I think Oborra had some preflop problems, and Strongpanyid is automated I think?

Most of the money in poker comes from mega adjusting vs fish and weak regs, so pure RTA not operated by a poker player probably won't win as much as the best humans still I think (well evidently if strongpanyid is winning 3bb/100)

I always tell my rec friends not to play online cash games. Spins or MTTs. And even then there’s no guarantee you are not getting cheated. Regs are such lowlife scum who will do anything for an edge (not all regs...obviously)

The reality is with online cash unless you are willing to work hard on your HUD /stats game you are at too big a disadvantage now. Better off playing the bloody slots

Obviously this is jus the tip of the iceberg of online cheating . Just for high level reg on reg wars blahblah **** off cheats . You’ve ruined online poker with your greed . Cheers for that

This is such a dumb argument. If you are playing solver RTA, you can play for hours, absolutely knowing you are +EV after rake. At high stakes rake is small, so you know you're +EV. That is huge! You know you are profitable! Your strategy against the best regs not using solver is better than their strat, and you make money against fish. It's a money printing machine.

has this kid Kruze state his case yet anywhere on any other sm platform?

No, not yet. He joked about it in a Twitch stream though.

If he'll publish a statement, it will most likely be some boilerplate rubbish. It might get a little spicey though if he comes up with accusations that the whistleblowers or some of the other streamers were kneedeep in it as well. Other than that, I guess there's nothing to see here anymore.

He said he would make a statement after he talks to pokerstars on tuesday, looks like no statement will be coming I guess since its been 3+ days? Maybe whatever excuse he was going to give stars to get his account back didn't work?

You said you would shut up!
Everyone would be happy do not hear your bullsh...

What has this guy been doing that's on the shady side?

I said I wouldn’t continue teasing the short feminine men here.

You still salty shorty? Go suck on that exclamation mark you used. If you want my attention so bad, PN me. Maybe I’ll let you taste the d and make you my little sissy.

That's a very homoerotic thing to say for such a tall, masculine man. I'd even go as far as suggest it might be flirting.

True. That is pure RTA. AI though which learns and implements strategy based on opponents is possible. I'd imagine the best cheaters using this technology incorportate hand histories from players they have played with into a database to improve decision making. I doubt many out there are that sophisticated as of yet but once RTA is accessible to lots of people then building on top of that will be what comes next and I have no doubt people are using or working on such a thing right now.

It is why I would say some of the sites have tried to move away from providing hand histories as if you can buy massive databases on your opponents and incorportate these with RTA well I don't think the end result needs pointing out.

A solver uses input parameters of ranges and strategic options (bet and raise sizings) and outputs a nash equilibrium game tree which is an unexploitable strategy, meaning that your opponent can not adjust his strategy and improve his ranges overall EV. A player using RTA with detailed, well-built trees (actually easier said than done) can not really lose to a human player in heads up situations. If the opponent is also playing a nash equilibrium strategy (cheating), the players will break even pre-rake.

Some people have the misconception that this means that it does not matter what the opponents strategy is and that all strategies will have the same EV vs a solver strategy, but in reality, human players lose tons of EV vs the solver all over the place by not value betting thin enough or for the wrong size and bluffing/calling incorrect blocker combinations.

There's many non-intuitive plays that happen with some frequency in solved game trees. A common/generalized one would be the high amount of turn donking that occurs. A more rare/specific one might be overbetting K-high on the turn and then check calling unimproved on the river. Basically plays that are difficult if not impossible to fit into a model of human poker logic. It's similar to watching a high level chess computer play; the solver/AI is looking at the game tree and preparing for scenarios with such breadth and depth that is unapproachable by the human mind.

There are differences between chess and poker at competitive level. While to play online tournament for significant sum of money you have to be vetted or at least well known and your ELO to be established, while in poker there are no conditions to fulfill, except buy-in which is a blessing and a curse in one, since you want for fish to remain in the pool but also you don't want anyone unknown who can't prove that he achieved those stakes naturally with out RTA.
Client detection is impossible to enforce (cameras, mouse detection, program detection ... ) everything is unreliable and spoofable since you are giving the control of the device to the client.
Also ghosting is real problem.

The only 100% way to verify the player playing online, is to have him show up on location and play on provided hardware (like a online casino on location like bookies in UK or some establishment like that).

I just hope his Twitch and YouTube channel also get banned, would hate to see he continue to make money on subs views etc.

The first example you gave is not in any way "unapproachable by the human mind". (I don't really understand what you mean by the second one).

Developing a donking range on a turn card that favours the check/caller's range is a very easy to understand spot both theoretically and practically.

Lots of regs have implemented this, and those who haven't, including myself, have most likely put it on the back burner because there is more low hanging fruit out there unless you are already one of the very best players.

