I have 1 million bought hands from 2014 (half from $3/$6, half from $5/$10 PLO 6-max), can send if it's any help.

oh man these last graphs are insane

it makes me sad now

so damn insane

Pasta, there is a bug in HEM, when you have a very big sample on a player/alias and you quickly switch off showdown winnings and switch on all-in EV, it shows the showdown winnings as all-in EV line (orange) and the non showdown winnings as a green line (actual winnings). Sometimes this can be fixed only by restarting HEM. Can you confirm a 100% that this is not the case, cause this actually looks a lot like it.

edit: I am not trying to counterfeit your arguments, I just don't want to believe that this is happening

yeah i know that bug, it's not that, ive watched trough those players one by one too.

here's same in numbers.

new ones.


legit regs.


already banned ones.

how is such a consistency in "running good" possible. Even if you know 4 and sometimes 8 additional cards, one has to have variance and not win every single hand like it seems from the graphs.

you see there's few at ~10bi ds's vs EV with new ones, and bigger downswings on already banned ones, new ones do have that high wr what it makes it impossible to have that bad runs. everything does match with everything with everygraph sofar. thiskind of variance is just not possible.

What is being done getting this to the sites?

Sick graphs you've been posting Pasta.

Great work. It still didn't prove anything yet, but it's worrisome for sure. Don't let the guys saying: "you don't know what you're doing" hold you back.

Thx for taking your time to look into this.

how the **** can someone still say that doesnt prove anything yet? odds for that being variance is less that 1/1billion, it's not even naive, it's braindead talk if you really mean it.

This.

Fortunately, Stars got just nominated as poker operator of the year, I´m confident their security team is already investigating this.



All jokes aside, I´m really baffled by some of the responses in this thread (e.g. Grethe or lautie). The fact that at least 2 of the suspected accounts are long-standing members of 2+2 doesn´t make it less likely that they are behind this collusion/bot ring. I´d argue that it actually makes it more likely, because of the knowledge of PLO you need in order to pull this off profitably and find these ev printing spots while simultaneously taking measures to go undetected. Besides, reading 2+2 is obviously a must for every bot operator because you are able to read up on ongoing investigations and therefore are able to pull your money in time or even obstruct said investigations.

Would be a sick plot twist if some of the doubters in this thread were actually in on the collusion and are just posting doubts here to obstruct Pastafiores efforts.

Not calling anyone out obviously, but would it really be that shocking at this point?

YES

this is ridiculous. excellent work pasta, you're doing a serious PSA here

pasta, 1st of all, thanks for the work you doing, everyone who is legit apreciate. seems you feel offended when anyone say this doesnt prove or don't believe in what you say.

i think you are very very paranoid with this bots/colusion stuff . if you ask me, you think theres bots on stars? maybe, who knows ? IF theres bots, in my humble opinion, its 2-3 accounts maximum . for ex, you listed one friend as bot, i talk w him often about plo, 100% sure not a bot .

this graphs doesn't prove anything . as you know, variance at plo is huge .

can you post this graph from the accounts you think are bots without any filter to us see how much this accounts are winning ? im curious because you asked me to send a screenshot to you (im urubu) with the screenames you suspect are bots. i have million hands databas, all this accounts togheter was like 100k hands , they were marginal winner, almost breakeven.... you saw the graph im talking about with accounts you listed.

another question, is this all about normal tables or zoom tables are included ? if i remember correct, you don't play much zoom... if normal tables are infested with bot and/or humans making colusion, why normal tables winrates are so much bigger than zoom winrates ?

another thing... even if you are right and theres 20 or 30 bots, games at stars still very soft and easy to beat (off course im talking about the stakes this suspect accounts play. not talking about 5/10+). you posted a graph w your 2016 making 13bb, there is guys making 15bb, guys making 8, 10bb...

posts like yours give impression that its a botfest , bots are invencible and nobody wins . its not like that .

