Will there ever be really great poker software ?

[Walter Pullis]

I own just about all major poker software. At best, it has only helped my game a little. I can't believe, in an era where a computer can beat a grandmaster at chess, that better software is not available. I would be willing to pay as much as $1000
for software that could really help my game.
Developers, get busy!

[KenProspero]

I'm not sure that Poker (or Bridge) aren't harder than Chess.

When you think of it, Chess is an 'algorithm' not a 'game'. By way of example, think of Tic-Tac-Toe. We call it a game, but in fact, when you analyze it, there is always one or more 'best' move. Learn the move and you win, or draw every time. Furthermore, it's a simple algorithm, I hear that there's a chicken who hasn't lost a game in years.

Chess is really the same thing -- there is no hidden information. Though it's millions of times harder than tic-tac-toe, there is always one or more correct moves. So, given a computer that can crunch enough calculations and do the crunching faster than a human mind, and the computer should win.

Poker, on the other hand is a game of imperfect information. When you make a bet, you usually don't know whether you have the best hand. You can estimate probabilities, but you may be wrong. Further, a certain percentage of the time it's optimal to make the 'wrong' (i.e., bluffing). So, what you have in poker is a situation where you need to apply game theory to come up with an optimal strategh.

This being said, do I think it's possible to come up with a computer program that could beat a good human. Yes, probably, however the task is a lot more daunting than it looks.

[klange]

I think my previous post disappeared (if not excuse the duplicate).

Ken you bring up the pertinent points in developing strong computer AI, the key to those is the imperfect information aspect, this makes poker a VERY difficult AI game to tackle. But this is part of the fun and challenge, we work on this everyday along with the University of Alberta and this is why we have so many bots, different styles, etc...

Poker adds another interesting element, short term luck. The best player can lose to an average player if they are unlucky enough, granted that in the long run the better player will come out a head. This factor adds a lot of excitement to poker, it gives most people some short term shot at success. Most of us couldn't beat Jordan one on one in basketball ever, even on his worst day.

Whether there will ever be bots that beat good players is open to interpretation, in the long run our good bots have winning records against the masses. Are the masses good? I'd say Yes, on average. Are they very good/great, i don't know?? Can vexbot beat very good players, yes, when played enough hands. Can it beat the elite? hmmm...

Anyway, Yes would be the bottom line answer, eventually bots will be winning players at any level. Remember Kasparov NEVER thought he could be beat by a computer.

Is this good or bad for poker might be another question. Used correctly it's good for poker, used inappropriately it could be bad. We for one company intend to use it appropriately.

Kurt
www.poker-academy.com (http://www.poker-academy.com)

[Evan]

[ QUOTE ]
I hear that there's a chicken who hasn't lost a game in years.

[/ QUOTE ]
There used to be in Chinatown but its not there any more.

[Luv2DriveTT]

Mmmmmmmm Chicken!!!!

TT

[Luv2DriveTT]

I'd like to add that although Poker Accademy (poki) is a great trainer application, it's real strength for good and semi-good players is the ability to create defined scenarios to replay over and over again. 2 days ago I played 72s from the small blind for two hours with Poki... against all types of simulated opponents (which seemed very real to me). This exercise game me the understanding just how to play this tricky hand safely and aggressivly at the same time. My results? It was a statistically break-even scenario. But I walked away with my heald held high, knowing I've accomplished something tremendous!


TT /images/graemlins/club.gif