I dont know of any real time thing.

who is sallywoo

Lets just have em play live for a nice big sample, along with the rest of the online superstars. And see how they do.

Oh wait.... live is too slow... and umm boring. ......

Translation: no ****ing way dude. I wont be able to have my computer program help me with the volume I need to overcome the variance cause its all math and computers are math and the computer helps me overcome the boredom cause it allows me to play a billion hands more than any human could possibly play and basically lets me profit waaay quicker than im supposed to. Then i can claim im the best and be on top of the darkside nerd community when in fact i know its all bull**** and i would probably be crushed live because my A.D.D. consumes my life and god damm it im entitled to this!!

Gl with OLP in USA... it aint goin nowhere...save this post

Buncha darkside nerd cheats imo... pff :middlefinger:

Tons of people itt have no idea how computers work. They can't do magic. I studied computer science at the undergrad level and have several non-poker friends who are currently studying AI and machine learning at grad school. Similarly I've played poker on a professional or semi-professional level for 6 years now, although I play nlhe and not o8.

The bottom line is though that in general all computers can do is to do things humans can do, but slightly faster. The game space of most poker games is ****ing massive. I know someone mentioned about how many states of chess there are, but in any game with a flop/turn/river there's the introduction of 52*51*50*49*48 = 311875200 states. Every time there are 3 possible actions increases the amount of states at an exponential level. The point is computers aren't fast enough to deal with this kind of information, even with the simplifications possible.

I also studied the Alberta HULHE bots pretty extensively (read through their entire 100+ page report and exchanged some emails with the author about things I wasn't sure about). Even their gto bot didn't attempt to actually traverse a game tree, it simply made simplifications, tried some different bucketing strategies, and then played against itself on what were essentially supercomputers for years and years before it was able to beat most human opponents.

So which is more likely? That a few guys who were already talented enough to beat high-stakes limit decided to pay people to write software that would take years and years to get good enough at this game, and probably magnitudes longer given I believe plo8 is a ****ton more complex than lhe (haven't really played either so I'm not sure, but I'd be inclined to think that since you get 4 cards and can only use exactly 2, and there are 2 pots, that these are all things that would increase the game tree on an exponential level). Or is it more likely that these guys worked really hard on their own game, maybe used some software commercially available to improve at plo8, and then beat someone widely considered a fish at those stakes?

Like I seriously don't think any of you realize that computers aren't as powerful as you think. I've written programs to help me here and there. I wrote something to color-code opponents based on info from my db, I wrote something to give me a crude rush hud right after rush was invented and before one was commercially available, and I've written some sql code to analyze my db a bit better than commercially available software currently can. Most guys I know beating 5/10+ don't even know how to write a hello world program.

I understand most of the nvgers want an excuse as to why other people are crushing nosebleeds and they can't beat 2nl, but it has absolutely nothing to do with complex computer programs that are solving the game for them. In fact the most cutting edge poker software out there today is the stuff that focuses on bumhunting, since that's way more profitable than actually getting good at poker. But the sickos at the top of the food chain who are crushing didn't get there because they have a program, they get there because they've studied the game a ton, and they also work with other people who are on a similar level to them, and this helps everyone get even better. It's pretty likely that these 3 guys talk strat all the time, are already above pretty much everyone else at this game, and are only improving and widening the gap between them and their competition.

@zach

You're right, the game space is massive for Poker, especially games like NLHE and PLO which are quite complex in comparison to LHE. In LHE it's possible to compute a pretty solid equilibrium if you use the right abstractions. For the more complex / more popular games, people now are just using neural networks and doing billions of simulations of the bot playing against itself and "learning" what good plays/bad players are. Then once the bot has a decent strategy another bot is passed its database and made to exploit it, so the original bot learns over time how to be less exploitable. This has been done for NLHE many times (lots of private programs exist in this area which we probably aren't aware of), and I think it would be possible for O8 as well. It's not unlikely that somebody somewhere has a GTO O8 bot programmed that's close enough to actual equilibrium to beat high stakes players.

That being said the most probable scenario is Sally just used some software to improve his game, and isn't actually cheating.

Solid post. Unfortunately the substance of it will be ignored by many, but I guess that's the fun of NVG. Conspiracies, Evil Geniuses, etc.

Zachvac is pretty much spot on.

I recommend watching a documentary on youtube about American chess player Bobby Fischer for the people that don't believe it's possible to become amazing on your own. Having the ability to track your leaks today with software and dedication that has always been around, people can become great if they are serious and motivated.

Well, there is no luck in chess so if they design something marginally better it's going to win 99/100. Hence the "crushing."