offended is wrong word, of course it pisses you off, if you have proved even more proof way, than grethe and schwein last time, that accounts to be colluding. if every human on the earth was playing on stars, there's still less than 0,01% chance for that even 1 of them to have graph like that. and some1 are still talking about variance, wich is just not possible!

why there was only 2-3 bots? that makes just absolutely zero sense, for example last time, how many bots got banned, even if all did not?

that friend of yours do fit even grethe's pattern, and it's same player on that screenshot i posted ~3months ago.

that comment just proves youre incredibly stupid/havent read the thread/youre trying to protect colluders. that graphs prove more than good enough.

if botters/colluders was most winning players, they would get busted faster too. all those accounts do beat rake, and it's not db of my games. and as you see, they have changed something in their gamestyles at ~same time.

read the thread, and ofcourse normal tables are easier to beat. there's colluders in zoom also, just like last time.

why all the worst regs have quited then? and why some longtime legit winning players used to win, does not beat rake anymore?

i have published my WR and stuff, ive said all the time that games are still beatable, but not even close for every1. there's at least 20 more accounts to bust, and it's a fact. youre arguing from bottom of feeling, ive spent tens of hours with this, and i do have some data.

you just do sound like youre trying to protect those, arguing with absolutely zero of concerete things, against 101% proof evidence.

Those graphs for sure prove that there is something shady going on.

Maybe I've missed it but did you take a look at the spots where suspected accounts are in the same hand? It should be the case that the more suspected accounts are sitting at the same table the bigger the difference in EV should be (they know more dead cards).

Additionally, I think you need some kind of software to effectively use this advantage and constantly make proper decisions knowing dead cards.

You don't need software to take advantage of knowing dead cards.

i don't have clear statement about that if they had somekind of software for playing. it's possible, and very likely, but i can't prove it for 100%, yet at least. but they're colluding anyway, wich is against the rules at least as much as using somekind of botting software.

Of course you don't! It will, however, make cheating a lot easier if you need to know dead cards/AIEV playing many tables simultaneously.

I don't have anything to add about thes players/graphs, but I like to thank Pasta for doing this hard work. It surely sucks for them if sombody is named here that is innocent, but at the same time it would be naive to think there is no cheating going on in online poker and I personally believe that AI/bots will ultimately be the end of online poker, so it is extremely important for us players to really keep our eyes open for these king of things.

+1 keep up the good work pasta!

Sent from my Nexus 6P using Tapatalk

Some findings from my old database:


Dates: 9/2014- 05/2015, PLO $100 (normal ring), 3M hands


So almost 720k hands out of 3M was played by the now banned bot ring.

Filter: Position MP,CO,BTN/Did 3-Bet =true/ All in on turn(or earlier)

Filtered to only all-in pots breaks it down to relatively small number of hands (no reason to include non-all in hands) but still sufficient sample for the purpose.



Banned bots graph:


"Reg" graph with same filters:




Chance of this run being variance with these numbers gets rounded down to 0% by Pokerdopes Variance Calculator.

Same DB, different filter:

Position SB, BB/ did 3-Bet = true / All in on turn(or earlier)



bots:


"regs"




The now banned bot team did run massive under EV when 3-betting IP and ran massive under EV when 3-betting OOP to the extent that is not explained by variance.

Is there other explanation to the all-in-ev disparity other than an advanced hole card sharing strategy?

freewilly, thanks for sharing this graphs. why this bots team run so bad at ev in 3bet pots ip ? if pasta theory is right, they should only run above ev, no ?

You got it wrong way.
The bot ring "ran bad" when 3-betting OOP.
When 3-Betting IP the bot ring was running over EV.

Both good and bad runs are so much off the charts variance calculator gives them 0% chance.
(Control group of random regs runs within the variance on these spots)


Pastas hypothesis is they are running under EV by purpose some spots. I have no clue how they would do this.
My database supports pastas findings though.