush has nothing to teach you — the money went in right and it went wrong. Meanwhile the small pot you won with a lazy line never shows up on that list, and that one had a real mistake in it.
House rule: a hand gets reviewed for decision quality, not for the result. A hand you won can be badly played and a hand you lost can be flawless.
The three filters that work
Sort your history by these instead of by money:
1. Hands where you hesitated more than three seconds. That pause is a spot you haven't solved. Tag it live — almost every client has a hotkey for marking hands mid-session.
2. Hands where you called the river. It's where money disappears quietly. Reviewing ten river calls a month tells you more than a hundred won pots.
3. Hands where you bet and got called. If you always get called, your bet wasn't collecting value — it was paying someone else's.
Those three filters produce five to eight hands per session. That's your material.
The 20-minute method, hand by hand
For each tagged hand, in order, no skipping:
- Cover his cards. If you've already seen what he had, you aren't deciding, you're justifying.
- Write down what you thought he had. Not "something strong": a concrete range. "Medium pairs, AJ, flush draws." If you can't write it, you just found the error.
- One question per street. Why bet here instead of checking? Why this size and not another? One question, one answer, one sentence.
- Write the correction in the present tense. Not "I should have folded", but "dry board, out of position, one pair and no draw, I shut down on the turn". That's a rule you can use tomorrow.
- Uncover the cards last. Only to check whether your range was close — never before.
Twenty minutes covers five hands done this way. More hands with less care doesn't pay: what fixes the learning is writing the correction, not the amount of history that scrolled past.
The notebook is half the work
Without a record, every session restarts from zero. With one, three weeks in you'll see the same sentence written four times — and that's no longer a hand, it's a leak.
Two columns is enough: the situation and the new rule. If you already keep a session tracker, add one line a day with that day's correction. By month's end you have the exact syllabus of what you need to study, written by your own game.
When to review, and when not to
Never hot. After a losing session what comes out of a review is blame, not judgment, and it leaves you worse for the next one. Sleep on it.
Don't review tired either, or with a table still open in another tab. Twenty clear-headed minutes beat two hours dragging yesterday's tilt around.
And you don't need a paid tracker to start: most rooms export hand history and every client lets you tag hands. When you want to move up to equity and review software, it's on our tools page.
The lesson
Good review means choosing worse on purpose: fewer hands, less dramatic ones, with the correction written as a sentence you can apply tomorrow. History won't tell you whether you're running well — that's variance and it can't be studied — but it will tell you exactly which street you're still giving money away on.
In the club those hands get reviewed by several people at once, which is when the reads you'd never spot alone start showing up. I want in on the club.