You're (Probably) Using Solvers All Wrong

Sponsored by

Hello, and welcome to the April issue of the Solver School Newsletter!

In this month's issue, I’d like to confidently declare that most people who use solvers are not using them properly to gain insights that can actually help them improve their understanding of the game and become better poker players.

This is an important topic that isn't discussed nearly enough. In today’s newsletter, I’ll explore it in more detail.

Before we jump in, today’s newsletter is sponsored by OddsJam!

How did @TheArbFather make $10,000+ profit in February?

His secret sauce: Arbitrage betting.

Instead of getting bettings tips from slick-talking handicappers, TheArbFather bets on both sides of an outcome to guarantee a risk-free return. An example of this is betting on a game total to be over 224.5 AND under 224.5.

This might sound too good to be true, but the reason this can happen is sportsbooks set their lines for games independently. Sometimes, they make mistakes and there are situations where FanDuel’s odds are different from DraftKings’ odds.

Unless you are a PhD-wielding, Python-coding, Excel-Wizard… finding arbitrage opportunities consistently has been out of reach. Until now…

OddsJam scans millions of odds every second and finds these need-in-a-haystack opportunities that you can bet on to secure a risk-free profit.

 Try it for FREE now.

Thanks to OddsJam for sponsoring. Now, onto the newsletter…

The Rise of Solvers in Poker

It’s crazy, but solvers have been around for almost a decade!

Solvers have evolved considerably over the years.

From early, niche tools like CREV to the advent of full-functionality applications like PioSOLVER, GTO+, and MonkerSolver to the powerful cloud-based platforms on the market today, like DeepSolver and GTO Wizard — solvers have come a long way.

Let’s state the obvious. With the company name of Solver School, I favor using solvers as a technology for poker study. I believe that the mass adoption of solvers is excellent for the game's evolution. They can teach us much about game theory and how to devise optimal strategies against opponents. For those who are passionate about exploring the depths of this game, such as myself, they are invaluable tools.

However, after watching countless poker analysis videos online, I believe that most people who use solvers are misusing them. Many videos fail to provide the proper context or the necessary caveats as they go through hand analyses. As a result, viewers often draw faulty conclusions and develop detrimental habits when trying to integrate these insights into their own game or when using solvers themselves.

Treating Equilibrium as “The Correct Answer”

The main problem comes from the practice of treating a single solver output as “the correct answer.”

Merely looking at equilibrium solutions and trying to memorize them to build strategies is not new. From the early days of PIOSolver, plenty of individuals took outputs as gospel and tried to incorporate outputs into their strategies.

However, this issue has worsened with the widespread adoption of tools with endless libraries of equilibrium solutions. The true value of solvers comes from running multiple solutions and testing how changes in input variables manifest differently in the resulting strategies. Now that most players don’t need to run solvers themselves, I worry that this valuable process is becoming less common.

When you open up a tool like GTO Wizard, you will find many libraries of solutions that are meant to model the equilibrium of various stack sizes, opening sizes, and table structures (e.g., 100 BB, 6-max, with 2.5 BB opens).

The ranges used to generate these outputs come from preflop solutions that people have developed by constraining the game tree (due to processing limitations). These preflop solutions are then used to generate the post-flop solutions that populate their entire database.

These solutions are great for analyzing general trends, understanding frequencies across aggregate reports, and identifying patterns across different scenarios. However, the critical assumption that often gets lost in translation is that they are based on strategies where all players play a GTO-like, equilibrium-based strategy throughout the game tree.

If you've ever played live poker, you know this assumption is never valid in reality. No one plays perfect GTO poker. Ranges are imbalanced, open sizes vary, players limp and overcall, multi-way poker happens regularly, bet sizing is often exploitable, etc. The resulting ranges you see postflop are often significantly different than those in an equilibrium solution.

In some cases, these deviations don't impact the solver output much. An opponent opening 15% vs 20% of hands may not change the fact that you have a significant equity advantage in certain spots, resulting in your ability to bet frequently. But in many scenarios, even small changes to the construction of a range or its shape can yield drastically different optimal strategies.

So when I see commentators in analysis videos trying to determine if they played a hand "correctly" solely by looking at a solver output, I think that's the wrong approach most of the time. Those analyses don't account for how the solver's equilibrium assumptions differ from the actual gameplay they might experience.

How You Should Be Using Solvers

I’ve spent some time discussing the wrong way to use solvers. So, what's the right way? The real power comes from examining the same scenario over multiple iterations, each time adjusting the inputs and assumptions to identify how those changes impact the outputs.

The levers you can pull include adjusting to look at different boards, altering your bet sizing options, or (most significantly) adjusting ranges and/or node-locking player actions.

I just finished a three-part series detailing how these various inputs impact the outputs. The hyperlinks will take you to Part 1, Part 2, and Part 3.

For example, instead of just looking at a solver's default output for a spot you played and stopping after you examine the equilibrium output, make that your starting point. After you understand the baseline, iterate based on your reads and knowledge of your opponent's tendencies.

What if they called with a much wider range preflop than GTO? What if they don't 3-bet bluff nearly as much as they should? What if you know they overbet certain board textures at a high frequency? Your solver can answer all of these questions with a bit of extra work with your solver.

By adjusting these inputs to match your observations, you can get much more useful information about how to maximally exploit your opponent. Just as importantly, you'll be training yourself to think through these adjustments in real-time as you play.

By the way, you can do all of this in GTO Wizard! You need the Elite tier subscription, but you can use its AI Solve functionality to adjust the game tree and node lock as much as you want.

Remember, even the most die-hard GTO advocates will tell you that playing a pure GTO strategy will not maximize your expected value. GTO play minimizes your exploitability, but to truly maximize your winnings, you need to identify your opponents' mistakes and adjust accordingly.

Closing Thoughts

I hope this gives you a better sense of how to use solvers to explore the game of poker. Tools like GTO Wizard are immensely powerful. However, their real power will be unlocked if you understand the limitations of looking at pre-solved solutions and know how to utilize the node-locking functionality.

If this content resonates with you, check out my flagship course, The Solver Masterclass. It will teach you everything you need to know to become an expert at using solvers to analyze and ultimately improve your game.

You can also check out the rest of my courses. I have several mini-courses — all included in The Solver Masterclass — that detail how to use solvers to build strategies, How to Use GTO Wizard, and starting ranges you can load directly into PIOSolver or GTO+.

You can also follow me on YouTube, where I share more solver-focused and poker strategy insights.

I appreciate you following along and reading the newsletter today.

Until next time!
Michael Lukich