Two questions, not one
Poker strategy conversations often force a false choice: either play “GTO” or exploit your opponents. Strong players do both. They begin with a strategy that is difficult to attack, then make deliberate adjustments when the table provides reliable evidence.
That creates two separate questions for every meaningful decision. First: what is my sound default against a capable, unknown opponent? Second: what should change because of what this specific player—or this player pool—actually does?
If you skip the first question, your adjustment may be a guess. If you skip the second, you can leave money on the table against opponents repeating obvious mistakes.
Balance gives you a floor. Evidence tells you when to reach for more.
GTO in plain English
Game-theory-optimal strategy is commonly used in poker to describe an equilibrium approach: a complete set of decisions that an opponent cannot improve against simply by changing their own strategy. In heads-up, zero-sum models, neither player can gain by unilaterally changing from an equilibrium when the other player stays there.
That does not mean every equilibrium action wins money in isolation, or that there is one magic move for every hand. Strong strategies often mix actions. The same hand might bet sometimes and check sometimes because the entire range—not one hand—is doing the defensive work.
Poker AIs such as Libratus and ReBeL demonstrated the power of strategies built through self-play, equilibrium finding, and real-time search. Their success is a reminder that robust poker comes from protecting the whole strategy, not from memorizing a bag of tricks.
What a balanced baseline gives you
Think of the baseline as your home address. You may leave it often, but you should know where you left from and why.
- A starting point when you have little or no information about an opponent.
- Protection against observant players who look for automatic folds, calls, or bluffs.
- A way to review decisions without letting the result rewrite the story.
- A framework for understanding which hands belong in your betting, checking, calling, and folding ranges.
- A reference point that makes an exploit measurable instead of emotional.
What an exploit actually is
An exploit is a purposeful departure from your baseline to punish a mistake. If a player folds too often, you bluff more. If a player calls too often, you value-bet more and bluff less. If a pool arrives at the river with too few bluffs, you make disciplined folds that would be too tight against a balanced range.
The word purposeful matters. Raising because someone looks nervous is not automatically an exploit. Neither is copying an unusual solver output without understanding the ranges. A profitable adjustment needs a target, evidence, and a reason the adjustment attacks that target.
Every exploit also exposes something. When you bluff less, a sharp opponent can fold more. When you overfold, they can bluff more. The adjustment is worthwhile only while the opponent is unlikely or unable to counter it.
When to stay close to the baseline
Unknown does not mean average. It means unknown. Until evidence arrives, a solid default prevents you from building a huge decision on a tiny sample.
- You have played too few meaningful hands with the opponent.
- The player is strong, observant, or capable of changing gears.
- The pot is large and your read is based mostly on a physical tell or one memorable hand.
- You cannot clearly describe the mistake your adjustment is attacking.
- Your current emotion—fear, impatience, or frustration—is driving the deviation.
When the evidence earns an adjustment
Good evidence repeats. It can come from hands shown down, sizing patterns connected to known holdings, a large online sample, or a stable tendency shared by a familiar player pool. The more extreme your adjustment, the stronger the evidence should be.
Start with the smallest change that captures the mistake. You do not need to turn every river into a bluff because someone folded twice. Add a few more bluffs first. You do not need to stop bluffing entirely against a loose caller. Shift the borderline bluffs toward checks and widen value carefully.
- Overfolder: increase pressure in spots where their range is capped and your line credibly represents value.
- Calling station: value-bet thinner, choose larger value sizes when supported, and remove low-quality bluffs.
- Underbluffer: fold more bluff-catchers when the line and population data both point toward value.
- Overaggressive player: protect checking ranges and allow their extra bluffs to put money in.
The BASE decision method
Run the method before important turn and river decisions. It takes only a few seconds once practiced. If you cannot fill in the evidence and mistake steps, return to the baseline.
After the session, record the adjustment separately from the outcome. “Bluffed because villain folded too much in three prior showdowns” is reviewable. “Bluffed because it felt right” is not.
- Baseline: What would my sound default range do here?
- Available evidence: What have I actually observed—not assumed?
- Specific mistake: Is this player folding, calling, betting, or sizing incorrectly?
- Execution: What is the smallest clear adjustment that attacks that mistake?
Do not confuse solver output with truth
A solver answers the game you gave it: specific ranges, stack sizes, bet sizes, rake, and assumptions. Change the inputs and the strategy can change. A chart or screenshot without its assumptions is not a commandment.
Use solver work to understand patterns. Ask which player has the nut advantage, which range contains more strong hands, what each size accomplishes, and which hands mix because their alternatives are close in value. The goal is not to imitate every decimal. The goal is to make your real-time decisions less fragile.
A three-session practice plan
This process builds two skills at once: a stronger default game and better discipline when departing from it. That combination is more valuable than trying to label yourself a GTO player or an exploitative player.
- Session 1: mark five uncertain spots and write only the baseline you believe applies.
- Session 2: record observed tendencies, but make no major adjustment without repeated evidence.
- Session 3: choose one supported exploit and state it before using it: “This player calls rivers too wide, so I am adding this thin value bet.”
- Review: check whether the evidence supported the adjustment, regardless of whether you won the hand.
Actionable takeaways
- Learn a defensible default before making large opponent-specific deviations.
- Treat exploits as evidence-based adjustments, not instincts dressed up as strategy.
- Require stronger evidence for larger pots and more extreme deviations.
- Make the smallest change that clearly attacks the observed mistake.
- Return toward balance when opponents notice, adapt, or remain unknown.
- Review your reasoning separately from the result of the hand.