GTO Poker Solver: How It Works and How to Use It
Last updated: May 13, 2026
A GTO poker solver is software that computes the Nash equilibrium strategy for any Hold'em spot — finding the mixed strategy that cannot be exploited regardless of what the opponent does.
Modern solvers (PioSOLVER, GTO+, Solver Free) produce strategies with less than 0.5% exploitability. Licenses run $50–$100 for entry-level to $250 for a full professional build. Every solve requires a starting hand range for both players, a board texture, and stack and pot sizes as input.
This page explains how solvers work, the 5-step workflow for extracting actionable insights, the limits of solver outputs, and how to apply solver findings without a license using free resources. For the underlying theory, see GTO fundamentals and hand ranges.
What Is a GTO Poker Solver?
A GTO (Game Theory Optimal) poker solver is a piece of software that calculates the Nash equilibrium strategy for a heads-up postflop situation in No-Limit Texas Hold'em. You provide the inputs — both players' hand ranges, the board texture, effective stack size, pot size, and which bet sizes to allow — and the solver returns the strategy that is mathematically unexploitable given those inputs.
Unlike a simple equity calculator that tells you win percentages for two specific hands, a solver works with entire ranges of hands simultaneously. The output is not "bet with this hand" but rather "bet 67% pot with AJ on this board 35% of the time, check 65% of the time." This mixing makes the strategy resistant to exploitation.
Solver exploitability = max profit a perfect opponent can gain vs your strategy
Target: <0.5% of pot per hand (near-perfect Nash equilibrium)
Solvers are primarily a study tool, not a real-time aid. You run them away from the table to understand how balanced, unexploitable strategies look in specific spots, then internalize those patterns for in-game decisions.
How Solvers Work — Nash Equilibrium and Iterative Algorithms
Most modern solvers use counterfactual regret minimization (CFR) or a variant. The algorithm works iteratively:
Both players start with uniform random strategies across all actions at every decision node.
For each iteration, the algorithm calculates how much each player "regrets" not having taken a different action — the counterfactual value of actions not taken versus the strategy actually played.
Actions with positive regret (would have been better) get more weight in future iterations. Actions with negative regret get less weight. Over thousands of iterations, strategies converge.
When neither player can improve expected value by unilaterally changing strategy, the Nash equilibrium is reached. The solver stops when exploitability drops below the target threshold.
This process is computationally expensive for complex trees (many bet sizes, deep stacks) but fast for simplified scenarios. A basic flop spot with 2–3 bet sizes solves in seconds on modern hardware; a full 3-street tree with many sizes can take minutes.
The 5-Step Solver Workflow
A disciplined workflow separates players who extract real insights from those who get lost in data. Here is the standard process:
Define Ranges for Both Players
Input the range of hands each player can hold at this point in the hand. Use real ranges from your preflop charts — garbage ranges produce garbage solver outputs. Start with common spots: BTN vs BB single-raised pots.
Set Stack Depth and Pot Size
Enter the stack-to-pot ratio (SPR) and current pot size. Solvers are sensitive to these inputs — a 10bb SPR spot will have a very different solution to a 30bb SPR spot even with the same board and ranges.
Select Bet Sizes to Allow
Define the bet sizes you want the solver to consider (e.g., 33%, 67%, 100% pot). Fewer sizes = faster solving. Beginners should start with 2–3 sizes per street. More sizes increase accuracy but exponentially grow the tree.
Run the Solve and Wait
The solver iterates using counterfactual regret minimization (CFR) until exploitability drops below 0.5% of the pot. This takes seconds to minutes depending on tree complexity and hardware.
Study Outputs by Board Texture and Action
Filter results by action — check node, bet node, call node. Look for patterns: which hands bet, which hands check-raise? Study 5–10 spots per session rather than trying to memorize every frequency.
The stack-to-pot ratio matters enormously for solver outputs — see SPR strategy for how stack depth shapes postflop decisions.
Reading Solver Output — Mixed Strategies and Frequencies
Solver output is expressed as action frequencies across your entire range at each decision node. Understanding how to read it is essential to extracting value:
Range frequency
% of total range
E.g., 'Bet 67% with 55% of range' — the majority of your range bets this sizing.
Hand frequency
% for specific hand
E.g., 'AhKh bets 100%, A2o checks 70%' — different hands play differently in a mixed strategy.
EV difference
bb per 100 hands
How much EV is lost by deviating. Small EV differences mean hands are close to indifferent.
Exploitability
% of pot
How much the current strategy can be beaten by a perfect counter. Target <0.5% for production use.
Solver outputs directly inform MDF from solver outputs and connect to equity realization in solver spots. Understanding EV and solver decisions helps contextualize why certain frequencies are chosen.
