Before you go off and pour over hundreds of hours of Mike Postle streams, why don't you contact a math/stats person. They can guide you on what factors you need to consider, what you are overlooking and if this is even worth doing. You could email professors at MIT, you could email the Bernie Madoff whistle blower, you could pm David Sklansky and Mason Malmuth on 2p2, you could tweet/email Bill Chen, author of "Mathematics of Poker", and Liv Boeree. I'd spend at least an hour or two contacting people and getting recommendations on how to proceed before spending so much time on what might be a fruitless endeavor.

would just like to say whosnext modding has been on point throughout this thread.

agreed, it's gotta super tilting getting reprimanded by someone with invisible dick hentai blowjob porn as an avatar too

The core argument to proving circumstantial cheating is not to look at $ or bb/100 results, but instead evaluate his decisionmaking in light of the fact that he seems to know his opponent's precise holdings.

This may be similar to the idea of a zero-knowledge proof in that it seems that Mike chooses to do things incredibly accurately in a way that could mathematically prove he has access the information demonstrated by the proof.

I don't think you understand what "god mode" is referring to.

Yea, that's true didn't think of that.

Well, that would be a good fix but for all we know they already have it in the docs. I kinda lean towards blaming Stones Live for these types of things since, to my knowledge, it doesn't really happen all that often on other PokerGFX powered streams and the operating side of Stones seemed to be a complete **** show.

I'm not quite sure what you're asking about this session. You're talking about the commentator Justin Kelly, and the timestamps show that he curiously left the commentary booth at some point and went to talk to Postle. Footage of that can be found in the shots of the table (30 minutes earlier in the youtube timestamps). The commentators also have the ability to alter some of the broadcast as it's being streamed, such as going full screen, or muting their own chat in order to highlight the table chat. They aren't seeing holecards at the same time as the players. They see them at the same time as the viewers (half an hour after the hands were played).
There is someone in the "peek room" (I think it was Lance in this case), that is just the other side of the partition in the booth. The guy in the peek room sees holecards live, but also can simultaneously monitor the delayed stream half an hour later, when the commentary team add their voiceovers. This means the guy in the peek room can also warn (by shouting out to) the commentators when there is an RFID failure on the hand they are about to see. (He's already seen it play out in realtime).
In terms of security, it's a bad system, because the 'peek room' guy should ideally be locked away with no method of instant communication (except in cases of emergency) to anyone in the building while a session is in progress. It's a viable system when a stream has technical issues, particularly for smaller establishments with fewer staff, when the guy doing the action tracker/holecards also has some responsibility for the smooth running of the stream with commentary. e.g. Sometimes the peek room guy also has to add overlay graphics to the delayed stream, or notify the commentators of other issues.

Yes, yes, I 100% agree. It appears to be both the easiest way and the right way to prove, in a court of law, that he was cheating. Zero knowledge proof was a very interesting read, thanks for sharing.

Even if it turns out that using all that data is not going to be fruitful for a statistical analysis, it would be useful if the hand histories were gathered and put in a format for importing into PT4 or HM2 for analysis by filtering on various situations, especially if they were separated by when his phone is in his crotch or not.

Think of all the queries you could do by filtering hands using one of those trackers. For example, finding out how perfect his river bluffs are:

Filter on all the hands where he has a weak hand on the river and has the opportunity to bet or raise. Compare the percentage of times he made river bluffs when they were expected to work (opponent has air or a weak hand expected to fold, which includes when they happen to hero-call) vs. when a bluff would be expected to fail (they have a legitimate bluff catcher or better).

If he bluffs near 100% when it's expected to work and near 0% when it's expected to be called when he's in god mode, but only some middling percentage for each when not in god mode, that would be more damning than any statistical evidence in my opinion.

And there are many other such situations that can be similarly used to make pretty clear even to a non-poker playing judge or jury that he knew everyone's cards. Basically, this allows for the evaluation of his decision making as zizek suggested.

Great, but what you should be doing is comparing the winrates of non-cheating sessions and cheating sessions. If you can show that Postle (and indeed anyone else at the table) wins at less than 20bb/100 in a dozen non-cheating sessions, with a "normal" standard deviation, and then show he constantly wins at 600bb+/100 in a dozen or so cheating sessions with a different standard deviation (he doesn't lose any big pots), you go some way to proving that his crotch-staring winrate is a level of magnitude higher than is normally feasible.
Comparing two cheating sessions doesn't prove much about possible winrates. You also have to show that no one (including Postle) ever won that much without cheating.

