Press the exploit button and the chart changes: every hand now shows one recommended action, and a line above the chart says how many big blinds that is worth over playing the spot the way the field does. It is tempting to read this as a set of instructions. It is better read as a measurement.

What the number means

The exploit is the best response to a fixed model of the pool. For each hand you could hold, the model plays out the rest of the hand against the field as the data says it behaves, tries every action, and keeps the one that makes the most. The pool in this calculation does not react to you. It plays the way it always plays, whether you have bet this flop once or a hundred times.

So the number is a ceiling. It is what the spot is worth if you knew exactly how the pool plays and the pool never noticed you knew. Think of it as the answer for your first hand at a table where nobody has seen you before. That is a real and useful quantity, and it is also not something you will collect in full over a session against people who are paying attention.

Why there is no mixing

In field view, most hands show a mix of actions, because the pool as a whole does different things with the same cards. In exploit view, nearly every hand shows a single action. This is not a simplification. Against an opponent who does not adjust, mixing has no purpose. Mixing exists to keep a thinking opponent from reading a pattern; against a model that plays the same regardless, one action is simply best, and that is the one shown.

The exceptions are hands where two actions come out so close that the difference is smaller than anything you could act on. Those cells show the actions split evenly. That is not the model being unsure; it is the model telling you it genuinely does not matter which you pick.

Reading the gap

The comparison line above the chart is the most informative part of the exploit view, more than any individual cell. It shows what the spot is worth if you play it the way the pool does (as field vs field), what the best response makes (full exploit vs field), and, in green, the difference. The green number is the point.

• A large gain over the field means the pool is leaving a lot on the table in this spot. That is where deviating from what you see everyone else do pays the most.
• A small gain over the field means the pool already plays the spot close to its best. Deviating buys little, and the cost of being wrong about the model is not worth it.

A spot with a large gap is a spot worth studying. A spot with a small gap is a spot you can play the standard way and stop thinking about. The exploit view is as useful for telling you where not to bother as for telling you where to press.

Each spot stands alone

The exploit is computed for the spot you are looking at. You arrive at that spot with the range the field would have there, and the calculation starts from that range. If you have already deviated from the field earlier in the hand, say by opening wider than the pool does, your actual range at this spot is different, and the exploit computed for the field's range is no longer exactly yours.

In practice this matters less than it sounds, because most deviations change the range at the margins. But it means the exploits in two consecutive spots are not automatically consistent with each other. A strategy that spans several decisions is a different calculation, and it is on the roadmap; for now, read each spot on its own.

When it looks too good

The exploit is a best response to the model, and the model is not the pool; it is a description of the pool with some error in it. Where the model is wrong, the exploit will find the error and use it, because from the inside a hole in the model looks exactly like a hole in the field.

The usual sign is a value that seems too high for the spot. A common case is bet sizing: if the model's reaction to a large bet is not different enough from its reaction to a small one, the exploit will recommend the large bet with hands that do not deserve it, because in the model nobody punishes the size. That is a fact about the model, not about the pool.

Treat an unusually good number as a question, not an answer. Open the EV tree for the hand, look at which pool responses are driving the value, and ask whether they look like something the pool really does. Often they will. Sometimes the tree will show you the seam.

What to do with it

Knowing the maximum exploit is step zero. Not because you should play it, but because without it you cannot tell how far the standard play is from the best play, and therefore cannot decide how far to move.

The question the exploit answers is "where is the money in this spot?" The question it leaves to you is "how much of it can I take before the table adjusts, and is it worth the cost when the model is off by a little?" Some parts of an exploit survive attention: the pool folds too much to a small bet on some textures and will keep doing it for a long time. Other parts are one-shot: they work until someone notices, then invert. Telling those apart is judgment, and the tool gives you the raw material for that judgment rather than the judgment itself.

Used this way, the exploit view is a map of where the pool is weak. Used as a script, it is a way to become predictable in exactly the spots where the pool is watching. The difference is entirely in how you read it.