Quote:
Originally Posted by iluvAA
Another question:
Does this optimum bluff to value-bets ratio only apply vs. an observant villain (one who will observe and exploit your weaknesses)?
In other words, if you play on an anonymous site (or zoom poker vs. a huge player pool), do you still have to place so many bluffs? After all, no one will notice and exploit you.
So you have population reads right? i.e the population plays x spot y way. In some spots you think the population calls too much, some spots maybe you think they fold too much. And then in some spots you think they call/fold approx correctly.
So lets say you get a river spot and you have the nuts/air (specifically your valuebets always win and never get XR for simplicity sake).
$100 pot, you bet $100 (allin). You'll have 1 bluff for every 2 valuebets right?
So your EV (assuming this spot they call correctly--50% of the time):
So 33% (1/(1+2)) of the time you're bluffing. When* you're bluffing you get called half the time and lose and the other half they fold and you win. So of that 33% of the time your EV with bluffs is (.5*100)+(.5*-100) where 100 is the pot you win when they fold and -100 is the bet you lose when they call.
So for your bluffing section it would look like .33*[(.5*100)+(.5*-100)]=0!
So when you're bluffing, you dont lose, you break even. That is the point of them bluff catching you--to make your bluffs break even.
So 66% of the time you're valuebetting.
They call 50% and fold 50%. (.5*(100+100)+.5*(100)). When they call you win their bet and the pot.
So your valubetting section looks like .66*[(.5*(100+100)+.5*(100))]
Together add them up, and that accounts for the exchange of $ for every situation. Notice that if you didnt bluff, and instead checked and lost your EV wouldn't change! Id recommend doing a math tree of exactly what amt of $ exchanges with what frequency and sub*frequency (of my bluffing range this happens this often etc.)
Sorry for the long winded reply, maybe that helps some folks internalize what's going on and why.