Review five hands.
Leave with one clear idea.

A review means revisiting a decision after your session to understand what makes it good or bad. The aim is to explain your next choice in a comparable situation. You can start with a hand history, a note made after a live game, or a hand reconstructed from the information you have.

Choose a family of situations: defending your big blind, firing a second barrel on the turn or calling a river bet. Write your question before looking at your opponent’s cards. “Which worse hands will call me?” gives you something to investigate; “Why do I always lose this pot?” mixes the decision with the outcome.

Record positions, player count, effective stacks, blinds and antes, known cards and each successive amount. State whether a raise is the total amount or an additional amount. Confusing these changes the pot and the odds. In tournaments, add the remaining places, pay jumps and any bounties.

Our five-hand selection is a suggested way to work. It limits the session to a set you can examine carefully. If two hands need all your attention, stop at two. The amount imported into a tool and the amount actually studied are different things.

Avoid selecting only bad beats. A winning hand can contain an inconsistent bet, and a losing hand can have been played well. Include a decision you thought was straightforward, too: explaining it lets you check whether your reasoning holds up.

The steps below follow the official Analyzer documentation. Start with one compatible hand to verify the process, then import your selection. Labels are shown in English so you can recognise them in the interface.

Importing and analysis are separate stages. An accepted file does not guarantee that every decision has a matching solution. Check each hand’s status before interpreting the displayed values.

EV, or expected value, describes an action’s average value under the calculation’s assumptions. A displayed EV loss measures a difference from a model reference. It is neither the amount lost in that pot nor a sum the software promises you will recover.

The Analyzer documents how hands are matched to available stack depths and bet sizes. Some hands may only be partially covered. If your opponent used a different size, check whether the matching changes the ranges reaching the decision.

Our method: write down your reasoning first, then compare it with the analysis. Had you identified the hands you wanted to call or fold? Does your card block some of those hands? Does your plan assume your opponent continues too often or folds too often?

An adjustment can deviate from an equilibrium strategy. To justify it, write down the observation available when you played and the mechanism you expected. “They showed me a bad hand afterwards” does not retrospectively justify the decision. If information is missing, keep the adjustment as a hypothesis.

An educational example created for this guide: in a cash game, you hold A♠ J♦ on A♥ 9♣ 6♦ 3♠ 2♣. You are heads-up on the river. The pot contains 100 before your opponent bets 50. You have at least 50 available to call, no other player can act and we ignore rake. We do not attribute a solver result to this hand.

You call 50 for a final pot of 200. The equity threshold is therefore 50 / 200 = 25 %. Above this threshold, calling has positive expected value relative to folding under the stated assumptions. Being able to calculate the threshold does not yet tell you your equity: that depends on the opponent’s hands that followed the entire line.

To isolate the reasoning, assume a hypothetical range with 12 weighted combos that beat you and 4 bluffs you beat, with no ties. Your equity is then 4 / 16 = 25 %, exactly at the threshold. With only 2 bluffs against the same 12 value combos, it falls to 2 / 14, or about 14.3 %. These numbers illustrate an assumption; they do not describe an observed frequency for your opponent.

Your review conclusion might be: “My pot-odds calculation was correct, but I had not defined the bluffs reaching the river. Next session, I will reconstruct the range from preflop, then test two bluffing frequencies.” That is a specific task you can check.

Use this worksheet during your next review. For each hand, write enough context to recover the situation, then a question and a note. A good note distinguishes established facts from what still needs checking. You can load our educational example to see the expected level of detail.

The worksheet stays on this page while you type. Download it to keep your work before closing or reloading the tab. It is a study-session notebook: your decisions receive no automatic score.

Choose one observation to explore. Reuse the same positional matchup and a comparable stack depth. Before checking a solution, state your action and explain which hands it targets. Only then compare your answer with the reference assumptions.

Change one parameter at a time: the turn card, effective stack or bet size. Note whether your reasoning changes and why. If you change the format, position and ranges simultaneously, you will not know what explains the difference.

At the end, ask what you can explain more clearly. A session may finish with a question still open. Return to the same situations in a later review to see whether you recognise the problem more easily, without treating a few playing results as proof of progress.

A notebook is enough to organise five hands. A tracker becomes useful when finding, filtering and replaying histories takes too much time, provided it supports their format. An equity tool helps compare a hand with a range; a solver answers a strategic question within a defined model.

GTO Wizard can make supported hands easier to study and connect them to a reference solution. Before selecting a plan, check your format, the function you need and your analysis volume. If you review only a handful of hands, a larger quota adds nothing by itself.

Start by testing the workflow with your own data, then read our subscription guide. For a tournament decision near a pay jump, continue with the ICM guide: the analysis must also account for the payout structure.