Unsettling Bot News

[mat]

isnt turbo texas holdem just like playing against a bot that game is very beatable i think for most decent players even on a tough line up

[D.H.]

[ QUOTE ]
isnt turbo texas holdem just like playing against a bot that game is very beatable i think for most decent players even on a tough line up

[/ QUOTE ]

Sure, that is a kind of bot. But saying that one bot is beatable doesn't really say anything about bots generally. Just 'cause you can beat John Doe who is a "poker player" doesn't mean that you can beat Gus Hansen who also is a "poker player".

[jpym84]

As far as I know, the University of Alberta has the most advanced poker bot out there (see http://www.cs.ualberta.ca/~games/poker/ and http://games.cs.ualberta.ca/webgames/poker/ ). The bot is surprisingly good, but definitely beatable, so I don't think theres anything to worry about just yet.

[tadams]

Programs that play poker well have been researched for some time now and there has been a lot of progress on them.

If you want to see how well a program can play poker go to:
http://www.cs.ualberta.ca/~games/poker/

However, it would be unethical to deploy such programs at a poker site without disclosing the fact that they are programs. In fact, no online poker site would knowingly allow such programs to reside on their site undisclosed. Why? It is simply not good business. If it was found out they would lose all of their human customers. However, sites may allow disclosed bots, but they would most likely be controlled by the site and not playing on real money tables. They would be used as advertising to get people to come play at their site. Online casinos already make plenty of money from the rake and there is no incentive for them to allow undisclosed bots. Therefore, you do not have to worry about online poker being taken over by bots.

[jpym84]

Lets just be glad that Darse Billings is too busy researching Ro Sham Bo to improve on that bot. Otherwise the online poker industry could be in some real trouble.

[Kevroc]

True, I must confess that I chose the subject title "Unsettling Bot NEWS" knowing full well that this wasn't news. It was just something I heard and was a bit bothered by it. I went with that subject title to get people to read it and respond...

It worked and I thank all of the posters for their input, I am optimistic that the games will stay good and remain honest. Also, when we say bots it doesnt hafta be a guy setting one up at home and having it play for him. It could be the websites themselves doing this.

Another side note... a player at Pacific Poker told me that the hands are rigged to induce action (such as whenever AA is out KK is or flush flops always pair the board on a later street so the flushes lose to boats etc.) This is probably just some guy that got drawn out on once to often but, hey whaddaya think about it??

Thanks!
KEVROC [img]/images/graemlins/laugh.gif[/img]

[BigBaitsim (milo)]

Now I gotta ask, you got notes on BigBaitsim or BigBeitzim?

[M.B.E.]

[ QUOTE ]
Lets just be glad that Darse Billings is too busy researching Ro Sham Bo to improve on that bot. Otherwise the online poker industry could be in some real trouble.

[/ QUOTE ]
Does Darse Billings play online? I've always wondered whether Spy328 on Stars is Darse (or someone else in the UofA group).

[BusterStacks]

Eh, ok some things I'd like to reply to. First of all, to any doubters, a bot is most certainly possile and very lucrative. The only real downside is the amount of integration and calculation required to make the bot a winning player. To the guy who said it would only be possible with "screen scrapes", um, you are retarded, turn on full chat and post a window of your game. You will see that the cards that are delt are in easy to understand text format, as well as all moves by all players. I mean, it's cool that you don't know about programming, but seriously, let's not give too much credit to the makers of poker software. It's horrible and absolutely buggy. I'm no expert in visual basic, but if I were to make a simple bot, here are the FACTS of what it could do:

a) run in real-time and undetected

b) incorporate specific fields from external programs (pokertracker) and utilize the data as decision making influence.

c) Keep a small database to interpret table texture and playing styles as well as recent trends.

d) make an educated move after considering hundreds of times more factors than any human could.

That being said, I think an actual well-programmed poker bot would beat just about anyone at any game. If you think you can trick a bot, you are fooling yourself. You may win incidental victories, but let's think about this for a second: Poker is not an abstract art, it is first and formost a game of technical strategy. The conclusions you draw about a player or playing style are based on static events that could easily be calculated by a machine. At each intersection requiring input, you have an amazingly small number of options: 3 (fold, call, raise). Now given that a computer can remember every single move it has ever seen you make just as well as the previous hand, do you actually think you could hope to beat it? I don't think so. A computer would would walk all over unless you changed gears literally every hand, at absolute random, at which point you would be playing incorrect poker in an attempt to throw the bot off your scent. I know we would all like to think that a person has some intrinsic ability to do something that a computer cannot, but my friends, it ain't poker.

[Richard Berg]

I agree with BusterStacks. Some games like Go have incredibly expansive search trees even after applying the best-known pruning strategies (which typically derive only from expert contributors' gut feelings). Poker is the complete opposite, even NL. With such a limited state space, its primary heuristics lie in backhistory. In other words, establishing the basics is trivial, and past that point the player with the most information wins. That's not a task on which you want to wager against a computer.