Poker solvers didn’t appear overnight. Their story spans nearly a century and involves two brilliant mathematicians, nuclear weapons, and an algorithm inspired by playing solitaire. Today, solvers are core tools in every serious poker player’s toolkit—but their origins are anything but ordinary.
The Origins: Game Theory and Poker on Paper
In 1928, mathematician John von Neumann published the Minimax Theorem—the first formal breakthrough in Game Theory. He later co-authored the seminal book Theory of Games and Economic Behavior in 1944 with Oskar Morgenstern, laying the foundation for strategic decision-making in economics, politics, and even card games.
To illustrate imperfect information games, von Neumann used a simplified version of poker: no real cards, just random numbers between 0 and 1, and one round of betting. This paper-based poker game was enough to prove the existence of mixed strategy equilibria—an idea that became the backbone of future solvers.
Enter Stanisław Ulam and the Monte Carlo Method
Stanisław Ulam, another legendary mathematician, grew up in Poland and was a known poker player in his youth. During World War II, he joined von Neumann and other scientists at Los Alamos to work on the Manhattan Project. With little entertainment in the desert lab, they played weekly poker games—sometimes even mid-experiment.
Ulam later devised the Monte Carlo method while wondering how likely it was to win a game of solitaire. Unable to calculate it traditionally, he chose a new approach: play the game many times and collect the data. The more he played, the more accurate his estimate became. Named after the famous casino in Monaco, the Monte Carlo method became a cornerstone of computational science—and eventually poker solvers.
Von Neumann’s Role in Computing
Von Neumann wasn’t just a Game Theory pioneer—he helped design the architecture of the modern computer. These early machines enabled scientists to move from theory to simulation. Ironically, poker wasn’t yet the focus of their computing efforts—they were simulating nuclear reactions. But the logic trees they used weren’t far off from poker decision trees.
One early Monte Carlo application modeled neutron diffusion inside a nuclear bomb. Think of each neutron’s path like a turn in a poker game: branches, probabilities, outcomes. This method changed how problems were solved—and would later power poker software.
The First Computer Poker Tools
While von Neumann and Ulam didn’t use computers to solve poker, their work laid the groundwork. In 1956, Ulam and others created one of the first chess programs, played on a 6×6 board (without bishops) to save memory. It beat beginners but not Ulam himself, who was a strong chess player.
It wasn’t until 1990 that we saw the first equity calculators for poker. These tools used Monte Carlo methods to estimate hand strength—dealing thousands of boards between two hands to get accurate win percentages.
The Rise of Solvers (2013–Today)
The first modern poker solver appeared in 2013, built privately by a developer known as Ol Ostrov and sold for around $200,000 to a few high-stakes pros. Among the early adopters were players like Kanu7 (Alex Miller) and Trueteller (Timofey Kuznetsov). It used Monte Carlo Counterfactual Regret Minimization (CFR), a groundbreaking algorithm for decision-making under uncertainty.
In 2014, Oscar Tammelin introduced CFR+, which accelerated convergence and improved accuracy. This method became the standard across most public and private solvers.
Then, in 2017, a major milestone: Libratus, a poker AI created by Tuomas Sandholm and Noam Brown, beat four top pros—including Jason Les and Dong Kim—over tens of thousands of hands in heads-up no-limit hold’em. It was poker’s equivalent of Deep Blue beating Kasparov in chess.
For more on how Libratus worked, we recommend this episode of the Lex Fridman Podcast with Professor Sandholm.
How Players Use Solvers Today
Most professional players use solvers to:
- Study optimal preflop ranges
- Analyze postflop decision trees
- Improve ICM spots in tournaments
- Train exploitative adjustments for specific player types
From tools like PioSOLVER, GTO+, and Simple Postflop, these platforms allow users to input hands and compute optimal strategies, using the same Monte Carlo-based methods pioneered by Ulam decades ago.
Final Thoughts
Poker solvers are the product of brilliant minds, strange wartime friendships, and simple card games. They represent the intersection of math, strategy, and software—and they’ve forever changed how we approach poker.
What started as a question about winning at solitaire helped shape nuclear physics, computer science, and ultimately, the way we grind today.
👉 Want to dive deeper? Watch our full YouTube video on this topic:
Let us know in the comments: When did you first use a solver?