A solver is a powerful piece of poker software that calculates ‘optimal’ strategies (called solutions) for user-inputted scenarios. The most commonly used solvers include PioSolver, GTO+ and Simple Postflop.
Solvers have changed the game since they became commercially available in 2015, raising the standard of play significantly.
A solver’s solutions contain a lot of valuable insights, but how exactly do solvers work? And is it worth your time to try and remember everything the solver does so you can replicate its strategy?
This article will help you understand the answers to these crucial questions and more.
How A Solver Works
Let’s start off with how a solver works in a general sense.
In short, solvers calculate an optimal strategy based on a handful of options (inputs) which are set by the user. These options are:
These inputs make a solver’s strategic output quite abstract. In a real game of No Limit Hold’em, you can’t possibly know your opponent’s exact preflop range. You also don’t have such a limited number of bet/raise size options, and leading out is always an option. And yet, the solver bases it’s output on these constraints.
You could say that solvers don’t actually solve No Limit Hold’em. But they do solve pieces of No Limit Hold’em, which can help you better understand the underlying mechanics of the game.
(Note: If you usein conjunction with solvers, you can actually get relatively close to the true strategy.)
There is one more important aspect that you need to keep in mind about a solver’s inner-workings…
The solver operates under the assumption that each player knows the other player’s completefrom preflop until the river. It is based on that assumption (plus the constraints listed above) that the solver comes up with an unexploitable (i.e. equilibrium) strategy.
So, if a solver’s solutions are abstract, what good are they? That’s the subject of the next section.
Note: Want to know how to play every hand in every common preflop situation?(for cash games, heads-up and tournaments) when you join the Upswing Lab training course.
How Should You Think About Solvers?
A solver is like a magic genie. You get exactly what you ask for — nothing more, nothing less. If you ask the wrong question (i.e. screw up your inputs), you may not get a useful answer.
In order to get accurate and useful strategic outputs from the solver, your inputs need to be as close to reality as possible. Once you figure that out, you should try to understand why the solver plays certain hands the way it does. (Or at least why playing a certain hand a certain way adds value to the overall strategy.)
When you understand why the solver plays the way it does, you will start to identify patterns/concepts. Since you can’t possibly expect to study every possible flop, turn and river in the solver, internalizing patterns/concepts for the ones you do study is critical.
What types of patterns and concepts am I referring to?
Provided that you don’t screw up your inputs, here are a few types of patterns you can learn from a solver output:
- Betting/raising range morphology (i.e. which hands you should bet/raise).
- Global frequencies (i.e. how often you should bet/check/etc)
- Which hands should be played with a (e.g. checking sometimes while betting the rest of the time)
With these outputs in mind, you can begin your exploration into figuring out the why behind them.
For example, in a ♥ or 7♣ 5♥ 4♦) was the solver’s preferred action with most overpairs.I explored when to check on the flop with overpairs. I ran a bunch of different situations and through PioSolver and found that checking on very connected and middling flops (such as 9♠ 7♣ 5
Here’s a solver screenshot from that article showing the high frequency checking strategy with overpairs (button vs big blind, single raised pot on a 9-7-5 flop):
So, what’s the why behind the solver’s strategy?
These connected flops are great for the big blind because he has so many straights and two pairs in his range. The solver responds by playing more defensively with overpairs.
Additionally, the highest overpairs are checked more frequently than the lower overpairs (i.e. AA was checked more often than JJ). This is because the lower overpairs benefit more from protection since there are more potential overcards that can come on the turn.
Both of these concepts hold true on the vast majority of very connected, middling flops. This is a great example of how you can discover valuable trends by studying a relatively small group of flops in a solver.
The Nodelock Function
If you don’t work with solvers yourself, you can skip this section. But I can’t, in good conscience, write an article about solvers without talking about this key feature.
The nodelock function allows you to lock-in a particular strategy for either player, and then the solver calculates an exploitative strategy against it.
Nodelocking enables you to discover the best possible strategy based on the way your opponents actually play, and you can do a lot of helpful experiments using it.
For example, suppose you are playing in position against an opponent who is both tight and passive. You don’t think she will play nearly as loose and aggressive as the solver does against your c-bet. So, you can do the following with the nodelock function:
- Switch some of her hands to check-calls to reflect her passivity.
- Take some of her weaker check-calls and nodelock them as folds to reflect her tight strategy.
These changes will have a drastic impact on the solver’s solution, which you can study to improve your strategy against this type of player.
By doing a lot of these experiments, you will start to see the patterns behind the solver’s outputs. This will allow you to understand how to make small changes to your strategy when you believe your opponent will play in a certain way. Small changes early on in the game tree have big implications deeper in the.
This tool can help you earn a lot more, but it can also allow you to crash and burn if your assumptions are incorrect.
For this reason, you should be careful about what assumptions you make. They should be less severe the earlier you are in the game tree. This is because people tend to play the flop with smaller deviations from GTO compared to later streets.
To sum up (and for those of you who were to lazy to read the whole article):
- A solver doesn’t actually solve No Limit Hold’em
- It gives the right answer to the question it was asked (that doesn’t mean it was the one you were looking for)
- Your inputs need to be very close to reality
- Focus on understanding why the solver takes the lines it takes
- Use the nodelock function to experiment with different strategies and take note of how the solver adjusts
Now that you know how solvers work, this article will be a valuable read for you:.
Till’ next time, good luck, grinders!
Note: Ready to join 6,000+ players currently upgrading their No Limit Hold’em skills?you will learn inside the Upswing Lab training course.