Exactly, you make a great point. The only problem is the amount of work it will take me to do this alone, if somebody could set up a discord group with 10-20 regs we could have this data collected rather quickly.

As someone else pointed out, linking to hands where someone made a hero-call doesn't prove that Postle wasn't cheating in that hand. You might be right that that was NOT a godmode session, and I'll certainly add your timestamps and notes to my database, but way back in the thread someone suggested that there were some suspicious hands in that session, so a further look is needed. As I've said before, it's important to work out which sessions involved 100% legit play, because it can help us work out the reasons why he didn't cheat sometimes. e.g. We can compare notes on the whereabouts of suspected accomplices.
Thanks again for doing your analysis. Keep it up!

https://youtu.be/DGaM1-HeXMI?t=13880
24/09/2018

Here we go again, watch this hand guys, i was checking out Mike's hand before that and the cardreader didn't read it at all.
This time here is different anyway, because the booth is not saying through out the hand, that he has another holding. At first i was being confused with this hand, because the pot in the end is being moved towards Mike on the table.
But looking closely again, you can see on the table which hand he had actually here.

This is interesting as well, didn't get to see this before:





So it seems they are keeping track of this somehow.

I'm not sure I do either. I'd define it as having knowledge of your opponents hole cards and adjusting your play to exploit that knowledge.

How would you define it?

I was trying to explain what i'm trying to look into ealier. I agree.
I'm not in mood to go through every single hand in detail, it's just exhausting, to watch and to write it after then. I will try to put some better notes on now.
And anyway, the spreadsheet should have god mode on or off already on every session. This is what i would do first of all.

I think you understand it perfectly well, and you were right in your earlier post when you said words to the effect of it being intermittent. (Some times, it's intermittent because the RFID keeps breaking). In some sessions, he'll play normally for over an hour at the start (presumably because his accomplice isn't in the peek room), and then switch God Mode on for a few hands, or maybe the whole of the rest of the session.

Thanks, one other thing I would add is that I think he intentionally pays off bets here and there particularly after doing a sequence of non-standard things that net him money. This is speculation but I feel very strongly it's true.

I guess the point I'm trying to get at is his win rate is that much more "impressive" given that he's not playing, or trying to play, perfect every hand.

as mentioned earlier -- if you have every hand history recorded down from every live session that postle played you can start looking at which stats are completely out of whack for a normal player to achieve and you will also have stats to every other player in game as a comparison point.

Its a lot of work to record every hand history from every postle session on stones live stream (As he played for over a year) ... but it would be very useful. you can take this data and produce many meaningful observations based upon the data.

So his method of trying not to get caught would seem to be that he only cheats for a small amount of hands at a time rather than the entire session? That makes sense. Thank you. Or maybe if he gets into a hand that starts to get a little dicey he goes into his crotch for answers.

Using bb/100 to prove he's cheating doesn't make much sense to me, I mean, he wasn't caught purely on his bb/100. He was caught overwhelmingly by observation of his play and bb/100 was only factored in afterwards on inflated numbers. Why would we try to prove, in court, that he was guilty by bb/100 when the argument is both weak and wasn't even what made most people think he was cheating in the first place?

I believe bb/100 and HH stats wouldn't do much in court, and even if it did, VerStandig couldn't use your findings, they'd have to make their own. Furthermore, most poker players, even good ones, wouldn't even understand statistical frequencies if they've never played online with a hud, let alone a jury of non-poker players. Regarding HH and the stats they could bring, wouldn't you have to transliterate each hand into the specific hand history format? That would literally take forever. Probably 10x as long as just manually getting bet sizes and winnings. If you want the stats because you want the stats thats cool, but is it worth it?

I think there's this misconception that a jury couldn't understand what is being explained to them regarding Mike Postle's hands because they may be non-poker players. But it is pretty simple what Mike is doing. It isn't rocket science. He isn't doing some high level GTO solver bullshit where sometimes this works and other times that works depending on the range and frequencies you've constructed for your opponent. This is straight up ABC superusing, if there is such a thing, and I think that's a lot easier to prove than using win rates and frequencies that a jury will have no concept of. I think it's cool to gather the stats if that's what you want to do, but just want to be realistic about the outcome. ArtyMcFly's idea seems to be the most logical way to go about it that's been posted so far, it's probably a good idea to ask other experts in the field as well. Good luck on your endeavor my friends.

