The bots are coming!

[Paul Talbot]

Does this worry any one else?

bots for on-line poker

[AAmaz0n]

Yes and no. I played against bots on the old IRC poker site. Most of them were awful, although r00lie was pretty good and could even chat a bit. If you came in with a weak hand that got showed down he would go "Bwahahaha, call 2 bets cold with K9? What a fool!"

Unless they are built with some good adaptation algorithms, they are bound to be predictable; a good player should be able to spot patterns against them.

I'm not sure how sites would be able to keep them out, maybe some kind of conversational Turing test to prove you are human?

Shauna

[scotnt73]

to blend in with party players all they need to do is buy one of homer simpsons drinking bird thingys placed over a mouse button.....call...call...call...call...call...bird falls over and you time out(DOH!) [img]/images/graemlins/grin.gif[/img]

[kg99]

haha, that's funny. one can get made fun by a bot...

[Lori]

No.

Bots suck at poker.

There are lots of references in the archives, which ill dig up when im not having dinner sometime.

Lori

[CORed]

[ QUOTE ]
Bots suck at poker.

[/ QUOTE ]

I think perhaps that should be modified to read " All bots designed so far suck at poker."

Because I am both a programmer and (I like to think) a pretty good poker player, I think the task of writing a 'bot that could play well is formidable, but not impossible. Calculating pot ods, etc. would be pretty easy. Where it gets a little trickier is that any 'bot using a fixed strategy could probably beat fish, but any decent player could proabably beat it, whether he knew it was a 'bot or not. Using game theory and a random number generator to mix up play would help here. The most difficult (and key) part would be adapting to opponents' play.

OTOH, i don't think it would be too hard to write a 'bot that plays well enough to make money off of, say Party .50/1.00 or 1/2. If you had enough accounts, and enough 'bots running, you could possibly make a lot of money.

[BreakEvenPlayer]

Pokerroom already has a nice selection of bots running in the $1-$2 games. They are all from various suburbs around Denver which makes me think that some smart kids from Colorado figured out a way to do it.

These bots are in any of the full ring games 24 hours, 7 days a week.

Maggie500
Softplay
Lunchmoney
Bangkok27
Opponent
Zagnut
TETRIX

I had doubts that they were bots for a while but as I observed it became more and more obvious. Besides the fact that their play is ridiculously predictable, they never chat, always auto-play, never were in the same room together, etc. There was one key thing I found out: Sometimes there connection would go out, and instead of clicking back in like 90% of players, they would be stuck out forever, until pokerroom booted the idle player. Then they would gone for a bit until whoever was running things could get them back into a game.

[Webster]

Not worried at all!

[fluff]

I was recently wondering how a Bot programmed with something like Sklansky's System (very simple algorithm to program) would fare. There are tons of NL hold'em freerolls (or almost free rolls) out there to probably make it worth while.

[TomCollins]

I have tried this system in a free roll before, and it did surpsingly well (I thought I would get called a lot more since it was a free roll). I ended up in 4th out of 400 players simply by going all in or folding. The one part of the strategy is if you in the big blind, and everyone limps to you. You see a flop... then what? I would go all in on TPTK or better, and fold the rest.

I was pretty bored, had already signed up for a free roll, and had other stuff to do at my computer, so figured why not. So I'd imagine the strategy would probably be a lot better than a lot of the fish I see in NL tournaments, but long term no better than break even.