Two Plus Two Publishing LLC Two Plus Two Publishing LLC
 

Go Back   Two Plus Two Poker Forums > >

News, Views, and Gossip For poker news, views and gossip.

Reply
 
Thread Tools Display Modes
Old 05-08-2015, 10:33 AM   #1276
mcb08
grinder
 
mcb08's Avatar
 
Join Date: Nov 2008
Location: Not where u think I am
Posts: 434
Re: WCGRider, Dong Kim, Jason Les and Bjorn Li to play against a new HU bot

The prof is on the stream with WCG, trying to explain why Claudico folded A4o to Doug's AIPF shove (~9K). Doug is trying to explain that it's never right to fold A4 in that spot, and the prof is trying to say that Claudico made the right play, because Doug's playing of his hand (99) broke the NE. Hopefully Sam can explain it to him later.
mcb08 is offline   Reply With Quote
Old 05-08-2015, 10:52 AM   #1277
ArtyMcFly
Carpal \'Tunnel
 
ArtyMcFly's Avatar
 
Join Date: Dec 2014
Location: Enchantment Under the Sea
Posts: 13,251
Re: WCGRider, Dong Kim, Jason Les and Bjorn Li to play against a new HU bot

Quote:
Originally Posted by restorativejustice View Post
I agree, but the last interview I saw with the Professor that I posted about he seemed to redefine things, ignore the mirroring, and say the contest was "fairly even" and that the bot was in "fourth" ahead of Jason.
I suppose he forgot to mention it was also in sixth, seventh and eighth.

Academics gonna academic.
ArtyMcFly is offline   Reply With Quote
Old 05-08-2015, 11:34 AM   #1278
e i pi
veteran
 
e i pi's Avatar
 
Join Date: Feb 2010
Location: ☃
Posts: 2,230
Re: WCGRider, Dong Kim, Jason Les and Bjorn Li to play against a new HU bot

Quote:
Originally Posted by sam
Nash equilibria are theoretically not guaranteed to exist in games with imperfect recall (even though they exist for full NLHE which has perfect recall), and there are no theoretical guarantees on our equilibrium-finding algorithm for them. But it does well empirically.
Quote:
Originally Posted by punter11235 View Post
How do you know how it does empirically if you don't measure how well they do at what they are supposed to be doing?
I think he means their algorithm empirically does well finding the equilibrium for their abstraction but there is no proof that they will always exist to be found.

I suspect your algorithm for finding max exploitability would be useful, using a few fixed bet sizes, but not rigorous enough for academia. All your own arguments about claims without rigor would equally apply to your own metric.
e i pi is offline   Reply With Quote
Old 05-08-2015, 12:00 PM   #1279
mack's
enthusiast
 
Join Date: Apr 2015
Posts: 65
Re: WCGRider, Dong Kim, Jason Les and Bjorn Li to play against a new HU bot

and here we had geniuses(tultfill) that play 25$ nl saying claudico is **** constantly and now mr rider confirms that he beat claudico for less than he beats most of the top regs he plays online..

Nothing worse than the lifelong micro players of 2p2.
mack's is offline   Reply With Quote
Old 05-08-2015, 12:08 PM   #1280
watergun7
veteran
 
Join Date: Jun 2012
Posts: 2,773
Re: WCGRider, Dong Kim, Jason Les and Bjorn Li to play against a new HU bot

lol mack's wtf is your problem?
watergun7 is offline   Reply With Quote
Old 05-08-2015, 12:16 PM   #1281
KingOfSwings
centurion
 
Join Date: Mar 2015
Posts: 109
Re: WCGRider, Dong Kim, Jason Les and Bjorn Li to play against a new HU bot

Quote:
Originally Posted by watergun7 View Post
lol mack's wtf is your problem?
He is ...|... aka rewket aka max86 aka a lot of other banned account. He created this new account a few hours after the last one was banned so he could keep trash talking people. Sick life uh?

