Statistics for a winning player

[Sciencekid]

I am interested as a relatively new player as to how many hands one would have to play (using poker tracker or equivalent) to determine whether he or she is a winning player at a certain level.

And also what percentage of sessions profitable players are winning.

[BobK]

If you use a fixed session length and an average table strength. Regardless of running well or poorly (as long as you are playing well). I've found that for limit holdem the long run (1000's of sessions) seems to be about two wins for each loss. With average win and loss sizes roughly equal.

Adjusting your session lengths based on whether you are locking up a win or chasing your losses may change that. I don't do either.

For me, a B&M session is around 6 hrs of play, no matter the number of table changes. 2-3 logical mistakes or impulse calls is enough to make me pick up and go home early. I'm not focusing well.

For NL holdem your mileage may vary.

[kimchi]

I think the answer would be more in the realms of calculating the likelyhood that you're a winning player after a certain number of hands.

I did this for myself a few months ago, which informed me there was a 99.something% chance that I was a winning player after 40kish hands....but I can't remember the link, and I got 16% for my stats module at uni.....

[kimchi]

OK, I found it.

This is from MrWookie:

[ QUOTE ]
Win rate and standard deviation: Well, I guess I'll put something in here about win rate. It's simultaneously the most important and the least important stat. Be happy with anything above 0, naturally, but I'll include how to calculate with what confidence you know you are a winning player. This applies to all poker, not just 6 max, and it has been covered elsewhere, but I'm including it here for completeness. First, get your standard deviation per 100 hands from the Session Notes tab ---> More Detail. Divide that value by sqrt(N/100), the square root of the number of hands you've played divided by 100. That gives you the uncertainty (standard error) of your win rate, i.e., the accuracy to which you know your win rate after having played this many hands. As you can tell, it'll take quite a while to converge to within even as much as 0.1. Anyway, divide your current win rate by your uncertainty. This result, which I'll call Z, can be taken to a normal distribution table, like this one. Enter your Z into the "Below" box, and that's the probability you're actually a winning player. For the lazy ones among us, here are the cliff notes: Z = 1, there's an 84% chance you're a winning player, Z = 2, 98%, Z = 3, 99.999%, and Z = -1, 16%. As you can see from those numbers and a little figuring on your own with your standard deviation, it doesn't take all that many hands to know for near sure that you're a winning player. It should be equally apparent, however, that it's relatively likely for a winning 2+2er to have a negative Z over a reasonable stretch of hands.

[/ QUOTE ]


link for stats entry

[Abbaddabba]

Sessions won/lost can be misleading.

A winning player who plays 10 hand sessions will lose more than half of them

A 2bb/100 winner who plays 1000 hand sessions will win around 80% of them.