More generally, lots of things that are now considered standard, such as betting range for 1/3 on certain flops, or overbetting certain turns, were once considered "counter intuitive PIO moves".

The game evolves, and the only concern we need to have is catching or disincentivizing cheats. Making good theoretical poker sound like some impossible to comprehend abstraction doesn't help anyone.

This was about fedor holtz i think? I thought that too. Why is he regards as shady ?

This is how to catch an perma ban.

Seems like everybody itt is getting shadowbanned.

love you my darling!

I think that the original poster who's question I was attempting to answer was interested in identifying unintuitive plays for the purposes of cheat detection. I mention the turn donk not because the play itself is impossible to find, but it is impossible to play truly accurately without an insane amount of work. You are correct in that implementing this strategy is not particularly high EV to study, and thus one would expect high inaccuracy in this node. So if you are a security team looking at a suspicious player's hand histories, and this guy is just nailing all these OOP caller 3b pot turn donks, it would rank to be quite likely the player is using RTA.

Most of these cheating guys using the tree database method probably just run massive scripts to produce their trees with a ton of sizing options on each street (so they can defend accurately against every bet size). It takes a lot of effort to go through multiple sets of trees, comparing EV's and pruning lines from certain flop subsets etc. The result is a strategy no human can approximate, and should be detectable if we are looking in the right places.

I'm not suggesting that you can say "oh he donked the turn, cheating for sure", but its a spot that guys are not going to be getting their frequencies right nearly as often as flop cbets or preflop or something.

I find your turn donk example flawed in multiple ways. First of all it will be extremely hard to get enough hands to have a sample size where you could say "yeah his freq here are so on point, either super human brain memory, or he is using rta"

I would imagine to have good proof you would need a lot of hands. That's very rare. Who in hs has 1million plus hands on each other? Only the super crushers like Linus bbb etc. Caus they play for a long time with each other.

Secondly it's just not hard evidence. As seen with some of the extremely best players, people will be able to just learn this stuff. It's not that you have to learn all the boards, combinations or what not. Obvethere are patterns in pios play. Mostly driven by some kind of range advantage, blockers, etc. Tbh I have never seen any kind of very very inhuman u logical play by pio. All makes sense, u just have to think more. Meanwhile pluribus did do some extrem weird ****, that pio didn't agree with.

And people are learning this. So u cannot just say from. Some random spot that this person uses assistance.

It's good that poker site has this data of all user hands. Ofc there needs to be some significant sample to get results. But people just knowing there is a possibility of getting banned and funds seized when using RTA is enough to deter them from using them. Mainly if the decision and detection system is somewhat obscured and evidence is presented to some kind of authority so player funds cant just get seized.

Africa Brass-you bring up an excellent point about myriad sizings as this, I recall reading in the textbin, was one tell that brought Kruse to the attention of some of the pool... Balancing multiple sizings well, even OTF. I think your right, originally, he's run a plethora of sizings on each sim for each street and then has to implement these sizings to reach an equilibria!

True enough, definitely just this is not enough to convict. I do think though, that given enough spots with clairvoyant hand histories, we can do mapping between a player's strategy and PIO solutions and determine the statistical likelihood of a player's accuracy being that high. The turn donk example is just to set a flag for the security team to look more into the player's hands, as surely they can't be asked to do full analysis of every player. There needs to be a simple filtering system so that a security team's resources can be most effectively utilized. Beyond that, player's probably need to play on camera and have the monitored gameplay compared with the suspect period of gameplay. I don't believe the issue at hand is if we can detect cheaters (100% with enough effort we can), but rather is how do we create an effective method of narrowing down candidates for investigation, especially as computing power cheapens and allows cheating software to be viable at lower and lower stakes.

And just to touch on your comment about "inhuman" play, yeah I agree that PIO's lines have some logic to them in all situations, but we are always reverse engineering this logic. We observe what the computer does and try to figure out some kind of simplified rule that covers most situations. Just try node-locking a turn spot with your best approximation of GTO strategy and see how much EV you lose, it is crazy difficult to play anywhere close to the level of a solved tree.

Yeah, an effective set of trees needs to have as many sizes as is computationally feasible. The full set in 6max already costs potentially 5 figures in server time to compute, so there are practical tradeoffs. I'm certain there are some sharp guys out there who can engineer a way to mimic more simplified strategies. Kruse seems to be more of a heavy-handed cheater with off-the-shelf type software. Not sure if anyone's looked into his multiway play, but usually that's a super good indicator if there's a large differential in skill compared to his heads-up pots, as building a full set of multiway spots for every texture would be an astronomically large computational task (if it's even possible? I haven't really looked into it).