Winning Player VP$IP 50%

[RunDownHouse]

Changed me mind. I might repost this later.

[Mr. Graff]

[ QUOTE ]

(1) Nobody has mentioned the possibility that these 40% players are actually long-term 30% players who have been getting uncommonly good starting hands. When looking at a huge # of subjects, how likely is it to find 2 or 3 people who have been getting significantly more good starting hands than expected?

[/ QUOTE ]
VPIP tends to average itself out pretty quickly in my experience. And given the number of hands these guys have played (8K+) I think it is extremely unlikely they are VPIP 30 players. For each of my 1K sessions my VPIP never swings more than +/- 2 percentage points.

[ISF]

To think you can draw any conclustions from the data you presented here shows a complete lack of understanding of basic statistics. Without a gigantic sample size no multiple comparison test will allow you to draw any conclusions. To test that the biggest winners are all high vpip is trivial. All you have to do is take the biggest 50 high vpip winners from one period. Take the 50 biggest low vpip winners from the same period. And see how those specific individuals do in the next period. i.e. a simple out of sample test. What everyone is saying and you dont seem to understand is that when comparing only the biggest winners over any one period is that you will inevitable get those with the highest variance. The real test is if those same people do will in the next period.
To draw any real conclusions you will need in excess of 20 people from each group I am guessing and at least 10k hands in each of the periods.
The standard error for that one guy means nothing. Say the estimate for his winrate yeilded a probiblity of his winrate being greater then 5bb/100 of 99 percent. Well as there are more then 100 guys like this you would have to expect that at least one would have these stats. Therefore it says nothing about the amount of luck present in his winrate. This is what leads to the need for a multiple comparisons test. An out of sample test similar to the one suggested above solves these problems.

[Lost Wages]

how bout the ones who don't run good go broke before they hit 5k hands you've observed....

Really good point.

Not many -4BB/100 losers are going to stick around 15/30 for long. The are either going to quit, drop down or change.

Lost Wages

[imported_excel]

[ QUOTE ]
how bout the ones who don't run good go broke before they hit 5k hands you've observed....

[/ QUOTE ]

Ok, Im looking at all players, in order who have lost the most, regardless of number of hands,

Top 5 losers, yes they play vpip 65,72,55,35,41, but there PFR's are 7, 5, 3, 7, 5

loose passives, not players im highlighting as winners. In fact seems the easiest way to loose.


Ok, Ill look at the big losers, 200 times the big bet of 30, so players who have lost over $6000,

How many are players who play vpip over 40 and pfr over 18, only 2 players, you think im been biased,

ok losers over $4000, which i have seen some of the good players i talk about losing this and more, but ending day up.

8 players, within above ranges have lost over $4000, 3 of those are extremely loose, +60 vpip, and 2 are bordline, over 56 vpip, so thats 5 out of 8 not in the range im talking, plus the dont increase the aggression with the extra hands, some of the winning players who do play between 50-55 vpip, are also far more aggressive than these losers.

leaving only 3 players who are losers, within the range of play im talking, 1 of those played only 2600 hands,



However under 30 vpip, those players much rarer, only 2 players have lost over $5000, with 9 have won over $3000, not many losers, but not many winners either,

compared to 21 players in the range im talking have won over $3000 with vpip over 40, prf over 18,

[Schneids]

Excel what about that PT screenshot I sent you awhile ago?

[AlwaysWrong]

Well 30%-40% is pretty extreme, obviously. But what if you look at many 1000s of players? Might you just select the ones that have had an uncommonly large number of good starting hands?

[AlwaysWrong]

Good Post.

Excel: Here's my suggestion, hopefully to be modified by people who know statistics better than I.

(1) Select three groups of 10 people each. 1 Group of the top loose-aggressive winners, 1 group of the top tight-aggressive winners, and 1 group randomly selected among all the other players (who say have played more than X # hands.. 500?).

(2) Track those groups over the next month and see how they do.

I'm sure this won't give us anything super-reliable, but it should be somewhat relevant. I picked smallish numbers to make it practical. Practical > theoretical perfection.

[RunDownHouse]

Matt from www.thepokerchronicles.com has a winrate of about 4BB/100 over 125K hands at 35% VP$IP. This comes from 5/10 (3.9BB/100) and 10/20(4.4BB/100), with about 60% of the hands at 5/10 and the rest at 10/20.

[stripsqueez]

my experience in the 15/30 6 max games at empire certainly supports the idea that the swings are bigger than 10/20

my VP$IP has never really changed much from around 24-25% - i did 20,000 hands at 7BB/100 and then another 20,000 or so at about -1BB/100 - the game was wild which is the basic description of anybody who played it

whilst the blind structure should drag your VP$IP up i dont think you could win long term in this game with a VP$IP of much more than 30-35%

a long time ago zeejustin posted about the super aggro guys who then proliferated the 10/20 game - these guys had stats like VP$IP of 30-40% and a massive pre-flop raise % - there was much debate about whether they were winners - i reckon they were although many of them have disappeared since

stripsqueez - chickenhawk