To answer your question: mental.
KingOfSwings is offline   Reply With Quote
Old 05-08-2015, 12:32 PM   #1282
Joshua
journeyman
 
Joshua's Avatar
 
Join Date: Sep 2002
Location: Monkey Tilt
Posts: 279
Re: WCGRider, Dong Kim, Jason Les and Bjorn Li to play against a new HU bot

I didn't see the streamed press conference but on the competition homepage the professor says it was a statistical tie:

But statistically, says Tuomas Sandholm, the CMU professor of computer science who led the Claudico's development, it was a tie, and the results of the experiment would seem to be inconclusive.
The "pot" won by the human poker pros—$732,713—sounds very impressive. But during the two week "Brains vs. A.I." challenge, $170 million was bet. The winnings amount to less than one-half of 1 percent of the total pot—a statistically tiny amount.
It's impossible to say that the humans clearly beat Claudico, according to Sandholm.


Link:
https://www.cs.cmu.edu/brains-vs-ai#May-8-2015
Joshua is offline   Reply With Quote
Old 05-08-2015, 12:44 PM   #1283
+rep_lol
banned
 
+rep_lol's Avatar
 
Join Date: Sep 2013
Posts: 14,904
Re: WCGRider, Dong Kim, Jason Les and Bjorn Li to play against a new HU bot

thats atrocious
+rep_lol is offline   Reply With Quote
Old 05-08-2015, 12:45 PM   #1284
restorativejustice
veteran
 
restorativejustice's Avatar
 
Join Date: Aug 2014
Posts: 2,753
Re: WCGRider, Dong Kim, Jason Les and Bjorn Li to play against a new HU bot

Quote:
Originally Posted by Joshua View Post
I didn't see the streamed press conference but on the competition homepage the professor says it was a statistical tie:

But statistically, says Tuomas Sandholm, the CMU professor of computer science who led the Claudico's development, it was a tie, and the results of the experiment would seem to be inconclusive.
The "pot" won by the human poker pros—$732,713—sounds very impressive. But during the two week "Brains vs. A.I." challenge, $170 million was bet. The winnings amount to less than one-half of 1 percent of the total pot—a statistically tiny amount.
It's impossible to say that the humans clearly beat Claudico, according to Sandholm.


Link:
https://www.cs.cmu.edu/brains-vs-ai#May-8-2015
The Prof is too proud to tell it like it is. I knew the word tie would be used, I just thought it would only be the case if the computer beat two humans instead of just one (and then that is only relevant if you ignore mirroring of hands).

I did not see the press conference, but the best way to explain it to layman would be the humans on average won over $9/hand each and every hand that they played against the computer over two weeks. Nice work if you can get it.
restorativejustice is offline   Reply With Quote
Old 05-08-2015, 12:48 PM   #1285
mcb08
grinder
 
mcb08's Avatar
 
Join Date: Nov 2008
Location: Not where u think I am
Posts: 434
Re: WCGRider, Dong Kim, Jason Les and Bjorn Li to play against a new HU bot

Quote:
Originally Posted by Joshua View Post
I didn't see the streamed press conference but on the competition homepage the professor says it was a statistical tie:

But statistically, says Tuomas Sandholm, the CMU professor of computer science who led the Claudico's development, it was a tie, and the results of the experiment would seem to be inconclusive.
The "pot" won by the human poker pros—$732,713—sounds very impressive. But during the two week "Brains vs. A.I." challenge, $170 million was bet. The winnings amount to less than one-half of 1 percent of the total pot—a statistically tiny amount.
It's impossible to say that the humans clearly beat Claudico, according to Sandholm.


