![]() |
#21
|
|||
|
|||
![]()
[ QUOTE ]
To any of you non-believers our there, below is a link of a paper written by 4 professors at the University of Alberta who wrote a computer program that plays world class poker. Hope this answers your question. http://www.cs.ualberta.ca/~duane/pdf/2002aij.pdf [/ QUOTE ] Do any of the HU specialist around here think this is actually any good? I think any decent short-handed player should be able to beat this thing in the long run. The results from there paper support this conclusion, and I think they have pumped up the ability of this bot somewhat in the press-releases etc. However, it is very, very good for a computer. This is the Heads-up bot here: http://www.cs.ualberta.ca/~games/poker/ Hint: It's over-aggressive with mediocre hands, this is when you have to make it pay. It will go too many bets with middle/bottom pair and high cards. It will call you down light sometimes. It has no memory, each hand is in a vacuum. Therefore it cannot hand read. If it could hand read, it would not be so overly aggressive with mediocre holdings. It seems to bluff at an almost perfect frequency. It is psedo-optimal from a game-theory viewpoint. Not optimal. I think there are some other exploitable traits.... |
#22
|
|||
|
|||
![]()
once u play enough hands with a computer u can see its pattern and then beat it easily. it wouldnt be able to successfully "change gears"
|
#23
|
|||
|
|||
![]()
Why not?
|
#24
|
|||
|
|||
![]()
yes,
1) it could use a random script, so that when there is a 50/50 decision, it just does something random. or 2) i hink computers are good enough that they could make it so you would have to play a lifetimes worth of hands to see a pattern. |
#25
|
|||
|
|||
![]()
The cpu would only do well playing against a 'good' player... if you matched the cpu up against a bad player I don't think it'd fair so well, jmo
|
![]() |
|
|