Hand analysis is where you keep what the session earns.

Hand analysis is what separates players who improve from players who keep making the same mistakes. Done right, it's a system: the right tools, the right hands, the right questions. Done wrong, it's posting "what would you do" on a forum and waiting for validation.

Set up the toolkit

Three tools. And one optional one.

Before you can analyze anything, you need a way to record and review hands. Three tools cover ~95% of the work — a fourth is the gold standard for serious students.

Online players Required

Tracking software

Beyond the built-in hand-history viewer. Lets you filter by position, action, opponent type — surfacing patterns you'd never find manually.

Examples

Poker Copilot, Hand2Note, PokerTracker 4

Everyone Required

Equity calculator

Plug your hand and a realistic opponent range into a board texture, see your actual equity. Running these calculations trains your intuition for at-the-table estimates.

Examples

Equilab, Flopzilla, GTO Wizard (free tier)

Live players Required

Hand notebook

Record hands while details are fresh: stack sizes in BB (not chips), pot sizes at each street, reads. Reconstruction is impossible from memory three days later.

Examples

Notion, Apple Notes, a dedicated paper notebook

Serious students Optional

Solver (optional)

Truly answers what's optimal. Steep learning curve, not necessary below mid-stakes, but the gold standard for spot-by-spot reference.

Examples

GTO+, PioSolver, Simple Postflop

Pick the right hands

You can’t review them all. Pick the ones that matter.

The skill is selection. Three types of hands are worth your time. Two are not — and one of those two is the biggest review-time waste in poker.

Hands where you lost the most chips

Especially if you're unsure whether you played correctly. The biggest losses contain the biggest lessons — provided you can separate variance from mistake.

Hands where you hesitated

Uncertainty at the table means there's something to learn. Even if the line was correct, the hesitation itself is a signal worth tracing back.

Winning hands you suspect you misplayed

Bet too small for value? Missed a third barrel? Slowplayed and lost a street? A winning hand played badly is still a leak — and the result hides it from most reviewers.

Don't only review losing hands

Results bias is the deadliest enemy of honest review. You can play a winning hand badly and a losing hand perfectly. Analyze decisions, not outcomes.

Don't analyze coolers

AA vs KK all-in pre. Quads over full house. There was no decision to improve on. These hands cost you money but didn't cost you EV — file them and move on.

The five-step process

Reconstruct. Compare. Conclude.

A repeatable process beats raw analysis. Five steps for every hand you review — from reconstruction to a written conclusion you can revisit a year later.

01

Reconstruct the hand, street by street

Write out the action at each decision point. Include stack sizes in big blinds, pot size, position, and any read on your opponent. The act of writing it out forces precision that mental replay never delivers.

Tip

Use BB, not chips. A $200 raise means nothing without the stake — at $1/$2 it's a massive over-raise, at $5/$10 it's standard.

02

At each decision, list your options

For every action you took, consider the alternatives. Could you have bet a different size? Checked instead of bet? Raised instead of called? Don't just evaluate what you did — evaluate what you didn't.

Tip

The decisions you didn't make are usually where the real improvement lives.

03

Run the equity

Plug your hand and a realistic opponent range into an equity calculator. If you called a river bet as a 25% underdog getting 4-to-1 pot odds, the call was correct — regardless of the result.

Tip

Always assign a range, not a single hand. "He had it" isn't a range — that's hindsight.

04

Evaluate your sizing

Bet sizing is where most players leave money on the table. Were you betting big enough for value? Small enough to get called? Did your bluff size make sense given the story your line was telling?

Tip

Sizing carries information. A 33% c-bet says something different than a 75% c-bet, and your opponent reads both.

05

Write down the conclusion

Identify the specific mistake — or confirm the hand was played well. Keep a running log. Over time, your notes become a personal strategy guide built from your own experience.

Tip

"Should have bet bigger on the turn" isn't a conclusion. "Should have bet 75% pot on a draw-heavy turn against a station" is.

The six questions

For every hand you review.

01

Did I take the right action at each street?

02

Was my bet sizing correct?

03

Did my opponent's player type influence my decision?

04

If I won, did I extract maximum value?

05

If I lost, was the loss avoidable?

06

Would I play this hand the same way next time?

Modern study partners

AI for the math. Forums for the meta.

In 2026, AI assistants are a serious hand-analysis tool — they run equity, walk through ranges, and explain reasoning at any depth. But the player pool you face tomorrow night is local knowledge. Use both — they're stronger together than either alone.

AI assistants

Specific, fast, infinite patience.

“Open to 5x preflop next time — your 3x let five players in cheaply. On the turn you have nine outs in position; bet 2/3 pot instead of checking. Your river fold was correct given his check-raise represented a flush.”

Available 24/7 · Run equity · Walk through ranges
Communities & forums

Lived experience and meta.

“This sizing is GTO-optimal but at $1/$2 live, the player pool calls way wider. Stations don’t fold flush draws to 2/3 pot — size up to pot. Also: I see this player on Wednesday games, takes him three sessions to start 3-betting back.”

Stake-specific · Player-pool reads · Tilt context
Use both Modern AI (Claude, ChatGPT, Gemini) can analyze hands well — it’ll run equity, evaluate sizing, identify leaks, and explain reasoning at any depth you ask for. What it can’t do is know your specific player pool. Use AI for the math and the theory; use forums and a coach for the meta. Together they’re a stronger study partner than either alone.
Build the habit

The discipline. The list.

Three things to do, three things to never do. The hand-review habits that compound — and the ones that quietly waste your time.

Use tools — gut isn't analysis.

Tracking software and equity calculators turn vague feelings into specific numbers. Without them, you're guessing.

Review winning hands too.

Winning doesn't mean playing well. Some of your biggest leaks live inside the sessions you finished up.

Post hands. Help with others.

Other players see your blind spots. Teaching is the fastest way to learn — diagnosing someone else's leak reveals your own.

Don't review while emotional.

Same-day session review is a tilt trap. Flag hands, sleep on them, review tomorrow with a clear head.

Use AI as a study partner.

Modern AI is genuinely useful for hand review — it runs equity, evaluates lines, and explains the why behind each decision. Pair it with forum reads for the player-pool context.

Don't analyze results.

Analyze decisions. A bad call that wins is still a bad call — the wrong-but-it-worked hand is the most dangerous one in your history.

The compounding Hand analysis is the highest-leverage activity in poker outside the table. One hour of careful review typically improves your win-rate more than ten hours of additional play. The session is where you earn it; the review is where you learn how to keep earning it.
The complete habit

Decisions. Toolkit. Conclusions.

01

Decisions, not outcomes.

Analyze how the hand was played, not how it ended. A bad call that wins teaches nothing; a good call that loses confirms the read was right.

02

Use the toolkit.

Tracking software. Equity calculator. Notebook. Three tools convert vague feelings into specific numbers — and specific numbers are the only thing you can fix.

03

Write it down.

Every review ends with a conclusion in a notebook. Over time, this becomes a personal strategy guide built from your own leaks and your own fixes.