Just looking over that 88 vs TT hand again. Looking over the potential error options.

The blinds are 5/5 with a $10 straddle on utg.
When you look back at the preceeding hand, from the size of the pot going to the flop $35x4 = $140, means that everyone else folded without putting a bet in.
So utg+1 folds, utg+2 folds, then it's Postles turn to be first to act.
So for there to be a persistency error, he would have to have folded 88 preflop from the Hijack, facing no action from anyone. Not even a limp.

He claims it was 78 in the next hand (in the 88 vs TT hand). So he folds with 88 preflop, but opens with a raise with 78? That would be his claim.

For there to be any persistency error at all, it must be that computer would have to register whatever hand it thought it was the first time around, during that hand.
There must be logs of all these hands right?

https://www.youtube.com/watch?v=DGaM1-HeXMI
24/09/2018

Interesting fold here

https://youtu.be/DGaM1-HeXMI?t=7719

What i also recognized throught some other streams i went along the last two day,
the chipcounts are not always perfect too, also people randomly rebuy during the sessions
and you can not see that somewhere, unless you watch out carefully when
people actually rebuy on camera. I found an example here, Mike's stack is somehow displayed
with 2.8k or so, the comms are saying that his down around 100 bucks for the day at this point.

Somewhere earlier in this stream i remember darkly they were talking about the
buyin rules, in the beginning with a minimum of 100 bucks and a maximum of a maximum 500 bucks.
And the during the session you are allowed to rebuy to the max stack at the table, something like this.

https://youtu.be/DGaM1-HeXMI?t=10954

This here is a good example on how loose Mike is playing, and there is those situations, where its paying out big times.
In this session he is really playing a lot of hands, trying to get on the flop almost all the time.
I was other sessions where he doesn't do that, or at least he only does sometimes, and it's also those streams, where
he is not or only sometimes in god mode. Could be a correlation here between max VPIP and god mode, because
knowing the cards pays off more on the flop or postflop.

https://youtu.be/DGaM1-HeXMI?t=11806

Just found another one of Mike's steals.

https://youtu.be/DGaM1-HeXMI?t=12126

This here is a good example what i also got to see again and again. Mike is standing and i was trying to get a look at him
when standing up and sitting back down again. Click the link to see what i happening then. It's pretty often coming a scene change then.

https://youtu.be/DGaM1-HeXMI?t=12957

Btw. JFK was in chat having day off (at least it was writen). Mike had an agression factor of 50 in this session.

Yes, exactly. I've only watched the first 4 streams though, I assume there are less and less of these hands as he gains confidence. He's really uncomfortable, awkward, and shifty at points during the first 4 streams.

The cheating is definitely evolving. Stream 2, he moves his chips. Stream 3, he has his left hand folded over his lap and moves it slightly while looking down. Stream 4, he's no longer sitting with his left hand in his lap but wearing sunglasses and back to just pulling his chair back and looking down.

It looks like he's trying to perfect the stare down but having a lot of trouble with it along with a lot of anxiety. He nervously looks over his shoulder something like 15+ times during these streams anytime someone walks into his peripheral vision.

Yes, the HHs would have to be put into one of the formats that could be imported into PT4/HM2, and that would be very time consuming. SrslySirius said he used to have to do this exact thing for some work he was doing, and he wrote a script that helps put it into the proper format. I believe he offered it for use for this purpose. He said it was still time consuming, but it helped.

Maybe it's not worth it, but if it showed that stark of a contrast between how perfectly he chooses when to bluff the river in god mode vs. not (as well as other situations), that would be the most compelling evidence in my opinion, especially because it would include every hand that matches the filter, not just a few cherry-picked ones.

I suppose this could also be done just by looking through all of the hands by their descriptions without having to turn them into HHs first, but once the effort was done to put them into HHs, the number of situations that could be quickly filtered on would likely uncover a lot of situations where he clearly knows the cards in a very convincing way.

Yea I don't disagree that it would uncover situations we may have overlooked, as well as, be useful in many other ways, but it's just such a gargantuan task to get the formatted HH that I don't think the payoff justifies the workload.