Link:
https://www.cs.cmu.edu/brains-vs-ai#May-8-2015
LOLZ. What an embarrassment. The humans would need a winrate of over 21BB/100 to get to 1% of the total pot, which would still likely be deemed to be a "statistically tiny amount". I wonder if 50 BB/100 would be good enough?
mcb08 is offline   Reply With Quote
Old 05-08-2015, 12:53 PM   #1286
polarizeddeck
newbie
 
Join Date: Apr 2015
Posts: 38
Re: WCGRider, Dong Kim, Jason Les and Bjorn Li to play against a new HU bot

someone correct me if I'm wrong but how could this be a statistical tie?

suppose the standard deviation is 200bb/100. if the mirror matchup cuts this in half to 100b/100, after 800 sets of 100 hands the sd of the average is
100/sqrt(800) = 3.5

winning at 9bb/100 is over 2.5 sds from 0 so it seems pretty significant
polarizeddeck is offline   Reply With Quote
Old 05-08-2015, 01:03 PM   #1287
bip!
Slow Pony
 
bip!'s Avatar
 
Join Date: Oct 2012
Location: not on urban dictionary...
Posts: 13,669
Re: WCGRider, Dong Kim, Jason Les and Bjorn Li to play against a new HU bot

Quote:
Originally Posted by ArtyMcFly View Post
I suppose he forgot to mention it was also in sixth, seventh and eighth.

Academics gonna academic.

bip! is offline   Reply With Quote
Old 05-08-2015, 01:03 PM   #1288
RalphWaldoEmerson
Pooh-Bah
 
RalphWaldoEmerson's Avatar
 
Join Date: Sep 2014
Location: san francisco
Posts: 5,700
Re: WCGRider, Dong Kim, Jason Les and Bjorn Li to play against a new HU bot

Quote:
Originally Posted by mcb08 View Post
The prof is on the stream with WCG, trying to explain why Claudico folded A4o to Doug's AIPF shove (~9K). Doug is trying to explain that it's never right to fold A4 in that spot, and the prof is trying to say that Claudico made the right play, because Doug's playing of his hand (99) broke the NE. Hopefully Sam can explain it to him later.
oh man! did anyone record this by chance?
RalphWaldoEmerson is offline   Reply With Quote
Old 05-08-2015, 01:04 PM   #1289
watergun7
veteran
 
Join Date: Jun 2012
Posts: 2,773
Re: WCGRider, Dong Kim, Jason Les and Bjorn Li to play against a new HU bot

Yes idk if the prof is delusional or just putting up a show for pr. 9bb/100 is a pretty big edge especially considering the duplicate matches.
watergun7 is offline   Reply With Quote
Old 05-08-2015, 01:09 PM   #1290
e i pi
veteran
 
e i pi's Avatar
 
Join Date: Feb 2010
Location: ☃
Posts: 2,230
Re: WCGRider, Dong Kim, Jason Les and Bjorn Li to play against a new HU bot

Prof needs to keep the grant money flowing. I don't think they set all of this up expecting the bot to lose lol.
e i pi is offline   Reply With Quote
Old 05-08-2015, 01:12 PM   #1291
e i pi
veteran
 
e i pi's Avatar
 
Join Date: Feb 2010
Location: ☃
Posts: 2,230
Re: WCGRider, Dong Kim, Jason Les and Bjorn Li to play against a new HU bot

Also for future bot matches consider going through all their published papers. Many of them are about improving on some known botting method by reducing the exploitability in some way. But these methods are still exploitable so their papers are basically a well documented list of the bot's leaks.
e i pi is offline   Reply With Quote
Old 05-08-2015, 01:13 PM   #1292
bip!
Slow Pony
 
bip!'s Avatar
 
Join Date: Oct 2012
Location: not on urban dictionary...
Posts: 13,669
Re: WCGRider, Dong Kim, Jason Les and Bjorn Li to play against a new HU bot

They used the hand mirroring to get 200bb fixed starting chips... but don't want to use the hand mirroring when considering statistical significance. Lulz

