20% of online players win, why do so many say only 10% do???

[Rubber Soul]

I am curious as to why so many people have been saying that only 10% of all players win (online). I have seen several published reports from poker tracker stats that state about 20% of all online players are winners... Now, there may be only 5% who win big, but I think 20% is the actual number of those who win... Some reporter at ESPN claimed he talked to poker site operators off the record at the WSOP and they said only 7% win... That just doesn't sound right...

[Paul2432]

Using poker tracker to determine the percentage of winning players can lead to sampling errors.

I have always heard 5-10% of players are winners.

Paul

[jman220]

Why would the poker sites lie to the ESPN reporter? If anything they would want to exaggerate how many winners there are, not downplay it. And pokertracker is not a reliable indicia of how many players are winners.

[eric5148]

[ QUOTE ]
Why would the poker sites lie to the ESPN reporter? If anything they would want to exaggerate how many winners there are, not downplay it.

[/ QUOTE ]

True, but they would also not be counting rakeback and bonuses.

[iluzion]

[ QUOTE ]
[ QUOTE ]
Why would the poker sites lie to the ESPN reporter? If anything they would want to exaggerate how many winners there are, not downplay it.

[/ QUOTE ]

True, but they would also not be counting rakeback and bonuses.

[/ QUOTE ]

Well, hes comparing the said number to PT stats, which also wouldn't include these numbers.

[dealer_toe]

It might even be lower than the 10% reported. That doesn't account for a winning player having accounts at multiple sites. So its actually lower.

[MicroBob]

[ QUOTE ]
they said only 7% win... That just doesn't sound right...

[/ QUOTE ]


[ QUOTE ]
I think 20% is the actual number of those who win.

[/ QUOTE ]


if it just doesn't sound right to you and you really THINK that it's more like 20%...well...that's good enough for me. it must really be 20%.

[PokrLikeItsProse]

There is actually a reasonable way to reconcile these two data points.

1) The 20% figure covers cash game players at levels where someone will use PokerTracker. It is quite plausble to suggest that there are a greater percentage of losers playing 0.02/0.04 limit hold em on some site compared to 10/20, but that PT will oversample higher stakes players. Also, the rake effect makes more marginal players losers at lower stakes and there are more low-stakes players than mid- and high-stakes players.

2) Presumably, those PT numbers do not include tournaments. To take one site as an example, PartyPoker seems to pay out to about 10-13% of the field under most cases. What percentage of tournaments-only players do you think end up positive over the course of the year? I would say that somewhere less than 10%.

3) Some players who could be winning players aren't sufficiently bankroll and normal variance causes them to go bust and end up in the red because they can't handle normal swings. For example, a player deposits $100-200 a month and chooses to play 50max or 100max NLHE online at least once a week, unless he is broke. This player will always be in danger of going bust because he isn't properly bankrolled for the level he is playing. Because there always exist some players who will stupidly play above the their bankroll, the number of players who have a natural positive win rate should always be greater than the number of players who are positive over the course of a year.

[stone_7]

I think there is another Huge point that is being missed in this ongoing discussion. Namely that the number of hands played by winning players will always be greateer than the number of hands played by losing players. Take the people on this site who have 800,000 hands at 2/4. SInce they are winning they can continue. SOmeone who loses .5bb/100 or slightly less than break even will have busted out their 300 bb bankroll in 60,000 hands meaning that a properly bankrolled slightly losing player will have busted out 13 times over before they play this many hands. Therefore 20% of hands are played by winning players but not 20% of players are winning players. These are 2 very different concepts and it is important to recognize this distinction when trying to understand these numbers.

[Snoopy76]

Also.....

The ESPN article stated that online sites reported only 7 or 8% of their players end up in the red. I would say that close to 50% of the people who play online are people who 'give it a try' deposit 20 bucks..lose it all and never come back. Based on that logic, of players who play regularly, about 14-16% are winners. Not a scientific theory, just a thought.

[Guthrie]

As long as I'm in the winning group and the losing group keeps playing, I really don't care what the percentages are.

[Uppercut]

[ QUOTE ]
Also.....

The ESPN article stated that online sites reported only 7 or 8% of their players end up in the red.

