How to Use a Poker Solver Without Wasting Hours

The most common solver mistake is treating it like a lookup table for one hand. You will never memorize ten thousand spots. What you can learn is the pattern that repeats across a texture. This page covers how a solver produces its output, how to read one spot, and how to turn that into something you can use at the table.

How do poker solvers work?

A poker solver takes a defined situation, two ranges on a board with set stacks and a fixed menu of bet sizes, and searches for a pair of strategies where neither player can gain by changing anything. That pair is a Nash equilibrium. The search is iterative: the algorithm plays the two sides against each other over and over, and after each pass it nudges every decision towards whatever would have done better. Early passes are wild; later passes change less and less. When the amount either side could still gain falls below a threshold (the exploitability), the solve is called converged.

The output is a strategy for every hand at every decision point in the tree. Most hands do not get a single action. They get a mix, say bet with some of their combos and check with the rest, because a hand that always did the same thing would be readable and therefore exploitable. Those mixed strategies are the reason solver output looks like percentages rather than instructions.

Two consequences follow. First, the answer is exact only for the tree it was given: change the allowed bet sizes or the preflop ranges and the numbers shift. Second, every action a hand mixes between has the same expected value at equilibrium; that is what makes the player indifferent between them. Both points shape how you should read the output, which is the subject of the rest of this page.

Solver+ packages this as a library rather than a calculator: the spots are solved in advance, so on your phone you browse converged solutions instead of waiting for a solve. The best poker solver app guide covers how a pre-solved library differs from running a desktop solver and what the free tier includes.

Reading a Solver+ spot, step by step

Every solver display shows the same kinds of information; the labels differ. Here is how to read one spot in Solver+ without getting lost in it. No specific frequencies are quoted because they depend entirely on the spot you load.

1. Confirm the spot. Check the positions, the preflop action that got you to the flop, and the board. Half of all misreads come from looking at the right board from the wrong seat.
2. Find whose decision it is. The strategy you are looking at belongs to the player who is acting at that node. On the flop the out-of-position player usually acts first; the in-position player's strategy is one step further into the tree, after a check or a bet.
3. Read the range-level split. Before any single hand, note how the whole range divides between the available actions, and which bet sizes are in play. The size that carries most of the betting range is the headline of the spot.
4. Look at hand classes, not combos. Group what you see: strong made hands, medium pairs, draws, air. Ask which classes bet, which check, and which mix. The explanation is usually one sentence about board coverage or protection.
5. Treat a mix as a mix. If a hand class is split between two actions, both are correct and your choice at the table should depend on the opponent. Do not read the larger percentage as the right answer and the smaller as a mistake.
6. Ignore the EV column for mixed actions. At equilibrium the actions a hand mixes between have equal EV. Comparing them tells you nothing; a difference you see is a rounding or sampling artefact, not an edge.
7. Change one thing and look again. Use the spot comparison tools to put the same board next to a different position or a neighbouring texture. Whatever moved between the two displays is the concept the spot is teaching.
8. Write the one-sentence rule and bookmark it. Solver+ lets you bookmark and annotate spots. The sentence you write is what you will remember at the table; the percentages are not.

Study by texture, not by hand

Load a board class (a dry ace-high flop, a paired board, a low connected flop) and look at what the whole range does. The lesson is usually one sentence: "on this texture the aggressor bets small with everything," or "here the defender check-raises a polarized range." That sentence transfers to every similar board.

Read frequencies, never compare EVs

At equilibrium every action a hand mixes has the same EV, so "line A shows 3.5 and line B shows 1.6" is a misread, not an edge. The real signal is how often each action is chosen and at what size. Frequencies transfer; per-combo EV numbers do not. The range-based bet sizing post shows why the sizing decision belongs to the range, not the hand.

Then drill it

Reading the pattern is step one. Playing the spot against the solved answer until you stop hesitating is where it becomes yours. That is the difference between knowing and doing, and it is the job of a trainer rather than a solver; the best GTO trainer app guide covers which one fits the leak you found.