Originally Posted by DenverDave
Given the new features, I wanted to ask something with regard to grades:
How are these grades determined? I am not asking for the exact calcuations mind you as I am sure that would make my head explode. My concern is that I may get a failing grade for something and I feel that this may or may not be correct. And is there any way to look at those hands for that failing grade? And for grades by position?
I want to get better. And even over say 30K hands, I am not raising enough from the CO per LB4, if you look at the hands and the tables I was playing, there could be a good reason for that. And would that be a part of the calculations?
I think understanding how this work will make LB4 more enjoyable and maybe be of benefit down the road!
Some of this is explained in the FAQ on our sites. The calculations themselves are pretty complex, but I'll try and simplify and explain as best I can. So there are stat range leaks, filter leaks, post flop leaks, etc... Stat range leaks (steps 2/3), are determined by peak winrates of players at your stakes based on sample size sustainability. So if most players are winning 3bb/100 within X range, and 2 players are winning at 7bb/100 at another range, leak buster will make an ideal range within the combined overall winrate for all winning players across that particular stat range, BUT it will also consider the two outlier players that are winning. It will not generally recommend that as a profitable range, but if you happen to hit that range, it won't score you low. It may warn you, or it may show it as a potential leak because the difficulty in maintaining that range.
Then based on some of the stat range leaks, it's going to look at specific post flop leaks that relate to it, and score those in conjunction with the stat range. Then based on sample size for all of these leaks, and how each of them impact general winrate, it ranks them hierarchically in order of importance. LB does a bit more than this, but this is the general idea.
To your question... "And even over say 30K hands, I am not raising enough from the CO per LB4, if you look at the hands and the tables I was playing, there could be a good reason for that. And would that be a part of the calculations?"
This is why sample size is important. Over 30k hands, 50k hands, and so on, people on average are going to face similar situations. So all of these numbers converge to form ideal ranges. If you just play say 500 hands, then the question you're asking becomes more relevant. But over enough hands, there's definitive winning ranges and plays for situations that net the highest EV LONG TERM.
A simple example is that we know VPIP is an important stat to winrate. Maybe over 500 hands you could have a VPIP of 45% and win, but over 50k hands, that will be near impossible. And we know, based on analysis, that every stat has ideal peaks and lows. So by analyzing this data we can create suggested ideal ranges, which again, are recommendations, and leak buster will show you that you can go below or above those recommendations profitably too, it's just not generally recommended.
There's a new deep level analysis in LB 4 that takes this to a whole new level and looks at the correlation between specific post flop plays, stats, and sample size. That's even a little over my head to try and explain in detail, but I will say it's pretty cool.
I hope that helps some.