[/ QUOTE ]

Not to nitpick, but ending up in the red is a BAD thing, i.e., a loss. Winning players end up in the black.

[Snoopy76]

But you did nitpick.....hope that made you feel better. You know what i meant.

I'd say that its pretty clear that while the NUMBER of players who win might be fairly low - between 7 and 14%, the VOLUME of hands played by the winning players is significantly higher than that.

[Webster]

Also - what do you consider a winning player. If a guy plays 500 hands in his lifetime and is in the Black is he a winner? Sure, but is he a winning players? not enough time to tell.

If you look at players with 20,000 or more hands you are still in the 50% area on whether they are winners or losers. Think of ALL the players that put in $200, play 200 hands and never play again after they bust out.

[WhiteWolf]

I think the reason is that we don't have a large sample size for most players in our PT databases. A small sample size means we will see a bigger deviation from the player's actual results, making some losers look like winners and some winners look like losers. Since most players are losers, we will misclassify more long-term losers as winners than the other way around, inflating the winning player percentage in our PT databases.

Here's a simplified scenario that illustrates this (warning, math follows):

I have 1000 players in the database. 90% of the players are long term losers, 10% are long-term winners. However, I only have 1000 hands on each players. Because of my small sample size, there is a 40% chance I will misclassify a long-term loser as a winner, and a 40% chance I will misclassify a long-term winner as a loser. Here are the results I would get:

Classified as Winners =
(.4 x 900 losers) + (.6 * 100 winners) = 360 + 60 =
420 winners, vs the true number of 100.
Classified as Losers =
(.6 x 900 losers) + (.4 * 100 winnners) = 540 + 40 =
580 losers, vs the true number of 900.

[Rubber Soul]

[ QUOTE ]
[ QUOTE ]
they said only 7% win... That just doesn't sound right...

[/ QUOTE ]


[ QUOTE ]
I think 20% is the actual number of those who win.

[/ QUOTE ]


if it just doesn't sound right to you and you really THINK that it's more like 20%...well...that's good enough for me. it must really be 20%.

[/ QUOTE ]

Because of the several poker tracker stats I have seen that show around 20%... It doesn't matter either way of course, I'm just curious as to what the real number is...

-

[intheflatfield]

Yes, this is true, however, for every player that plays at that level there more than enough losing players that continue to play at a losing level just for entertainment value, plus The number of new players at sites are likely to continue to increase at such an rapid level as to make the number hands played by the few people you are referring to statiscally insignificant. I don't have anything concrete in the way of numbers, but I don't think I'm too far off base.

[Cosimo]

[ QUOTE ]
Classified as Winners =
(.4 x 900 losers) + (.6 * 100 winners) = 360 + 60 =
420 winners, vs the true number of 100.
Classified as Losers =
(.6 x 900 losers) + (.4 * 100 winnners) = 540 + 40 =
580 losers, vs the true number of 900.

[/ QUOTE ]

Good stuff. I was about to do the same thing.

Further: you'll only see a fraction of the players at your limit. I posit a strong correlation between number of hands played and winnerness; losers don't generally multi-table, they play only a few hours a week, and they might play for a few weeks then give up. Therefore, you'll see a larger fraction of the winners in your DB but a smaller fraction of the losers. Losing players come and go without PT ever picking up a hand on them, but the winners are more likely to stick around and get seen by your database.

[Rubber Soul]

[ QUOTE ]
I think there is another Huge point that is being missed in this ongoing discussion. Namely that the number of hands played by winning players will always be greateer than the number of hands played by losing players. Take the people on this site who have 800,000 hands at 2/4. SInce they are winning they can continue. SOmeone who loses .5bb/100 or slightly less than break even will have busted out their 300 bb bankroll in 60,000 hands meaning that a properly bankrolled slightly losing player will have busted out 13 times over before they play this many hands. Therefore 20% of hands are played by winning players but not 20% of players are winning players. These are 2 very different concepts and it is important to recognize this distinction when trying to understand these numbers.

[/ QUOTE ]

Excellent post... Makes perfect sense...
-
Also, I should have stated in the beginning I was talking only about cash game players. I truly believe (as others have said) that less then 10% of tournament players are winners...

-

[Rubber Soul]

Now, the next question is, if only 10% or so win overall, what would you speculate the winning percentage is for regular 2+2 members (for online cash games at various levels)?

