I am combing through my data and over a pretty big sample (165k hands), I am winning at approximately 6bb at my low stakes games.

The one glaring stop where I see to bleed money is the call vs 3-bet stat. In case the stat is not clear, this is anytime I defend a 3-bet either OOP or IP. I am trying to dig through this and see how to improve it, but it seems that it is almost destined to be a losing stat. What is achievable in this regard?

I also combed through some data and I am losing when I defend the BTN vs the BB at around -700bb/100. I am defending a tight range and I ran an aggregate analysis of the equity (I play on a site with all hole cards revealed) and the average pot odds I am being offered are 2:1 (needing to realize 33% of equity to breakeven). The aggregate equity I had when I defended was about 40%. So even with rake of 5%, I should be more than fine to call here. Especially given that I am defending in position.

I just wondered what is realistic in these spots and if I should temper my expectations when I look at this stat.

can you exclude the possibility that you are doing something wrong in your analysis? maybe filter for the stronger parts of your ranges and see how they are doing. also keep in mind that you only have to beat the ev of folding which is like -200bb/100 for a minraise.

sorry if this is obvious to you, but i don't know how experienced you are with data analysis.

Yes

First of all, just because we have 40% eq doesn't mean we can call a 40% pot odds bet, or whatever we need after rake. It's a huge mistake to think that we can.
The population at low stakes tends to have very obvious flaws. They 3bet and open the BTN with ranges way tighter than they should. We can't respond to those bets with some kind of MDF because they are simply not bluffing enough. So maybe look at what you are getting 3bet and stolen by. They should be opening around 40% of the button and 3betting around 10% of hands before I go into defense mode. Any less and I'm exploitative folding every time they do that

Calling a 3-bet is destined to be a losing proposition, yes. But your aim when you press CALL is to finish the hand with a stack size that is better than it would be if you just folded to the 3-bet.
i.e. If you open for 2.5x, then when you call a 3-bet, you want your average loss rate to be -250bb/100 or better. If it's -700bb/100 over a large sample, you're calling too much, or playing badly in 3-bet pots.

Here's a snippet from one of my databases and it shows that, overall, I could probably call even more 3-bets, as I only lost 1.11bb per hand in this sample:



I may well have a problem in the SB, however, as I was losing quite badly (-5.4bb per hand) when I called 3-bets in SBvBB over a tiny sample. If your number is -700 for BTN, you almost certainly have issues that need addressing. Make sure you use the EVbb/100 stat to remove some of the runbad/variance. Note that you really need a much bigger database than mine or yours to come to some firm conclusions. (My numbers are from 125,000 hands, which is nothing really).

It should be noted that the standard deviation in this filtered situation-- calling 3bets --- is going to be very high and as a result looking at your winrates in sample sizes like you are and trying to come up with conclusions is very misleading.

This is the most important thing to consider, IMO. It's very likely that your analysis means nothing, OP. It doesn't mean that you shouldn't look for leaks in 3-bet pots. In fact 3-bet pots BU vs. blinds happen so much that it's well worth studying. However, your sample size is likely not significant.

I wish I had PT4. I was thinking this same thing pretty recently.

If I RFI with a high suited connector which is going to miss a lot then get 3 bet and call I am probably putting myself in a bad spot without strong post flop play. Same thing with lower and middling pairs looking to set mine. I do that 8 times to win once and then I need the villain in that one instance to offer the reverse implied odds I need. Otherwise I am just winning a pot basically. Then account for villain folding maybe 1 out of 3 of the remaining times they missed the flop or turn with overcards.

I'll tell you that you should fold significantly more than you currently do OOP to 3bets as pretty much every single person's database I look at calls too much