Its been a long time since I played limit but vpip under 24% is now fishy?anything below 24% VP$IP is somewhat fishy
Its been a long time since I played limit but vpip under 24% is now fishy?anything below 24% VP$IP is somewhat fishy
Yes. Or very suboptimal is the better word.Originally Posted by Checkmate
Btw. I just imported like 200k hands from my own and am running Step 4 on blind defense and it has been loading like 30 minutes and I want to go to bed :P. It would be nice with an "Cancel query"-button like I think they have at holdem visions since it takes so much time.
Sound for vid on preflop play is messed up for me - no sound at the beginning, but two different streams later.
And as el helado already mentioned: it's odd that Bryce's recommendations for openraising would be judged as pretty bad under step2.
Thanks for your feedback, it's appreciated. I understand your comments completely. The numbers aren't worked based on what we believe things should be, or the "common" belief of poker players. They are actually gained by running tens of millions through separate a program we have that we call LB cruncher. It creates peak mean ranges, and calculates where the major drop offs in winnrates occur.
I know from having worked on no-limit and limit, there are some stat areas which have been a little surprising to me as well. The ranges are the recommended highest winning ranges, and doesn't mean that something slightly outside of that stat isn't winnable. The way the scoring works is dependent on how much outside of that stat it effects your win rate.
These ranges are not from 5 years back, but are from recently played games within the last several months on sites like Pokerstars, Full Tilt, iPoker, etc... This is the same process that we did for no-limit. Nothing is different, but that doesn't mean we won't review some of the areas you were talking about.In general it appears that you are using ranges etc from 5 years back when people were playing 24/18-style while now most winning players play 30/20. What I would do is to check out the biggest winners at PTR, get HHs from the topp 20 winningest players and create an average. And the result would be something very different from these values.
Its kinda hard to pay for a product (with a supergood intent and that could be amazing) that are outdated even with such basic numbers as Vp$IP/PFR.
It should not take that long for such a small sample size. We can add a cancel, but in DB's of 3mill+ hands on reasonable computers it should take less than 5 minutes. Can you try running it again another time, let me know because we may want to look at your log file.
Actually you did find a bug in the high range for 6-max, and not sure how it was missed. It's supposed to be 34.1.
I'll post the actual ranges here, this isn't guess work or opinion, it's just the real winnrates of players in these ranges for 6-max (all recent 100% farmed data).
The computer is old (1 gb ram and 3,5 ghz single core) but it took a looooong time. Cancel-option is always good.
About the values. Allright if the figures was wrong then sure. But its not statistically correct to group ALL players with certain ranges and compare, or at least its not optimal. If you group winrates based on vp$ip alone then obviously a bad 30/0 would influence the good 30/20-player. And certain groups that doesnt have that wide of a spread (say vp$ip 18-22-people) then there is not the same mix of playertypes within the range.
What would be more correct (again) would be to get the most winning players (long term) and get some sort of average from that, even thou that has its flaws. And if you look at PTR and look at the biggest longterm-winners then you will get a VP$ip of 28-34.
Do you have any comments on the other figures i was talking about (except the vp$ip that was a bug)?
Btw, i like the product but I would prefer values that is considered correct in terms of winning players to be the measurement rod. Sorry for the bad english btw, im in class.
Yeah we will add it. During testing we never had it take more than 5 minutes on any system, so didn't anticipate it would be an issue. Even though your system is older, your sample size is small enough that it should be ok. I'd recommend vacuum analyzing your DB and it should speed things up for you quite a bit (all of your stats).
I understand what you're saying, and the way the ranges are constructed is actually pretty complex. I can't speak directly about it, but like I said there are some areas I'd be a bit surprised. There are some players that can crush the limit games with very low vpips (you'd be surprised). That being said we know nothing is going to be initially perfect, and we do diligently work on making necessary changes to improve the quality.About the values. Allright if the figures was wrong then sure. But its not statistically correct to group ALL players with certain ranges and compare, or at least its not optimal. If you group winrates based on vp$ip alone then obviously a bad 30/0 would influence the good 30/20-player. And certain groups that doesnt have that wide of a spread (say vp$ip 18-22-people) then there is not the same mix of playertypes within the range.
What would be more correct (again) would be to get the most winning players (long term) and get some sort of average from that, even thou that has its flaws. And if you look at PTR and look at the biggest longterm-winners then you will get a VP$ip of 28-34.
As far as your other stats, we'll go back through the areas you were talking about, and we're always open to comments. I can't say if any things else will be changed simply because we are basing the ranges on massive amounts of data.
No problem. Feel free to let us know anything else you think about it as you get more time to play around with it.
Btw, i like the product but I would prefer values that is considered correct in terms of winning players to be the measurement rod. Sorry for the bad english btw, im in class.
Okay. I play like 40-70k hands per month and datamines every major site that has mid/high FL SH-action and when I get to the office then I could probably get some examples for you. If there is a lot of players with low vpip with a nice winrate then
A) They are either playing at microstakes or
B) They dont have a large enough samplesize that will give a winrate that is close to their true winrate. I dont have the figures in my head but I believe your true winrate with like 1 bb/100 after is like +/- 3BBs after 50 000 hands so their true winrate would probably be within +4 to -2 BB/100 and if you are making comparisons where the difference is like 0.7 bb/100 and 1.0 bb/100 then you can draw your own conclusions.
Anyways. Philuva is probably the only big winner that is tight and have a low VPip (i think he is around 25-26 vp$ip for 5-6handed) that I can think off.
Posting Permissions