-

[rgschackelford]

It's a combination of things. Mostly, people who go out of their way to keep records on poker trackers are probably taking poker seriously. They aren't playing for the hell of it, they are looking to make money. Second, I would figure that people like to lie to themselves about wins and losses, and not count some losses, or exaggerate wins. Don't look at me, though. I once finished down 40 cents and marked it down. Third, keeping records will help you get better in choosing time of day to play, games to play in, limits to play at, and other tangibles. I'm, honestly surprised that poker trackers don't have more than 20% winners. I would expect 50% winners. Well, that's my 2 cents.

Rusty G.

[Victor]

according to my pt at 10-20short 43% are winners.

[Paul2432]

[ QUOTE ]
[ QUOTE ]
I think there is another Huge point that is being missed in this ongoing discussion. Namely that the number of hands played by winning players will always be greateer than the number of hands played by losing players. Take the people on this site who have 800,000 hands at 2/4. SInce they are winning they can continue. SOmeone who loses .5bb/100 or slightly less than break even will have busted out their 300 bb bankroll in 60,000 hands meaning that a properly bankrolled slightly losing player will have busted out 13 times over before they play this many hands. Therefore 20% of hands are played by winning players but not 20% of players are winning players. These are 2 very different concepts and it is important to recognize this distinction when trying to understand these numbers.

[/ QUOTE ]

Excellent post... Makes perfect sense...
-
Also, I should have stated in the beginning I was talking only about cash game players. I truly believe (as others have said) that less then 10% of tournament players are winners...

-

[/ QUOTE ]

I believe in the long term, less than 1% of players are winners. Consider a multi-tabling pro who plays the 15/30 game. Suppose this player plays 8760 table-hours per year, or one table-year per year. Over the course of a year the house will rake around $800,000 from this table ($100/hr) and the pro will win $250,000 ($30/hr). The rest of the table must lose $1,050,000.

I think a fairly reasonable assumption is that the average losing player loses no more than $5000/yr. If this is correct, it would take over 200 losing players to sustain one pro.

Of course, at any given moment the percentage of winning players is much higher.

Paul

[tripdad]

i know this is probably a dumb question, but how can pokertracker be so wrong about this? i have a pretty sizeable database, and from all limits, PT says that 42.4% are winners, and 57.6% are losers. now, i know that i don't have a huge amount of hands from a single player, but it still seems like the number should be a lot closer to what you are all saying (that 20% or less are winners).

having said that, i do have a player in my database who i know for a fact is a winning player, as a loser over 1800+ hands. so, i know the numbers can be misleading, but i just wouldn't think they would be THAT misleading.

cheers!

[KyleM]

well, lets say that hypothetically you only sat at tables with 7 random players, and 2 players that only pressed "called" (they called to the showdown 100% of the time). your database would should the majority of players as winners, because the majority of players will win at tables with these 2 terrible players. your database will be affected by your table selection, and your own play (if you were a huge loser it would also show an inflated number of winners). also, your sample size could be too small.

[jtr]

Tripdad, this old thread (http://archiveserver.twoplustwo.com/showflat.php?Cat=&Board=&Number=1390393&page=&view =&sb=5&o=&fpart=) answers your question I think.

Exactly the same sort of question starts off the thread: "Hey, my PT database says 40% of people win, what's up with the 10% estimates I hear about?" About ten posts down I chime in with some statistical and simulation analysis to show exactly why the PT number can be so misleading. Basically it's the dread problem of small sample sizes once again.

Hope you find the older posts useful.

[Zetack]

[ QUOTE ]
i know this is probably a dumb question, but how can pokertracker be so wrong about this? i have a pretty sizeable database, and from all limits, PT says that 42.4% are winners, and 57.6% are losers. now, i know that i don't have a huge amount of hands from a single player, but it still seems like the number should be a lot closer to what you are all saying (that 20% or less are winners).

having said that, i do have a player in my database who i know for a fact is a winning player, as a loser over 1800+ hands. so, i know the numbers can be misleading, but i just wouldn't think they would be THAT misleading.

cheers!

[/ QUOTE ]


Its because of the huge amount of short term luck involved and the bitch-god Variance.

