OK, What am I missing here?

CardSharpCook · #1 08-27-2005, 02:46 AM

OK, My PT database show that 46.6% of the players (20/40) are winners. Is there any reason why this would be wrong? I understand that gambler's ruin might take out a few would-be winners. However, shouldn't this still mean that some 45% of players are winning players? There is also a question of how much they win, but AGAIN, 45% of all players at this level are winners, right?

ike · #2 08-27-2005, 03:37 AM

I think it should be much less than that, probably 15-25% or so. The problem is that whether or not a player "should" win, over the sort of sample sizes you'll have for most of them, maybe a few hundred hands, often less, its approximately a coinflip whether or not they do win. Does this make sense to other people?

CardSharpCook · #3 08-27-2005, 03:45 AM

At 20/40 the avg player loses .76BB/100. Just to be sure you understand what this means - every hundred hands, the avg player loses .76BB to the rake. Now, the question is, what's the variance? What's the SD?

You may have something with the coin flip theory. But there should be some way to figure this out.

TStoneMBD · #4 08-27-2005, 04:14 AM

of course there is no way to figure out what the true number of winners and losers there are in a game other than looking at historical statistical evidence. its possible that there are no winners in the game if everyone plays at the same level as everyone else and its also possible that there are 9 winners in every game. its possible that 99% of players are winners and they all feed off of the 1% who cap every street and fold to the river bet.

blumpkin22 · #5 08-27-2005, 04:42 AM

[ QUOTE ]
its possible that 99% of players are winners and they all feed off of the 1% who cap every street and fold to the river bet.

[/ QUOTE ]

True, I think this is what's going on.

veganmav · #6 08-27-2005, 07:49 AM

hey by the way i played with you today on the 15/30 game

partygirluk · #7 08-27-2005, 08:14 AM

Imagine a game with 10 players. Every round 5 players are randomly selected to be winners, 5 losers. The winners all gain £10, the losers lose £11. After 1 round, PT type program would show 50% to be winners. After 3 rounds, there are guaranteed to be winners (I believe at least 4). After 20 games, it is still likely that there will be a couple of winners. However, in the long run 100% are losers. The same goes with poker - you don't have enough hands on all your players, you are nowhere near the long run, hence the discrepancy.

krishanleong · #8 08-27-2005, 11:27 AM

[ QUOTE ]
OK, My PT database show that 46.6% of the players (20/40) are winners. Is there any reason why this would be wrong? I understand that gambler's ruin might take out a few would-be winners. However, shouldn't this still mean that some 45% of players are winning players? There is also a question of how much they win, but AGAIN, 45% of all players at this level are winners, right?

[/ QUOTE ]

Your sample size is just too small. Variance rules win rates up to an enormous number of hands.

Krishan

QuikSand · #9 08-27-2005, 01:10 PM

Another matter is that a "player" is not clearly defined in this sort of effort. One fish whom you actualy seek out to play against time after time will be in your database (perhaps with many, many hands) as one player... while ten good players whom you prefer to avoid will show up as 10 players. I don't know what correlation there might be between good/bad players and the length of time they stick with one online handle - but that could be a factor as well.

mosta · #10 08-27-2005, 07:03 PM

[ QUOTE ]
At 20/40 the avg player loses .76BB/100. Just to be sure you understand what this means - every hundred hands, the avg player loses .76BB to the rake. Now, the question is, what's the variance? What's the SD?

You may have something with the coin flip theory. But there should be some way to figure this out.

[/ QUOTE ]

The rake is .75BB/10, not /100, no? 7.5BB/100. and that's only if it stays a 10-player table.

This has come up recently. Note Rudbaeck's posts in this thread :

> My poker tracker shows me that 39% are winning and 61% are losers.

Haven't you been on here long enough to know that these two datapoints actually support each other?

Your session results will form a bell curve around your long term true win rate, the steepness of which depends on your standard deviation. The peak of the curve coincides with your true win rate.

Now picture this in your mind, put the top of the curve at -3BB/100. Notice that a wide part of the curve is in the black?

----------------------------------------------------------------

but I'm still not quite satisfied either.

PS I also see you a lot in 15-30--you play more 20-40 than that?