How many hands do you need to be comfortable with the data?

[King Yao]

Data:

A: 1,900 hands +1.49 BB/100
B: 2,600 hands +3.44 BB/100
C: 1,400 hands -3.29 BB/100
D: 2,200 hands +5.73 BB/100
E: 9,300 hands +4.81 BB/100
F: 1,000 hands -7.43 BB/100

Total: 18,000 hands +3.21 BB/100

(hand totals may be off because I rounded)

A & B are the same limit. C & D are the same limit. E & F are the same limit.

If we believe that 18K hands is enough to say that I am a +1.50 to +4.00 BB/100 player, then what can we say about the results in C & F? If you were shown those results only, and they had 1400 and 1000 hands respectively, you probably would have thought I was a very poor player. But given the full evidence, you probably wouldn’t come to that conclusion.

My overall VPIP is 41.5 (this may be higher than normal because I have played a decent amount of heads-up and 3 player games). But the VPIP for those two losing limits is 34.1 and 34.8, significantly lower. Maybe this shows I was unlucky with good starting hands in these two limits. Or maybe it means I played more 6-handed games in these limits.

The other interesting number is that in F, I had a Won $ at SD % of only 43.9%, while the overall is 51.6%. Maybe this also reflects that I was very unlucky with the board as well in that limit.

If you only saw the results for C & F, maybe you could guess I was unlucky with bad starting cards (but without the overall VPIP, maybe that’s tough to tell), and unlucky in F with Won $ at SD %. However, it would be tough to conclude that with 15K more hands, the overall BB/100 would be such a big difference. So what does this mean? Does this mean that 1K hands is worthless? Maybe.

Can it be then extrapolated that looking at the results 100 hands of KTo UTG in a 6-handed game is even more worthless? If so, how many hands would we need to really tell if that is a good raisable hand or not? 1000? Maybe, but what if KTo was as unlucky as I was in C & F? 5000 hands? Maybe, but I’m not sure I can get that in an entire lifetime of playing every second of my waking hours.

Now what do you do if you see another players stats of 50 hands at -10 BB/100 with a VPIP of 50. Is that worth anything without actually watching him play? Is it possible just watching him play 5 hands will give you more information about the quality of his play than 50 hands worth of stats? Or even 200 hands worth of stats?

Thoughts?

[Grisgra]

To be clear . . . where do these numbers come from? How did you decide to divide them up?

[King Yao]

sorry for the lack of information.

those numbers come from my own play at 3 different limits. A, C & E are at the same place, B, D & F are at another.

A/B are the highest limits out of all I listed, E/F are the lowest.

[baronzeus]

I think only maybe the 1M+ players can give us even close to solid information on hand values. I just don't think 100 or even 500 instances of an easily varied hand like KTo or JTs is good enough to determine how good the hand is. Then again, I'm probably wrong statistically here.

And I also wouldn't make any judgments on anyone based on under 10K hands. Even 10K hands is pushing, but if I see someone losing something like 4BB/100 over 10K hands, I think that they have some leaks. The same thing holds true for winning players over 10K hands. Unless they are running super hot, it's hard to see someone winning 4BB/100 over 10K without having some strong points to their game.

But less than 2K just doesn't mean anything at all. I've been running hot like the sun recently, and am about 11BB/100 over 2K hands. I know this isn't going to last, and I know I'm not necessarily playing better than during my 8K break-even streak.

[Kirg]

[ QUOTE ]
A & B are the same limit. C & D are the same limit. E & F are the same limit.

If we believe that 18K hands is enough to say that I am a +1.50 to +4.00 BB/100 player

[/ QUOTE ]

I don't really believe you can say this, I think compiling hands and winrates from different levels is pretty useless. Every level does play differently. You might be a theoretically great winner at 3/6 SH, a decent one at 5/10 SH and a tiny loser at 10/20 SH. You can't then add all hands together and use them to see whether you are a long-term winner/loser if you're only playing 10/20 at that point. Unless of course you will be dividing equal time between the 3 levels in correlation to how you played each to get these numbers in the long run too. This is even more true if the 3 different batches come from different games (SH and Full ring, omaha and HE)

[King Yao]

I believe you have a point about the problems of commingling different limits.

[danderso8]

you post this in the shorthanded forum, so I assume they are from short games. I believe the long run in short is even longer than in full games, because the number of hands you play and times you bluff has to go up...in other words

(more hands committed $ to)/100 & (more $ committed per hand played)

will both bring your variance up.

(I have heard arguments that suggest the smaller pot sizes at short nullify this effect, but I don't think that is a big enough factor.)

--dan

[Justin A]

This is why I never put stock into another players winrate that I have stats on. I only look at their other stats.

[helpmeout]

I dont think you can take much from this data.

Since you play a lot of 3handed/HU you know that 18k hands is pretty insignificant.

I dont look at other peoples winrates when I play them because it means nothing. 50 hands of stats is going to give you a general idea at full ring because there are much less hands you can play. 3handed 50 hands means nothing, the hand ranges are much higher so you need many more hands to get an idea of how a person plays.

[sam h]

It depends on the question you want to ask. I think winrates are the hardest to make inferences about. You are definitely right that you need an ungodly amount of hands for this, making accurate assessments of individual hands in certain spots very difficult.

But there is another large problem that doesn't get nearly enough play around here. This is that when people talk about their long-term winrate, what they often fundamentally want to know is what their expected earn is going forward. And this is pretty much impossible to ascertain, because the data isn't reflecting the present, its speaking to the averaged past.

Fundamentally, the data is telling you about your aggregate winrate over the course of a long time, during which time it is not just likely but almost inevitable that game conditions, your resources (PT/PV), and your own ability have changed. Now if you're analyzing a 30K hand sample from the past few months, maybe these changes are not so large and the assumption that the "averaged past" is a good indicator of the present is not so problematic. But let's say you were only comfortable (in a statistical sense) with a sample of something like 500K hands. For many people, this probably brings them back to the days before playerview, when the games were somewhat different, and probably back into time periods when they themselves just weren't nearly as good.

On the bright side, this means that for players who have shown dramatic improvements in skill relative to changes in the difficulty of the games they play, it is likely that their winrate underrepresents their current expected earn.

On the other hand, it is very difficult to ascertain with real confidence what that earn is.