If you watched any given table for, say, 100 hands you would almost certainly see anywhere from 2-5 winners at the table, even accounting for the rake. But then you don't follow those same players from table to table over time to see how they do. Its possible that some, or even all of those players are long term losers. But over the short term, even with the rake, some number of people at almost every table will be winners.

Multiply that affect out over many observed tables of different players and you will see a lot of apparent winners who simply are not winners over the long term. But because it would be very surprising to see very many tables that nobody beat the rake over a small number of hands, you see a lot more apparent winners than there actually are.

Pokertraker, for the most part, is a serious of these short term snapshots. If you could get the complete hand history for those winning players, many many of them would turn out to be losers. You've just recorded a small sample of their hands win they were winning due to the short term luck involved.

--Zetack

[MicroBob]

right.


i use a different game in my analogy.

picture a casino with 1000 blackjack players all playing generic basic strategy so they are at roughly a 1-2% disadvantage.
They are all going to continue playing no matter how much they are up or down.

After 10 hands: 500 players are 'up' a little bit and 500 players are down a little bit

After 100 hands: +450, -550 (and some of the 45 winners are up a lot)

After 200 hands: +350, -650
After 500 hands: +250, -750

After 1000 hands: +50, -950

After 3000 hands: +10, -990

You could run the same kind of example in gigantic casino with 1 million blackjack players and get about the same results.
After 10 hands it would appear that the house is just dealing out an almost 50/50 game.


So....if everyone only played 100 hands or so and then left you would be led to believe that 45% of all blackjack players end up winners in the long-run.
but this isn't the case...they were just on the positive side of variance.

After enough trails EVERYONE playing a -EV game will end up losers.


Obviously poker is different in that some players are +EV while others are not.
I just use this to demonstrate the aspect of how it can appear that 45% of the players are actually beating a game when, in fact, it's just due to a small-sample size allowing them to hang-out on the positive side of variance for a little while.

[ggbman]

[ QUOTE ]
right.


i use a different game in my analogy.

picture a casino with 1000 blackjack players all playing generic basic strategy so they are at roughly a 1-2% disadvantage.
They are all going to continue playing no matter how much they are up or down.

After 10 hands: 500 players are 'up' a little bit and 500 players are down a little bit

After 100 hands: +450, -550 (and some of the 45 winners are up a lot)

After 200 hands: +350, -650
After 500 hands: +250, -750

After 1000 hands: +50, -950

After 3000 hands: +10, -990

You could run the same kind of example in gigantic casino with 1 million blackjack players and get about the same results.
After 10 hands it would appear that the house is just dealing out an almost 50/50 game.


So....if everyone only played 100 hands or so and then left you would be led to believe that 45% of all blackjack players end up winners in the long-run.
but this isn't the case...they were just on the positive side of variance.

After enough trails EVERYONE playing a -EV game will end up losers.


Obviously poker is different in that some players are +EV while others are not.
I just use this to demonstrate the aspect of how it can appear that 45% of the players are actually beating a game when, in fact, it's just due to a small-sample size allowing them to hang-out on the positive side of variance for a little while.

[/ QUOTE ]

This is an excellent illustration of the issue at hand here Bob, very nicely done.

Gabe

[jtr]

[ QUOTE ]
This is an excellent illustration of the issue at hand here Bob, very nicely done.

[/ QUOTE ]

Agreed. Nice one, Bob.

[jman220]

[ QUOTE ]
[ QUOTE ]
[ QUOTE ]
I think there is another Huge point that is being missed in this ongoing discussion. Namely that the number of hands played by winning players will always be greateer than the number of hands played by losing players. Take the people on this site who have 800,000 hands at 2/4. SInce they are winning they can continue. SOmeone who loses .5bb/100 or slightly less than break even will have busted out their 300 bb bankroll in 60,000 hands meaning that a properly bankrolled slightly losing player will have busted out 13 times over before they play this many hands. Therefore 20% of hands are played by winning players but not 20% of players are winning players. These are 2 very different concepts and it is important to recognize this distinction when trying to understand these numbers.

[/ QUOTE ]

Excellent post... Makes perfect sense...
-
Also, I should have stated in the beginning I was talking only about cash game players. I truly believe (as others have said) that less then 10% of tournament players are winners...

