The birth of AI poker? Letters from the 1984 WSOP | Hacker News Reader

The birth of AI poker? Letters from the 1984 WSOP

poker.org

55 pointsindigodaddy26 comments

It still pains me that in Colorado one has to play on a sketchy unregulated poker site against what is likely a large number of bots while people can gamble on the lottery, play daily fantasy, and sports bet.

I miss pre-black friday.

feels like solvers killed online poker and it can't come back. Theres just no technical solution to prevent using solvers to real-time assist

That being said , its kind of the golden era of live poker right now. Games are growing everywhere.

I really doubt live poker is as big as it was twenty-five years ago, when Phil Helmuth was a household name and Hollywood were casting Matt Damon in movies about the sport.

Live poker is significantly more popular now than ever. Every major tournament has seen record participants, Vegas has bigger poker rooms than ever before, and I'd say anecdotally local poker clubs are packed compared to anytime I can recall.

that's a shame, the coverage is 100x worse than it was.

the ESPN2 streams suck, they seem like they don't know what table they're watching and the commentary is usually below-basic pop-culture and memery, and the WSOP commentators are equally childish and unprofessional.

poster was right though, it seems far from what it was as far as general non-poker interest goes.. maybe the increased size of the poker hall/tournament attendance is evidence of another effect; gambling tends to go up in poor economies.

my .02c: i've seen a lot of my favorite casinos close their poker rooms or convert them in the past five years. my neighborhood games are all mostly dried up, and all of my cohort I network with about poker stuff is essentially still just enjoying 10-20 year old Poker After Dark eps. The coverage sucks and only the huge games or private tables are worth watching, and that's a whole other cash grab. The personalities are largely non-existent, and the ones that try angle don't do that great a job.

It all sounds like sour-grapes nostalgia, and maybe it is to a degree, but it's a common opinion that poker is in a rut lately.

ESPN2? I thought the live coverage is only on PokerGO for the last few years, with the packaged shows broadcast later on CBS Sports channel?

We must have been frequenting very different households.

It was people of a certain age and mindset I think.

At the peak before black Friday, it was pretty routine for 3 or 4 people I knew from work to be on Full Tilt at the same time and I only really knew about 15 people at this company.

Isn't it trivial for online poker providers to cheat, i.e. manipulate the cards you receive, and have a fake bot player at the table that can be made to win, etc. ?

Short term, yes.

Long term, people store their hand histories and this shows up plainly in analysis.

That only catches a subset of ways online poker rooms can cheat.

The server knows what cards everyone is holding. Even if the cards were randomly assigned and weren't changed after the fact, users have no logs of the order of cards remaining in the deck. Its pretty trivial to have software that selects community cards that usually lead to a larger pot.

Wouldn't that show up in a statistical analysis of the community cards? How is your algorithm modifying the community cards advantageously but preserving randomness such that over a large sample size every card shows up at the same frequency? Although it wouldn't be exactly the same, presumably some cards that are less often bet preflop, like a 2, would show up at a slightly higher frequency in the community cards, but still.

The much simpler way to cheat is to just give some players more information. Or, run bots that take up guaranteed payout seats in tournaments and such, which I've heard rumors of happening on certain sites. Or both.

Yes, statistical analysis would reveal unfair deals of any variety.

And yes, both the types of cheating you've mentioned have happened.

Some players getting more information is called "superusing" -- see Absolute Poker scandal.

Empty seats being filled with bots in tournaments with a guarantee -- ACR

Fake players or predefining winners would work as well.

My point was simply that an online casino could seem completely legit even if you can compile audit logs of every players' hands at the table. Controlling the community cards is completely undetectable and more than enough to push larger pots, and therefore larger rakes.

It's not undetectable.

Most pots already hit maximum rake.

How would one detect it?

As far as I'm aware, you would have to know the full list of cards in the shuffled deck before the hand was played to know they didn't change the community cards.

Because it shows up in analysis of millions of hands.

The community cards are subject to the same frequencies as the hole cards.

(And you can see them more often, so they're actually easier to analyze)

That's not exactly true. It's a non-trivial but not exactly difficult task to design a fair shuffling cryptographic protocol that every participant can validate after the fact.

On the other hand, that still doesn't prevent cheating in the form of the server providing information to some participants via a different channel. There's nothing cryptography can say about out-of-band communications.

So maybe fair shuffling is cute but ultimately pointless.

My point wasn't that a fair, auditable system couldn't be built. Only that we don't have that today, and I'd add that online casinos are incentivized to not build that.

I think you can make an analogy with Casinos and their incentives to cheat. They could do all sorts of things and there are plenty of gambling scams that do those things, but most legit looking Casinos are already making money hand over fist, have published odds on their favour and against yours, can kick out anyone who seems to be doing well, have all sorts of non-cheat tricks to squeeze money out if you, and are risking serious reputation and legal damage if found out.

Doubtless there are scummy poker games, but for most of them the money comes plenty easy, and the existential risk of faking cards to increase pot sizes just isn't worth the marginal benefit.

Of course everyone, winners, losers, and impartial bystanders will see these patterns in completely random deals so every site will be accused eventually.

That's very optimistic :)

Any online or electronic gambling. Any at all. You have to expect it to be crooked.

This applies to sanctioned sites, sketchy sites, or physical machines.

The incentive is just too damn high for it not to be a cheated system made up of black boxes.

Some of the cryptocurrency casinos pioneered having cryptographically signed random sequences that are revealed after the game is over. That way you can confirm that the game was fair. It's not a very popular feature, however, as it's not a major selling point for most people.

I fail to see how that helps considering all digital casinos likely use a similar form of pseudo random number generation and the crypto "guarantees" won't prevent people from using verifiers during play.

That only prevents a small percentage of ways to cheat.

I use to play on Full Tilt with guys from work all the time.

Quite often I would be at a table with someone I know and chatting on the side.

We wanted to beat each other though for bragging rights so never colluded. Thinking about what could be done though if sharing card information is really bad. There are so many spots that knowing for certain 2 cards have been removed from the deck would be an absolutely massive advantage.

I can't even imagine the schemes real cheats have come up with.

Reply on news.ycombinator.com