Mastering Solver Inputs

Hello and welcome to the November 15 issue of the Solver School Newsletter.

In the last newsletter, I wrote about solver algorithms and how they work to generate solutions. Today, I’m moving beyond the conceptual and will outline the inputs required when configuring game trees within solvers.

To generate strategy solutions, solvers require inputs from the user. These inputs allow us to model the poker scenario we want to study as best as possible. The adage of "garbage in, garbage out" applies here — the quality of the solver output depends heavily on accurately configuring the inputs.

The Solver Input Screen

Below are a series of configuration screens for several popular solvers on the market. While each UI may look different, the general inputs required to build a game tree are consistent across these solvers.

The Core Solver Inputs

Solvers require certain inputs to be entered to generate strategic solutions. These inputs include:

• Stack-to-pot ratio - The ratio between the player's remaining stack size and the current pot size. In the solver, you will enter the effective stack and the pot size, but the ratio is what matters.

• Community cards - The cards on the board, like the flop, turn, or river cards. This is needed to generate postflop solutions.

• Bet sizing options - The sizing of available bets on the current and all future streets, building out the game tree structure of strategic options.

• Ranges - The combinations of hands entered for each player, representing what they could hold. More accurate ranges yield better solutions. Ranges can be weighted.

It's crucial to nail these core structural inputs. The stack-to-pot ratio and bet sizing drive the underlying geometry and shape of the solving space. Ranges fill that space with strategic possibilities through potential hand combinations. The bet sizing configuration provides the paths that can be taken to showdowns. Poor inputs here can lead to trees and solutions that fail to reflect reality.

Of these three, ranges have the most significant impact on outputs. The solver can only simulate outputs for the hands we enter, so our ranges directly determine the generated strategy. Unfortunately, this is also the most challenging part to get right. It’s challenging enough to understand your range accurately. No one can assign their opponents with 100% precision. As a result, it’s important to test multiple ranges to see how sensitive the output is to changes in your opponent’s strategy.

Studying and discussing ranges with other poker players will pay dividends in your ability to build accurate range inputs. This is something that comes with repetition and time. Even minor tweaks to ranges to reflect reality can significantly alter outputs and recommended strategies.

Expanding the Inputs for Realism

Most solvers provide additional input parameters beyond these basics, allowing us to model real poker situations more closely. Standard configuration options include:

• Node locking - Fixing frequencies of actions for specific hands at certain decision points. This is extremely useful for testing exploitive ideas or modeling specific players. It can also test specific strategies to identify the best responses and defenses.

• Custom bet sizing - Customizing bet sizing on individual streets to mimic strategy. Adding complexity on earlier streets exponentially grows the game tree size, which can considerably impact processing time and memory requirements.

• Weighting hand combos - Reducing the likelihood of being dealt certain hands in a range, allowing you to fine-tune ranges based on previous actions or player tendencies.

• Game parameters - Factoring in elements like rake, tournament payouts, and chip values that affect EV calculations and make solutions more realistic.

While not always mandatory, expanding inputs makes your solver work more closely mirror your actual gameplay. You can account for real-world dynamics like specific opponents rarely triple barreling or being unlikely to fold top pair. The goal is to minimize the gap between your abstract solver model and what you face at the tables.

Experimenting with Inputs to Derive Strategy

Solver outputs are a direct result of the inputs configured. The same scenario with slightly different input values can produce meaningfully different strategies. The output solution provides a prescription based on the specifics of the input configuration.

This relationship is critical because it allows us to test theories and derive strategies by tweaking inputs. We can answer questions like:

• How does an opponent with a wider range impact our optimal strategy?

• What if we increase our flop bet size from 1/2 pot to 3/4 pot?

• What if we fix a player to only continuation-bet their entire range?

By manipulating input variables like ranges, bet sizings, and frequencies, we can experiment, gain insights, and derive optimal approaches. Making small, controlled changes allows us to isolate the impact of that specific adjustment.

Of course, there are far too many combinations of inputs to test them all exhaustively. Here is where understanding fundamentals helps guide your experiments so you can reach meaningful conclusions efficiently.

Looking Ahead

In next month's newsletter, we'll dive into unpacking and analyzing these strategy outputs that result from configured inputs. Understanding how to interpret solver outputs is the other half of the equation in leveraging these tools effectively.

I’m also considering additional formats for issues of the newsletter, including a quick analysis of a hand via video, text, or a combination of both. If you have a hand or scenario you’d like me to examine using solvers within a newsletter, email me at [email protected].

For more free solver content, check out the rest of the Solver School website. It is filled with video and written content about solvers and analyzing poker with data.

I have several training courses for sale on the Solver School website. If you want to take the next step in using solvers, check out my flagship product, The Solver Masterclass. It contains everything you need to know to become an expert at using solvers to analyze poker data.

You can also follow me on Twitter and 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