Latest cliffsnotes on Absolute soulreading.

[apefish]

Solid Josem.

I do have one question to clarify what we are seeing. Did you come across or plot any of the suspected accounts sessions that came after the update but were not included elsewhere in these threads?
I read what you wrote- I just want it to be clear for people taking a look.

It is a pretty ridiculous picture.

[Josem]

[ QUOTE ]
Solid Josem.

I do have one question to clarify what we are seeing. Did you come across or plot any of the suspected accounts sessions that came after the update but were not included elsewhere in these threads?
I read what you wrote- I just want it to be clear for people taking a look.

It is a pretty ridiculous picture.


[/ QUOTE ]

The only "plot" of the cheaters VPIP/winrate is the red dot.

That red dot is my guess at their BB/100 vs winrate. I don't know what their actual BB/100 or VPIP was.

If anyone can clarify, I'll update the graph as soon as possible.


Also, below is a list of other stats that I can easily graph if there's anything else people would like to see:
(the formatting of the text has no meaning)

(i made it an image file since I didn't want to destroy the readability of the thread... if you want the list in text, let me know)

[ChrisV]

Josem,

That's a nice pic, but I'm unclear on why you narrowed it to 500+ hand players. The cheating data we have was over 193 hands, correct? Why not take samples in the 200-250 hand or so range?

[apefish]

The cheating data was over more hands than that I think.

Hang on- I'll go look and then edit in the number I find.

And Josem- as the person who sifted thru the stuff to post the notes a couple times I dont think that approximation of either vpip or winrate is that far off. Even if you make it like 85 and something less on the rate it's still basically in the same spot on that plot.

edit: see next post I made for info on hands played.

[Josem]

[ QUOTE ]
Josem,

That's a nice pic, but I'm unclear on why you narrowed it to 500+ hand players. The cheating data we have was over 193 hands, correct? Why not take samples in the 200-250 hand or so range?

[/ QUOTE ]

i thought there were at least 500 hands on the cheater.

[apefish]

I've gone back and looked. It looks pretty apparent there are at least more than 300, although how many more than that I don't know because I have zero idea on the dates of the differing data. There could be more than 400-500 unique hands in that data- but again I cannot tell with certainty.

Even including the data for players with 400-500 hands I cannot imagine the picture as we see it changing. What you may find is more players in the greater outliers- but that's about it.
I'll look later to see if I can tell which shots are unique to come up with a number for hands played- but I'm pretty sure the talked about number was north of 400.

[Josem]

the PT screenshots in the second post in this thread give these figures for NL:


hands vpip BB/100
103 86.41 311.36
138 96.38 470.37
193 93.78 468.96
8 87.5 821.56
28 92.86 1098.48
95 89.47 462.92



assuming that there is no duplication, that works out to a VPIP of around 92.21 (each vpip contributes to the average according to the number of hands that it represents... ). that is for 565 hands.


similarly, the Big Blinds/100 is 475.74. Obviously, my wild guess of 275big blinds/100 was way off.

I'll update the graph immediately to reflect these corrections (note that this will change the graph in the preceding post, so some of our discussion won't be so clear - for future reference, the red dot was previously at 275bb/100)

also, i'm introducing an assumption here: that those PT stats were displaying Big Blinds/100. If they're showing Big Bets, then the win rate should be twice as big - so if anything, I'm erring on the conservative side by 50%.


edit: I've uploaded the Excel worksheet with these calcs to http://www.absolutepokercheats.com/w...dvpipcalcs.xls
It doesn't have any macros in it, and I take all care but no responsibility for viruses etc.

[apefish]

Seems about right about the hands numbers. But I'm still not sure those are all unique hands.
It seems pretty clear we are talking at least 300-400 unique hands with some reasonable assumptions about those shots. And if the shots were provided as session shots from the same source or from people playing the sessions- then the number is close to the sum of them all.


edit: and I wasn't saying your vpip or win rate numbers were off. I was pre-empting anyone who wanted to argue they could be skewed too high by showing that even reducing them to numbers nobody could argue against does nothing to reduce the ridiculousness of where they fall on that plot.

[Josem]

even if some of the hands are duplicated, while the data would not be perfect, I don't think there's likely to be any inherent bias created.

losing hands are just as likely to be duplicated as winning hands.

of course, if someone had compiled all the HHs into one PT database, that'd automatically filter out all duplicates. Obviously, if that was accessible, I'd use that, but in the meantime, I'm confident that these figures are a reasonable proxy for the real thing.

There's always going to be a margin of error, and it doesn't fundamentally matter whether the cheaters had VPIP of 95, 92, or 89; or win rates of 450, 475 or 500 bb/100. Even the most conservative figures show the cheaters winning fives times faster than ANY OTHER player in a sample of millions of hands, with unmatched levels of VPIP.

[apefish]

[ QUOTE ]
even if some of the hands are duplicated, while the data would not be perfect, I don't think there's likely to be any inherent bias created.

losing hands are just as likely to be duplicated as winning hands.

of course, if someone had compiled all the HHs into one PT database, that'd automatically filter out all duplicates. Obviously, if that was accessible, I'd use that, but in the meantime, I'm confident that these figures are a reasonable proxy for the real thing.

There's always going to be a margin of error, and it doesn't fundamentally matter whether the cheaters had VPIP of 95, 92, or 89; or win rates of 450, 475 or 500 bb/100. Even the most conservative figures show the cheaters winning fives times faster than ANY OTHER player in a sample of millions of hands, with unmatched levels of VPIP.

[/ QUOTE ]

Agree completely. That's all I was trying to help say.
Again- nicely done.