That said I don't necessarily agree that it is wayyy wayyy more complex. Poker can be pretty complex too, but from what I understand the LHE HU bots are basically unassailable.

@Zach

Well sorry, you talk about PLO8, not LO8, which says a lot about your understanding of the game itself. You are obv out of your profession when talking other games than NLHE. Your post was pretty interesting, but since you don't even know which game we are talking about here, you of course didnt add any relevant content to this discussion.
Playing 2-card poker and knowing a lot about computer is nice, but won't help the case when we are talking about Limit Omaha8.

To all others who havent play LO8 heads-up themselves, I can tell you that it will be easier to do a bot for it than for LHE, since naked bluffs nearly never happen don't happen on low boards, cards can be described as high, low, and Aces (H,L,A), and ranges will quickly be very definable. Now having a programm that helps to see our equity against Gus' range just 1-2% more precisely right on the spot, this would earn a ****load of money.

Once again Zach, the difference from PLO8 to LO8 is about the same as LHE to PLHE. No matter how many so called computer-experts explain that it is oh so difficult... Try playing the game you talked about in a heads-up setting, before making educated guesses

Zach:

-What about situation-specific programs? Most of what you talked about is having a general GTO-ish bot it seems, but I don't see why there couldn't be programs that take specific inputs (your range, his estimated range?, these flops etc) and spit out far better estimates than humans, or at least very useful info that humans couldn't suss out by themselves.

-I think it's more likely than you that people are pursuing this privately. I absolutely agree that these 3 sickos (and others of similar skill level) are where they are because of hard work and not cheating. I don't see reasons why people wouldn't be hiring developers to build programs for them though. I know if I was a professional playing for such big money I'd be finding all sorts of specific (even very simple) problems then trying to get programmers to come up with dynamic solvers for them. I don't consider this cheating if you're not using it in real time.

I definitely agree with your general point though, and would be interested in what you have to say. I'm a programmer by trade but know next to nothing about AI/ML and don't have any friends in the field either.

edit: also I think you overdid it with the 'computers aren't that powerful' stuff. Despite all of that, and despite you only writing things that can occasionally help you out, there have still (almost certainly) been bots capable of beating 5/10nl right? Surely that amount of expertise and resources applied to programs that are simpler and more aimed at aiding humans in study would be pretty deadly. Just imo.

You must be new around here. Welcome to NVG !

Probably overlooked the word 'gossip' in the forum title

I guess the real question is, If Sallywoo was out in the forest using an 08 program and Gus wasnt there to play her, would she still be turning a profit?

Maybe I missed it but I haven't seen it mentioned that maybe Gus just fell into the fairly common pattern of playing an opponent that has an edge on you, then running bad which leads to tilting and playing bad, which leads to the opponent having a bigger edge, then running bad again, which leads to tilting and playing even worse, and so on, rinse and repeat, resulting in a perfect storm death spiral. I mean, they haven't even played 100k hands against each other have they?

Fwiw I did realize it was limit not pl I just mistyped. That said I know just as much about plo8 as I do about lo8 and that's the rules but little to no strategy at all so your post is well taken.

Like I could be wrong and welcome proof from those that are knowledgeable. It's just really annoying to see people who don't play poker on a high level or know anything about programming/ai to be posting about "zomg nerds making programs to cheat Gus and other real pros out of money". If they were able to write programs that in itself took a ton of work and is extremely admirable because it goes beyond most phd-level ai and machine learning based studies. Solving poker is not something you could just pay some guy in India $2/hour to program, and it seems the average nvger doesn't get that.

naah. a bot does not have to know the holy grail. it only has to be better than $HumanPlayer. please show how all these bazillion of game states are easier to compute for represent in a human brain.

Look into professional GO (an asian board game)

I think your original post on the difficulty of creating such a program was spot on. Especially when it comes to the part about players who are already at an extremely high level in terms of game theory and application deciding they need to cheat by hiring someone to create what would in fact be a very difficult program to write with no initial guarantee that it would actually perform better than they already do.

It then takes an entirely new jump in logic to assume that these players would then need to use it against poor old Gus in the midst of his 10 mil losing streak.

Gus needs to get a lot more specific as to what he's accusing Sallywoo of. Otherwise the speculation in this thread is likely to get absolutely ridiculous. To see what some people are willing to believe is happening at this very moment with poker/computer programming, people here should take a stroll through the Is Poker Rigged thread.

is go played online ..for serious money? then i predict that you will be proven wrong in very short time (assuming you are right now).

nice post!

I stopped reading here. You obviously are full of ****.

Why do you think the software needs to be developed?

Database.
Decision tree software.
Probability calculators.
Quantitative analysis.