|
|
#141
07-24-2007, 03:26 PM
|
|||
|
|||
Re: Second Session Just Started
12:16pm Phil gets 72o on the first hand and folds preflop to Polaris' raise.
12:12pm Jonathan is giving the introductory talk so we'll be under way shortly. 11:38am Okay, your resident blogger is back. The room is quiet at the moment. The UofA team is deciding what bot they should field in the next session. There's some match strategy that goes into making this decision, so that's what is going on right now. Ali showed up a little while ago to ask some questions. Phil and Ali received the logs and some of our analysis (DIVAT) to evaluate their game for today. The problem with poker is that it is so noisy that it is difficult to really understand where you're at. Since we want the humans to be at their best, we gave them the logs and analysis so they could be as prepared as possible for the match. The match should run on time in about 20 minutes. 
|
|
#142
07-24-2007, 03:35 PM
|
|||
|
|||
|
Re: Second Session Just Started
Damn, I want to play against these calling stations.
12:34pm Phil's bankroll after 50 hands: -$15. On hand 34, Phil flops a set of queens and gets paid off to the river. Polaris shows just ace high. 12:26pm Phil's bankroll after 25 hands: -$155. Early in this match, Polaris seems to be getting the short end of the cards. He's gotten several straight and flush draws but none of them hit by the river and he was forced to fold to river bets. On hand 12, Phil finally wins a hand where he had top pair of jacks and Polaris bluffed into a flush card with an underpair of threes. Phil seems pretty timid early on here -- not a lot of raising going on. On hand 17, Phil rivered trip nines and got a call from Polaris' holding jack high. Two hands later, Phil tries to check-raise the river holding top pair of jacks with a busted straight draw, but Polaris didn't bet out with his pair.
|
|
#143
07-24-2007, 03:52 PM
|
|||
|
|||
|
Re: Second Session Just Started
To be fair Polaris rivered a pair of deuces on that set of queens hand.
Also tilly is sweating it up in the third row. I was begging for a kkJ board so I could yell, " I THOUGHT HE HAD POCKET KINGS"
|
|
#144
07-24-2007, 04:06 PM
|
|||
|
|||
|
Re: Second Session Just Started
shameless plug: bu there is another live blog going on Poker Academy's site:
http://www.poker-academy.com/man-machine/live-blog.php
|
|
#145
07-24-2007, 04:30 PM
|
|||
|
|||
|
Re: Live Updates of the Laak vs Computer Match HERE
[ QUOTE ]
2000 hands against one guy won't prove much... But as Todd explained... This is really more of a scientific research experiment... Than an attempt to prove anything. The WSOP FT winner was decided by about 300 hands. These Alberts guys are the leading edge... In certain specific poker algorithms. That said... I have probably read a dozen papers on AI and poker... And, in general, many academics based conclusions on fairly small samples... Because they are not committed or creative enough to generate very large samples. [/ QUOTE ] does this form of computer speaking remind anyone that teacher from southpark? all it needs is mmmkay instead of the three periods at the end of each sentence
|
|
#147
07-24-2007, 05:25 PM
|
|||
|
|||
|
Re: Live Updates of the Laak vs Computer Match HERE
[ QUOTE ]
can anyone comment on the accuracy that duplicate will elicit? [/ QUOTE ] I'd be interested as well. Someone chimed in way earlier in the thread to basically say that the results will be conclusive, but they didn't say much more. It looks like U of Alberta has put a considerable amount of effort into analyzing variance and thinks the results will be conclusive. It looks like the guy who is doing the updates has a 100 page Master's thesis on it on the site. Quote from the thesis: [ QUOTE ] In this thesis, a statistically unbiased estimator, called... (DIVAT) ... is described. The tool provides post game analysis of headsup Limit Texas Hold'em decisions, while filtering out most of the luck elements of the game. [/ QUOTE ] The thesis has lots of pretty graphs and stuff so maybe someone involved wants to comment, or else maybe someone here can read it and report.
|
|
#149
07-24-2007, 05:30 PM
|
|||
|
|||
|
Re: Live Updates of the Laak vs Computer Match HERE
[ QUOTE ]
[ QUOTE ] can anyone comment on the accuracy that duplicate will elicit? [/ QUOTE ] I'd be interested as well. Someone chimed in way earlier in the thread to basically say that the results will be conclusive, but they didn't say much more. It looks like U of Alberta has put a considerable amount of effort into analyzing variance and thinks the results will be conclusive. It looks like the guy who is doing the updates has a 100 page Master's thesis on it on the site. Quote from the thesis: [ QUOTE ] In this thesis, a statistically unbiased estimator, called... (DIVAT) ... is described. The tool provides post game analysis of headsup Limit Texas Hold'em decisions, while filtering out most of the luck elements of the game. [/ QUOTE ] The thesis has lots of pretty graphs and stuff so maybe someone wants to read it and report. [/ QUOTE ] the computer and the human team plays each set of cards against the same set of cards. So one side cant get a good or bad run of cards.
|
|
#150
07-24-2007, 05:32 PM
|
|||
|
|||
|
Re: Live Updates of the Laak vs Computer Match HERE
[ QUOTE ]
the computer and the human team plays each set of cards against the same set of cards. So one side cant get a good or bad run of cards. [/ QUOTE ] Yes, but for instance, Phil Laak suddenly decided on Hand 250 that he was going to be aggressive for awhile. Presumably Ali is doing his own thing. What if Phil's side gets a good or bad run of cards at that very moment? Since they have studied this, presumably U. of A. has some reasoning as to why this will not affect things.
|
 |
|
|
Linear Mode
