[MasterLJ]

[ QUOTE ]
[ QUOTE ]
the stats are not identical. their VPIP ranges from 13.64 to 14.15. this may not seem like much to many, but for an fr nit, this is huge, and there is no way the range would not be tighter over 100,000 hands if the accounts were being run by the same bot program.

ftp investigated these accounts for a month. does no one here even concede the possibility that ftp may have more information exonerating the accused than what so far has been made available here?

[/ QUOTE ]

As to your first point, there are 2 explanations. First, even over 100k hands, hand quality distribution would not be even. I haven't run the numbers, but from my experience this is a very tight grouping, not a loose one. Can someone do a monte carlo sim or a regression analysis to show this? And second, in was clearly stated in the very first post of this thread that the bot owner will take manual control from time to time, and this also is enough to explain the discrepancies.

As to your second point, yeah I'm open to the possibility we are wrong. But FTP needs to show up and do some explaining, because they are looking bad right now.

[/ QUOTE ]

I developed one a couple years ago. You need 1000 samples of each hand "type" to start seeing some decent convergence (but not strong). That's 169k hands if you look at it as AKs vs AKos, 87s and 87os, etc etc etc.

After 100k hands it is absolutely possible the VPIP could be slightly different.

I've always been curious as to how you can get your computer to auto-input information (ala clicking the mouse etc). I know you can wrap DirectInput (A MASSIVE undertaking), but that should be very easy to detect.