Or maybe not. An image that a neural network fails to recognize need not to be an unpleasant one https://towardsdatascience.com/break...s-f4290a9a45aa

Still, it'd be better to have half the recs than all netowrks shut down

Protect Ya Necks Kiddies

If that's a PM you received on 2+2, please forward it to me.

wow
online really is on it's way out if this is accurate

Cross-posting this recent Twitter thread here.

I Turned £1 Into £220 with the Help of an Online Poker Bot (by Hayden Vernon for VICE)

https://www.vice.com/en_uk/article/b...eview-winnings (Jun 26, 2020)

"So, without the patience or aptitude to beat all the Russian quantum mathematicians hanging around online poker rooms these days, I turn to more nefarious methods and download a shady-looking poker bot to play for me. The website I get it from looks like it's going to give my computer hepatitis, but the bot's easy enough to use and even comes with different profiles depending on the style of game you want to play..."

"I begin by using the bot in some small stakes cash games on Bet365. Progress is steady – it doesn’t seem to be doing amazingly well, but it's not making stupid bets like I did. I double my £10 deposit after a few hours and peak at £60 a couple of days in..."

"To play for money on PP Poker, you first need to find an agent who will let you buy into a club. It takes me a little while, but I search a few poker forums and find one based in the Philippines that’s accepting new players. I exchange £30 for 1,800 chips for one last blowout with the bot – then quickly wish I'd never bothered. No matter what I try – low stakes, medium stakes, tournaments – the bot loses. My chips soon vanish and it's time to call it quits..."

Tweets by: David K. Lappin (Unibet Poker Ambassador), Jon Sofen (CardsChat), TheGameKat (RedChipPoker)



---

Tweets by: Hayden Vernon (VICE), Haley Hintze (Kick Ass Poker), Jason Wheeler (poker pro)




_____

Tweets by: Christian Zetzsche (PokerNews, WPT live events reporter), Hayden Vernon (VICE)

lol actually a good comeback

hahahaha a great comeback indeed

It's almost like you're a shill trying to distract from how obviously viable this is.

You don't think people using solver software can make slight alternations in play to "throw off" the site? Not that they would even need to, because maybe they're also using a program that recommends adjustments to solver strategies based on reads.

This is basically what all players are doing in a roundabout way when they look at their hud stats and formalizing it into a program isn't exactly rocket science.

You could do this for every point on every decision tree. With enough hand histories you could inevitably spot holes in almost any humans game and you can make the program basically tip you off whenever you're in one of those spots where they've shown the tendency to deviate from gto to a statistically significant degree. And sure, ebb and flow of the game matter, but it doesn't have to be perfect (although you can absolutely measure, say, a persons inclination to bluff after having a bluff get called).

You don't even need a huge sample size on any given player - you can make informed estimates based on analogous situations or supplement with hands from players who have very similar hud stats.

The biggest reason id be skeptical is because there just isnt that much money in poker these days and regs wouldn't play them. They'd basically just be like the other regs, sitting around waiting for a single fish to show up, and then maybe they'd skin that fish slightly faster than the others.

That's not really how it works. A perfect GTO bot could tell you its exact ranges and frequencies for every spot and you would still never be able to beat it.

Imagine if you play Rock-Paper-Scissors against and opponent who upfront tells you he will completely randomize his strategy to play rock 1/3, paper 1/3, scissors 1/3. It doesn't matter how "unpredictable" you are or what "reads" you have on him, you'll never win over a large sample.

Same thing applies to poker except on a much more complex scale.

I heard there's glasses 4 live poker that show you your opponents cards and the board pre flop when you put them on.

At the end of the day he's still a Vice "journalist"...

vice journalist is an oxymoron

If people are succeeding at cheating at live poker (and not just Postle) it is incredibly naive to not think that a vast amount of winning players are cheating online. And using programs that practically make all your decisions for you is cheating its not just a tool like a HUD. We have seen all sorts of cheating in poker, be it Postle, Potripper/Russ Hamilton online, those 2 jerks playing that live final table at Partouche I believe who were signalling every hand to each other etc. It wouldn't surprise me if nearly every poker room had players colluding/signalling hands/soft playing or straight out cheating via marked cards etc. There is a reason why numerous people have been victims of violence over card games/gambling. Whether deserved (most of the time) or not.

What no one posted yet is the following. What if some of the top top pros, cheated in the last couple of years with the realtime charts/assistance that we have seen. When you do this for 2-3 years, everyday, 24/7h eventually you will just get really really good, caus you passivly will learn the gto freq etc.

Maybe some of the pros just stopped using the assistance and become really good. I would already have some names in mind, but just guessing so obv makes no sense of discussing.

imo this is what happened to some regs.

Just imagine playing everyday with the gto solutions in front of you, 500k hands over a time span, and then you will have learned on the side, without the need of grinding pio solver after every session.

albert, this has actually been discussed ad nauseum to the point where few of the top players haven't been accused of that - particularly the heads up ones

sauce, haxton, & polk specifically have had a ton of dream machine accusations sent their way

Just remember, all those long pauses that suddenly became prevalent among the poker upper echelon were "so they didn't give off timing tells".

i don't think it's a fair comparison to something like hand signals.

i think risk of getting caught is a pretty big factor and by that standards using real time solver software is likely to be far more common, where the only added barrier is the complexity of creating a rig that can solve hands in real time.

Never understood the whole real time solver thingy. It's so much more easier to just solve as many boards, betsizes, turns and rivers as possible, save all this in a database, and then on the fly sjust show the charts with what to do, freq etc.

I wouldn't wonder if in some Russian poker lab there are all already these databases with 10000flops turn river solved. Yes I know you only need to solve a certain subset, but if u have the resources why not solve for all of it.

Yes much easier, 8 tabling and always look up the solution instead of having it displayed in an instant on the table.

Congrats for that comment!

i wouldn't be so sure that it's easier. someone whose good with hardware may be able to put together a rig that can do a full solve in 10 seconds for 10-20$k within a few days. Running sims and categorizing them is obviously not that hard, and i'm sure some people do have "cheat sheets" for something like sizing on different board texture for different stack to pot ratios, but doing this for anything close to an exhaustive set of criteria that you'll be faced with would be very very time consuming, and would never be perfect.

the difficult is all in building hardware to run it quickly enough. and i have no idea how difficult that is, but i would bet the farm that you can build a computer with the tech that's available today that would do it quickly enough to use the results in real time - it may just be very expensive.

are u unable to read, no queno english? "display the charts" u clearly have no clue but ok thx for that valuable comment

he's saying that the real time solver woudl give you exactly the answer you were looking for, as opposed to having to sort through a large numbers of charts to find the right one... which would obviously be very difficult since you couldn't summarize all possible info on anything close to one page.

when you actually look at all of the things that are factored into why solvers do what they do it becomes obvious why this is not an easier way of doing it.

Yes, I'd say so easily. Seen a guy on a forum who has a custom rig (circa $23k), he is a consultant data scientist. It is used for software development for machine learning. He was spending $10k a month with AWS to do comparable work. I was looking for (a lot cheaper!!) ideas for a home build as I do some data analysis as part of my work, my laptop hates it.

The rig he has built I wouldn't even know how to get max value out of. I am sure if someone used a machine like that for real-time poker work it would meet mission, and I am confident people do, and have been doing for sometime. There has been too much smoke around it over the years for there not to be a fire.