F Paulsson
euro love
Silver Level
I agree that there's a sampling bias in Zach's database, but I have to stress that even when we do manage to datamine objectively, the only thing our sample tells us is how many people are winning players for that period of time. But maybe that's what we're discussing. How many of the people currently playing are currently winning.
The real problem, and I took the liberty to check this logic with my math genius wife, is that the player base we're looking at isn't static. There's a shift in population, and I contend that the shift is overwhelmingly that random players enter, but losing players leave. In statistical terms, we're not sampling the same population every time. Therefore, our samples are not reliable.
A way to go about this, which would still contain some margin of error, is to datamine every table across all stakes for a longer period of time. That way, people don't get to move up or down (as Zach mentioned) and escape our proverbial net for when they finally stop playing.
The real problem, and I took the liberty to check this logic with my math genius wife, is that the player base we're looking at isn't static. There's a shift in population, and I contend that the shift is overwhelmingly that random players enter, but losing players leave. In statistical terms, we're not sampling the same population every time. Therefore, our samples are not reliable.
A way to go about this, which would still contain some margin of error, is to datamine every table across all stakes for a longer period of time. That way, people don't get to move up or down (as Zach mentioned) and escape our proverbial net for when they finally stop playing.