Quote:
Originally Posted by Bob148
If there's no benefit to the value range ...
This is where "GTO" and Nash Equilibria come into play.
The way to do this analysis is not to start with what people do do, but what they should do. For this river, calculate what percentage of hands should be value bet, what percentage should be bluffed, what percentage should be called, and what percentage should be folded. Then, repeat for every possible river and take a weighted average of all outcomes to calculate what should be done on the turn. And then the flop. And then preflop. This is what the bots do.
In reality, people don't play what they should. And the direction of their change dictates the direction of your adjustment - e.g., if they call too much, then you value bet more and bluff less. But the magnitude of the change is super hard to figure out, because it's not in equilibrium.
Now flip it around. In an exploitative environment neither player is at equilibrium. You can say that a non-equilibrium value bet with no bluffs is worth E{VB}, and that by adding a certain level of bluffs it is now worth E{VB+B} = E'{VB} + E{B} - and calculate the difference between E{VB} and E{VB+B}. But the more apt comparison is E{Nash Equilibrium} and E{VB+B}, because it includes a baseline level of bluffs (otherwise you're comparing to a fictional universe where you never bluff, and even the worst poker player knows that's not true).