Quite often, the first step in learning a new skill is the hardest, and it's no different when you dive into the world of GTO and poker solvers. Despite our constant effort to improve Deepsolver's functionalities and interface, there will always be some complexity to overcome whenever you try to understand crucial concepts of optimal poker strategy.
We've already covered the topic of navigating basic Deepsolver's functionalities, along with taking a look at the GTO vs exploitative strategy debate.
This time, we'd like to offer a few practical tips you should consider whenever you analyze and implement poker solver strategies.
Remember, you play poker versus humans
Forgetting this basic fact is the most common and trivial misconception made by newcomers to the GTO world. They dive deep into simulations, set up optimal (often very wide) preflop ranges, use multiple bet sizes and frequencies with a particular hand, and take them for granted.
In reality, the GTO poker strategy is impossible for humans to replicate, so even the best will deviate from it, let alone your average opponent in low- to mid-stakes games. Despite many poker players improving, the population is nowhere near playing optimally. Remember that so you won't get carried away by theoretical deliberations and overcomplicate your poker solvers' study.
Always take a closer look at the preflop ranges
Every poker hand starts preflop, and that's what you should evaluate first.
How do you define your opponent's range? Is your opponent more on the tighter or looser side? Are they likely to open more or less than the solver suggests?
Answering these questions will help you adjust the ranges and make them as accurate as possible. Eliminating the hands that are unlikely to be present in your opponent's range will also make it easier to draw valuable conclusions.
Which player hit the board better?
When you come up with final ranges, it's time to think about the board. Which player's range hits more frequently? Whose range has the equity advantage, and which player has the range advantage?
Whenever you are about to run a simulation, take a while and guess how the equity of both player's ranges will split; it's an excellent exercise that will sharpen your in-game intuition.
The primary example is the UTG vs BB scenario on a high board, like AAQ, which will favor the player opening the pot heavily. The reverse situation will occur on a board like 765, which is assumed to be an excellent board for the BB player. Of course, most boards fall into the "in-between category", but the more time you spend on this toy game, the better your overall range awareness will be.
Which sizings to choose from?
The natural consequence of the board structure and the preflop ranges are the sizings preferred by the solver. Of course, the fewer sizings are in your strategy, the easier it will be to implement.
With time and experience, choosing appropriate sizings will become easier, but how can you determine which sizing the solver prefers in the particular situation? Start by running a simulation with various bet sizes (in the case of Deepsolver, the max is 5), then rerun it with only one or two preferred sizings. The expected value of both strategies will often be very close, and you will be able to downsize the elaborate strategy into one, including one sizing, making it much easier to implement.
Which hands bet and why?
When you make a bet, you almost always do it either as a bluff or for value. Usually, it's easy to come up with hands that want to bet for value (it's finding a threshold between a value bet and a bluff that can be tricky) since the choice is intuitive.
Things get much more interesting, when you look for hands to bluff with. Whenever you analyze the sim on the flop, take a look at hands that solvers choose to bluff; while some of them may be unclear at first glance, most will fall into a few common categories, like having a lot of good turn cards to continue aggression, blocking value or having so low equity, that they won't ever win at showdown.
Understanding the qualities of hands that solvers recognize as good bluffing candidates in certain spots will make finding them in the game easier, primarily when a good bluffing opportunity manifests itself.
What is your opponent actually doing (and how do they differ from GTO)?
You can safely assume that every opponent you face will differ from GTO in some way. The sooner and more accurately you identify your opponents' tendencies, the more effectively you exploit them.
Let's take the simulation for the 765 board mentioned above in the UTG versus BB scenario. The optimal poker strategy assumes the BB should lead approximately ⅔ of the time. When you're the in-position player, ask yourself how many of your opponents would choose to do so. We bet that only a tiny percentage of the player pool is close to the correct lead frequency.
As a result, you have to include this deviation in your simulations and node lock your opponent's decisions.
In this exact case, if the BB leads as they should, the c-bet frequency for the IP player should equal around 53%. However, if we assume that BB never leads (which strengthens their range significantly), the IP player should continue betting only 10% of their range!
As you can see, a simple adjustment drastically changes the strategy and increases your edge in spots where many players fail to adjust their strategy to the real-life environment.
An unexploitable GTO strategy is one thing; applying it is something completely different
However potent it is, GTO poker strategy should primarily be a guideline, filtered by what's happening in your particular game. Mindlessly obeying solvers' input may improve your win rate slightly, but it won't drastically improve you as a player.
The key to increasing your poker skills is improving your understanding of why some hands should be played a certain way and adjusting what you do to what your current opponents are up to in real time.
The best way to do so is to work with Deepsolver, a modern poker tool with every functionality a proper solver should have.