|
 Originally Posted by boost
 Originally Posted by Lukie
Right, because poker is so incredibly complex that only humans can even begin to understand it's sheer complexity and awesomeness.
el
oh
el
Nobody who has posted in this thread plays better than an extremely well-written bot would. Period.
On a somewhat related note, there are chess programs available for download (e.g. Fritz) that can beat the best players in the world at chess, a game that has been studied IN-DEPTH FOR HUNDREDS OF YEARS. And don't give me that perfect information argument either, it's a load of shit.
I dont think anyone believes that it is impossible to create a bot, just a hell of a lot harder than it is to create a chess bot. A poker bot that could pwn the best players would border on true AI.
I disagree, especially on the AI comment. Researchers are having a hard time conveying commands to machines that a baby or a pet would be able to understand. Let's not go down the "true AI" route. We're talking about a game (specifically poker, but also applies to chess) that relies heavily on numbers and mechanics, something that computers happen to excel at.
Read: http://www.livescience.com/technolog...re-robots.html
Chess, in it's modern form, has been played for over 500 years. It is the most popular game of its kind in the history of the world. We're going to say that it was in the late 90's that chess computers started becoming better than the best chess players in the world. I'm going to use '97, when Deep Blue defeated Kasparov, 3.5 to 2.5.
Exciting implications:
The best chess players in the world are certainly better at chess than the best poker players in the world are at poker. Give poker 500 years and you'll be where chess is at today (of course it would not take nearly this long in practice, largely owing to the advent of the computer and modern society, but you get my point. I'm also ignoring poker's own history, shoot me).
My guess is that anything more than a passing thought about developing a poker bot never really occurred until online poker. In addition, the understanding of poker has improved immensely since that time.
The game of poker is in its infant stages compared to chess, which for our intents and purposes we can just consider solved. It's close enough. I think this is part of the reason why it appears so easy to create a chess bot relative to a poker bot. Chess has been around and studied for such a long time. You can read all about the history of computer chess if you want, but it paints the same picture. http://en.wikipedia.org/wiki/Computer_chess
Let's start with this. NLHE, there are 1326 card combos we can have. By default, this leaves villain with 1225 card combos that HE can have. After the flop, villain can have 1081 card combos. Let's cross check all of our possible holdings (adjusted for preflop numbers, of course) with all of villain's possible holdings and run the equity calculations for all of them. Boom, done in a fraction of a second. Of course what to do with all of this information is a challenge, but it's something a group of great players and world-level programmers could hammer out given enough time. Google could do it easily. This is already far more powerful than some goofy ass 80th level abstract thinking that you're giving our world-class poker players credit for with your AI comment.
I could delve far deeper into this but I'd rather not. Time, unfortunately, is going to prove me right on this one.
edited to add: It's an irrefutable fact that winning bots have long existed. I'm guessing the person(s) that designed weren't exactly my description of "great players and world-level programmers", or "google"... lol.
|