This is getting interesting. I'm at work at the moment so can't spend much time on this atm but I'll think about it this evening and I try to write you a detailed response. 
Notifications
Clear all
Just a small disclaimer. The probability of brain farts in this post is moderate.
Well, but with smaller sizing like 2,5bb you are giving 30% pot odds and that gives your opponent really good price to call with wide range.
Okay let's start with this. You are absolutely right that we are giving villain better odds. For example if you just minraise on SB, villain on BB should probably defend with 80% range. But here the key thing is that they simply do not defend correctly. Not vs standard sizings and especially not vs smaller sizings. I have done some population analysis on this and it really seems like our sizing won't affect their ranges much. For example, vs 3bb+ steal BB will fold about 59% according to my analysis. If the sizing is between 2.1bb and 2.9bb (most often they use 2.5bb), then they will fold about 53% of the time. So, by decreasing your open size by 1bb you lose just about 6% in fold equity.
If we just look at this, I think it is insane to risk more money preflop if we can make them fold as much by risking less. Also, our autoprofit threshold with 2.5bb open would be 62.5%. In that case we could open any 2 cards even if we have 0% equity. In reality we always have more than 0%, so if villain folds 53% we are close to a situation where we can just steal super wide. But what if we increase our sizing to 3.5bb? Then we need 70% folds. I believe this makes it worse for us pre.
Okay, then we can make easily autoprofit postflop with cbet. But from my stat cbet F success (34,36) people are overfolding more than 9% of the time (on 1/3cbet), so basically the more money I put into the pot preflop the bigger will be the profit postflop.
Sure, but you risked money preflop to get there.
If you make villain fold immediately preflop, you will get 1.5bb profit right away. If you bet 1/3 in to a 7bb pot and villain folds 35% of the time your EV there is 1bb. If we assume you have 0% equity, you would actually want villain to fold pre instead of getting to the flop where you have a profitable bluff.
Also, because you risked 3.5bb preflop to get a marginally profitable bluff spot OTF, you are losing money from the whole hand. Example: You raise pre to 10bb and villain calls. OTF the pot is 20bb. Now if the pot moves your way 35% of the time, on the long run 7bb of the pot will be yours. But you paid 10bb to get that. Villain can let you make a profitable bet OTF and you are still not able to print a lot of money if you risk a lot of money to get to that spot.
I'm a bit too tired to run different calculations but what will happen when we face a 3bet? If villain keeps his defending range quite similar vs 2.5bb open and 3.5bb open, I assume we will face 3bets quite frequently too. If we open very wide with weak hands we have to fold more and we will lose one extra BB there.
Agree and your calculations are right if the opponents folding similar percents vs 2,5bb and 3,5bb. Well from my experience or I have just some kinda feeling, I used to steal for 2,5bb from SB and I feel like now people fold significantly much than before. Yeah, I know that it can be not that relevant, but I feel like there is a bigger deviation than 7%. But if you are right with the numbers, then why we don't use 2bb or 2,1bb sizing? Then it should be the cheapest and max profitable way.
I think we are talking about stuff on which we really need exact numbers to recognize which one is max profitable. And I mean numbers from such huge database. As default 2,5bb sizing is absolutely fine, but if someone is using 3,5bb I don't mark it as a mistake. We should rather focus and talking about bigger leaks which are more relevant, because I think the deviation in winrate between 2,5bb and 3,5bb raise PF is really inappreciable.