-

[/ QUOTE ]

I believe in the long term, less than 1% of players are winners. Consider a multi-tabling pro who plays the 15/30 game. Suppose this player plays 8760 table-hours per year, or one table-year per year. Over the course of a year the house will rake around $800,000 from this table ($100/hr) and the pro will win $250,000 ($30/hr). The rest of the table must lose $1,050,000.

I think a fairly reasonable assumption is that the average losing player loses no more than $5000/yr. If this is correct, it would take over 200 losing players to sustain one pro.

Of course, at any given moment the percentage of winning players is much higher.

Paul

[/ QUOTE ]

Someone doesn't have to be a "pro" to be a long term winning player. Just because all pro's must necessarily be winning players (or bust), the majority of winning players do not have to be pros.

[Sniper]

There was a recent discussion on this topic in the Internet Gambling forum...

http://forumserver.twoplustwo.com/showflat.php?Cat=&Number=2909887&page=12&view=coll apsed&sb=5&o=14&fpart=&vc=1

Here's a repost of my comments there...

[ QUOTE ]
If you examine all the factors, its fairly easy to see where all these numbers are coming from...

Rough expectation for long term players:
Ring Games - 40% win / 60% Lose
SNG - 20% win / 80% lose
MTT - 5% win / 95% lose

Other factors to account for...


The winning expectation numbers move up a bit when you account for extra $$$ from RB, Bonus, Point programs, Freerolls, etc.


The losing numbers move up alot when you account for the large number of players that try to play and lose their initial deposit and never return.


If you consider the relative number of people that play ring vs tournaments, its not to hard to understand how a casino mgr would be quoted as saying winners are less than 10%, but a ring players PT database shows 40% winners.


[/ QUOTE ]

[Cosimo]

[ QUOTE ]
Over the course of a year the house will rake around $800,000 from this table ($100/hr) and the pro will win $250,000 ($30/hr). The rest of the table must lose $1,050,000.

I think a fairly reasonable assumption is that the average losing player loses no more than $5000/yr. If this is correct, it would take over 200 losing players to sustain one pro.

[/ QUOTE ]

If there aren't any winning players at all, then there still needs to be 160 losing players (per table) each losing $5000 a year to sustain the rake. Each shark that comes along (assuming that they're actually making $250k a year) needs 40 more losers. If the shark isn't four-tabling for 40 hours a week, the number of needed losers goes down.

Are there four people at each 15/30 table taking down $30/hr? I don't think so; there's probably one or two. Call it two. How many more hands a year does a shark put in than a fish? I think it's on the order of a hundred. The shark puts in 80 table-hours a week; the fish maybe ten, for a few weeks at most.

.13BB comes off the table on each hand. There's eight fish there subsidizing this loss, so they each lose at the rate of 1.6BB/100h. This is the fish-loss-rate required to support two sharks and the house. If the fish buy in for 50-75BB, that's nearly 4000 hands of play before they bust out, or around 60 hours. If they play 10 hours a week, that's 6 weeks. Just to support the house (ignoring the sharks), a fish that buys in for $2000 will last for over a hundred hours. That's a lot of poker to a fish.

For every fish that buys in for $2000 (66BB) and busts out in two agonizing sessions, there's another fish that breaks even for three months -- and an endless supply of fish that think they're winning but spend 6 weeks til the cards finally catch up.

[Rubber Soul]

[ QUOTE ]
Tripdad, this old thread (http://archiveserver.twoplustwo.com/showflat.php?Cat=&Board=&Number=1390393&page=&view =&sb=5&o=&fpart=) answers your question I think.

Exactly the same sort of question starts off the thread: "Hey, my PT database says 40% of people win, what's up with the 10% estimates I hear about?" About ten posts down I chime in with some statistical and simulation analysis to show exactly why the PT number can be so misleading. Basically it's the dread problem of small sample sizes once again.

Hope you find the older posts useful.


[/ QUOTE ]

-

Thanks for the link... I have seen three poker tracker stats (from friends) with over 250,000 hands each that show winners at around 20%... That is why I started this thread...

-

[tripdad]

thanks for all the replies, guys. i appreciate the time you spent explaining and providing links, etc...

cheers!