Free Solver Resources (No License Needed)
You do not need to spend $250 to start learning from solvers. These free and low-cost tools cover the most common study scenarios:
PioSOLVER
Serious students & professionals — deepest analysis, largest tree support
GTO+
Mid-stakes regulars — accurate outputs, lower cost, faster tree building
Solver Free
Beginners & casual players — no install needed, limited tree depth
Simple Postflop
Recreational learners — simplified UI, built-in prebuilt ranges
Tip: Start with prebuilt databases
Many coaches and training sites publish prebuilt solver databases (GTO Wizard, Upswing, Run It Once). These are solved trees you browse without running any software — ideal for learning patterns before investing in a solver license.
Applying Solver Insights at the Table
The goal is not to play exactly like a solver — it is to build intuitions from solver study that improve real-time decisions. Here are the most practical applications:
Identify your bluffing hands
Solvers reveal which hand categories make good bluffs (good blockers, low equity vs calling range, high card removal). Learn the structural logic, not specific combos.
Learn bet sizing incentives
Solvers show when to bet small (polarized boards where large sizes are incentivized for nuts) vs. large (boards with big range advantage). Apply these size selections systematically.
Understand check-back vs. bet thresholds
Solvers define which hands should check-back the flop in position. Learning these thresholds reduces EV leakage from auto-betting every strong hand.
Calibrate your value-to-bluff ratios
Solver outputs confirm the mathematically correct ratio of value bets to bluffs for each sizing. Use these as anchors when constructing your betting ranges.
The Limits of Solver-Based Play
Solvers are powerful tools, but they have meaningful limitations every serious student should understand:
GTO ≠ maximally exploitative
Against opponents with known leaks (over-folding, over-calling), departing from GTO and targeting the specific leak is more profitable than playing unexploitable.
Garbage ranges in = garbage solution
A perfectly solved tree built on incorrect ranges gives you the wrong strategy. Range construction is as important as the solve itself.
Multiway spots are unsolved
Most solvers only solve heads-up postflop trees. Real multiway pots require heuristic adjustments — see the solver output as a directional guide only.
Human memory is limited
You cannot memorize every frequency. The practical benefit comes from pattern recognition and principle extraction — not rote memorization of mixed strategies.
The most effective players combine solver study with live reads and exploit-focused adjustments. Use solvers to build your default strategy, then deviate based on observed opponent tendencies.
Key Solver Terminology
Frequently Asked Questions
What is a poker solver?
A poker solver is software that computes the Nash equilibrium strategy for a given Texas Hold'em situation. Given two players' hand ranges, a board, stack sizes, and allowed bet sizes, it finds the mixed strategy that is least exploitable regardless of what the opponent does. The output is a set of action frequencies for every hand in the range — not a single 'best' action.
How do GTO solvers work?
Modern solvers use counterfactual regret minimization (CFR) or similar iterative algorithms. They start with random strategies, then repeatedly calculate how much each player regrets each action and update accordingly. After thousands of iterations, both players converge on strategies where neither can improve by unilaterally changing — the Nash equilibrium. Solvers stop when exploitability drops below a target threshold, typically <0.5% of the pot.
Do I need a poker solver to improve?
Not necessarily — especially at micro and low stakes. Most opponents at those stakes have significant exploitable leaks that a deviation from GTO will beat more effectively than playing unexploitable. However, solvers become increasingly valuable at mid-stakes and above, where opponents are harder to read and exploit. Free web-based solvers let you experiment with common spots without any cost.
What is the best poker solver for beginners?
Solver Free (free, web-based) is the best starting point. It requires no installation, handles typical postflop spots, and lets you experiment with ranges and bet sizes at no cost. Once you've developed a workflow and are solving spots regularly, upgrading to GTO+ (~$100) or PioSOLVER (~$250) gives you more tree depth, faster speeds, and scripting capabilities for batch analysis.
How do I use solver outputs at the table?
You can't memorize every frequency — that's not the goal. Instead, study the patterns and principles the solver reveals: which board textures favor the in-position player, which hand types prefer checking versus betting, when over-betting is incentivized. Build mental models from repeated solver study, then apply those models at the table. Many players keep a 'solver journal' of insights organized by spot type.
What are the limits of GTO solvers?
Solvers find unexploitable strategies — not maximally exploitative ones. Against a specific opponent with known tendencies (e.g., folds too much to flop bets), the GTO strategy may be significantly less profitable than a targeted exploit. Solvers also assume both players play optimally with the given ranges, which is rarely true in practice. Additionally, entering incorrect ranges will produce a perfectly solved solution to the wrong problem — garbage in, garbage out.
Related Topics
Calculate equity before you study a spot
Enter hole cards and board — RiverOdds gives you real-time equity data so your solver sessions start with accurate numbers.
Open RiverOdds Calculator →