Poker Bots

[StacysMom]

There is a post where ppl are arguing about the ethics of using a bot. As a non-bot user, im sure u can guess what I think about them and there users. But, please don't argue ethics here.

It seems to me that playing against bots is a reality. WHat I am interested in is:

A, What is the best way to detect a bot? I have heard chat wtih them... but what are some signals that make you suspicous in the first place?

B. If you are confident u know a player is a bot, is there an adjustment u can make in your game to beat it? ie A strategy that is different from that used against humans.

[AncientPC]

A) When they don't engage in chat, but it could be just a non-English speaking player, a mute, or they could be busy multi-tabling.

Party Poker is going after the users of WinHold'em. I'm not sure exactly how but a few of them have had their accounts frozen.

B) No. As it is, WinHold'em allows for each player to program the bots playing algorithm. Unless you had a look at the bots algorithm you wouldn't know exactly what loopholes it has.

However, I bet that bots are very susceptible to bluffing when they don't have the nuts. Otherwise a bot that always calls with the second best hand is more than welcome at my table.

Also, because of bluffing, it is much more difficult to program a bot for NL as opposed to limit where the game is a bit more mathematical.

[emonrad87]

I personally could not be worried in the least about bots. First, I play on Party, where they are actively going after bot users (by taking screen shots of their screens). Second, I believe that bots are not great poker players because of the huge human aspect of poker. It would take a genius programmer a long time to build a bot that could play as well as a good human. And there just aren't that many of those out there. I'm not worried.

[Rudbaeck]

[ QUOTE ]
B) No. As it is, WinHold'em allows for each player to program the bots playing algorithm. Unless you had a look at the bots algorithm you wouldn't know exactly what loopholes it has.

However, I bet that bots are very susceptible to bluffing when they don't have the nuts. Otherwise a bot that always calls with the second best hand is more than welcome at my table.

[/ QUOTE ]

The basic setup of WinHoldEm is more NoFoldEm. Actually it plays alot like a maniac. It makes it's fundamental analysis on the assumption that all opponents hold completely random hands. You can tweak thresholds though.

You can tweak the ruleset, but it's hell to do so. No offline testing, no way to find out several parameters like position etc. I don't need to worry about an opponent who doesn't understand the difference between being UTG and on the button.

To straighten out the ruleset you need to be able to program without ever making mistakes and to boot you will have to do so in right about the dumbest homebrewed expert system language I ever saw. And then you have to live testing.

I'm willing to bet that almost all non-colluding WinHoldEms running out there are pissing away money, and that the colluding ones are only break even.

WinHoldEm has no way of knowing if the player who raised it on the turn is a maniac or a rock. With tptk and a small pot I'm probably reraising the first and folding against the second.

Now if a good expert system mated with PokerTracker and had an unholy baby I'd start to worry slightly. That could be made to play like most winning people here do. (We use well defined rulesets to derive our plays, and the variation comes mostly from profiling of the opponent.)

[Rah]

That's not true. Maybe they don't play as well as humans, but there are players who have programmed their bots to make 4 bb/100 on Party 0.5/1. If you think it requires a lot of skill to win at low limits, you are wrong. I'm positive you can make a few bb/hr just by only playing premium hands and TPTK and doing so aggressively until resistance is met and then calling down. This is not optimum strategy, but it would be profitable nevertheless on low limits.

[Voltron87]

I don't think bots are good enough right now to be consistent big winners in a variety of games and limits. I think right now simple tight bots can beat soft loose 2/4 3/6 games, but I am sure they can't beat people who read 2+2. Right now if you're a winning player I would not worry about bots.


I agree with the poster who is worried about a v strong player with computer skills making a bot. I would guess that it is possible for a professional player with a computer background to create a long term winning bot at loose 510 games.

[Rah]

I agree. I don't think the bots of today can beat higher limits than 1/2, but lower limits are populated by winning bots as we speak.

[G Money]

These bots were born from the concept of the Ho. A bunch of beeatches running around making money. I ain't ever seen a bot put a smile on a man's face though.

Word

G

[solucky]

Look to chess and look to backgammon. the bots are nearly unbeatable there. And the backgammon - bots have a brain. If someone have finished a neural net based bot that remember any of your moves bluffs..... it will beat you. And i am sure they are playing or will play soon. At sample www.snowiegroup they start to create one. The only option the sites must stop to support software like pokertracker.....

Wolfgang

[Exsubmariner]

As an aside to this discussion, I 4 table Party at the lower limits. Today, I thought I might make the leap at trying a fifth table. The software refused and gave me a message that 4 tables was the max. Looks like the days of bot 20-tabling are over at Party.