Quote from: arisatox on December 25, 2017, 04:42:20 PM

Quote from: Hell-raiser on December 25, 2017, 07:52:04 AM

Quote from: MadZ on December 24, 2017, 11:11:40 PM

Perhaps I missed it in the article, but these hands were played online, not live, right?

Also, I think it's pretty interesting that it wasn't until the 30k hand mark that the AI started to become consistently profitable, and it wasn't until 80k hands played that this profitably was statistically greater than 0. I think this suggests that the AI's ability to win isn't from some sort fundamentally better strategy against all opppnents, rather, its improved ability to learn from opponents and exploit them over a large sample size. This bot can't just sit down at a random online table and start crushing, which is a good thing.

Yeah, it may well be so for the first time, but after a few months of continuous playing the AI should be able to squash all competition at once, at the first game. It is basically the same with chess, though AI doesn't seem to be used there so much. Chess bots use, for example, opening and endgame databases so they don't need to play every game from a cold start, so to speak. This roughly corresponds to the level of AI "expertise" in playing poker acquired after a number of games.

This is what we humans call experience.

It can do well as a tight player setup but people bluff and it has to learn how each individual plays before doing well against them. Chess is all open and public, while poker is mostly hidden and is about tells, so they're two different things.

It may be so in a face-to-face game, where physical presence is required, but poker is played mostly online these days, as many other games, for that matter. So how are you going to bluff in such conditions? Further, it actually cuts both ways. Just as you can say that players can get advantage by reading their opponents' facial expressions, there is no "face" to read in AI. While the AI can still do that.

Are you sure you will be able to deceive it?