Players Quit Quickly Due to Heavy Losses?
Sun, 02/07/2010 - 21:47 — maggie
X axis is bb100. y axis is number of observations. For players with 1-100 hands. Most of the players are losing more than 100BB every hundred hands they play.
I have been looking at the problem of collating statistics from the PokerFTP database.
At time of writing I am having memory issues, which are either from storing data for thousands of players in RAM, or from using arrays to store player profits for each hand.
I believed that one method for solving this issue would be to exclude statistics for players with smaller numbers of hands. For example, players with under 100 hands of data wouldn't have particularly significant statistics anyway, and excluding them would therefore be useful.
For players with 100-500 hands
However, looking across the full range of data in the database...
- Players with under 100 hands of data make up 43% of players.
- the BB100 chart for players with under 100 hands of data shows that vast majority of them have lost over 100BB per hand.
Therefore most players who have under 100 hands have probably quit due to losing their whole bankrolls.
The proportion of players with exceptionally low BB100's (under 100) falls as the number of hands increases.
For players with more than 1000 hands.
However, a potential problem with this result could be that it is skewed by the higher number of micro limit players. For future reports, where results are grouped by criteria such as blind size, I shall need to do survivorship bias tests for each result group.
We can conclude that a player's hand count greatly affects their overall result. And since the 1-100 hand player group clearly contains the greatest number of losers and players, excluding them will give any analysis a massive bias away from the least successful players.
X axis is bb100. y axis is number of observations. For players with 1-100 hands. Most of the players are losing more than 100BB every hundred hands they play.
At time of writing I am having memory issues, which are either from storing data for thousands of players in RAM, or from using arrays to store player profits for each hand.
I believed that one method for solving this issue would be to exclude statistics for players with smaller numbers of hands. For example, players with under 100 hands of data wouldn't have particularly significant statistics anyway, and excluding them would therefore be useful.
For players with 100-500 hands
- Players with under 100 hands of data make up 43% of players.
- the BB100 chart for players with under 100 hands of data shows that vast majority of them have lost over 100BB per hand.
Therefore most players who have under 100 hands have probably quit due to losing their whole bankrolls.
The proportion of players with exceptionally low BB100's (under 100) falls as the number of hands increases.
For players with more than 1000 hands.
We can conclude that a player's hand count greatly affects their overall result. And since the 1-100 hand player group clearly contains the greatest number of losers and players, excluding them will give any analysis a massive bias away from the least successful players.
[ Financialcharting.com
|
Borhon
]
About maggie
Personal Bio
Maggie is a statistical analyst from the city of London.
She has been playing Poker for over 5 years.
Syndicate