The prof wants his work to go down next to deep blue as a milestone in AI.. ego is your answer. He can't have that milestone if the bot lost, so he just lies - nice. I get that the performance is pretty close to being a bot that wins, but it is not a bot that wins. Claudico won't be the one to be remembered. Maybe next year's will be.
bip! is offline   Reply With Quote
Old 05-08-2015, 01:15 PM   #1293
restorativejustice
veteran
 
restorativejustice's Avatar
 
Join Date: Aug 2014
Posts: 2,753
Re: WCGRider, Dong Kim, Jason Les and Bjorn Li to play against a new HU bot

Quote:
Originally Posted by e i pi View Post
Prof needs to keep the grant money flowing. I don't think they set all of this up expecting the bot to lose lol.
Bet you're right.

Seeing this competition and then listening to the Prof "interpret" the results reminds me how all the global warming climate models have PROVEN wrong on the high side when the predictions are compared to the reality of temperatures each and every year forward.

Academics have trouble admitting when they are wrong. I bet the Prof thinks that Claudico gave up $9/hand to the humans because of something the humans did wrong in not playing "properly."
restorativejustice is offline   Reply With Quote
Old 05-08-2015, 01:17 PM   #1294
Wasp
enthusiast
 
Join Date: Feb 2010
Posts: 70
Re: WCGRider, Dong Kim, Jason Les and Bjorn Li to play against a new HU bot

Okay the competition seems ended so maybe Bjorn Li will be okay if I comment it again.

That was exactly I tought and stated here before the competition started (after I read the rules):
- if the computer wins it wins and this will be a good advertising for a multimillion dollar business
- if the computer loses it will be declared as a tie so this will be a win which is a good advertising .... (you know the drill)

If the computer folds every hand it loses at 75BB/100. If it plays like a fish it loses about at 30-40bb/100. The computer declared as a good player though. Winning 5bb/100 over a regular is pretty good in long term and 80k mirrored hands is an okay sample.

Suggesting that this is a tie it is a lie. Plain and simple. We are playing poker we know what 10bb/100 means. It means anybody that able to reach this can play for living if the loser player is willing to play.

10bb/100 means the computer was crushed in regwar. Thats all, gg and better luck next time Claudico (and CMU).

So that wasn't a tie, and we know it. But telling lie is bad for us. If CMU doesn't admit the lose I think we can explain to the media what the winning rate means and how bad it looks if a head of the developer team lies about the result of an experiment on public.
Wasp is offline   Reply With Quote
Old 05-08-2015, 01:19 PM   #1295
bip!
Slow Pony
 
bip!'s Avatar
 
Join Date: Oct 2012
Location: not on urban dictionary...
Posts: 13,669
Re: WCGRider, Dong Kim, Jason Les and Bjorn Li to play against a new HU bot

Quote:
Originally Posted by e i pi View Post
Prof needs to keep the grant money flowing. I don't think they set all of this up expecting the bot to lose lol.

I gotta believe they thought the thing was unbeatable.

Tartarian stomped the bot competition - this stomped Tartarian... so he figured it was near perfect.

And I keep harping on it - but hand mirroring was likely set up to justify 200bb fixed strategy as much as it was for statistical reasons (props to them for going all the way to 200bb though).
bip! is offline   Reply With Quote
Old 05-08-2015, 01:23 PM   #1296
200zoomgrinder
journeyman
 
Join Date: Jan 2014
Posts: 366
Re: WCGRider, Dong Kim, Jason Les and Bjorn Li to play against a new HU bot

WOW at tie. I thought Sam was saying that they would accept 95% confidence. At std dev = 160bb/100 (quite high for HUNL I think) and winrate 9.56 humans are still 95% confidence to be winning.

Using equity chops usually reduces std dev by about 20% @ 100BB, maybe a bit less @ 200BB since you get AI pre river less. So maybe 10%. Mirror I would guess reduces (significantly) more than equity chops, but again maybe lets say 10%. So conservatively the adj. std dev is 129.6 This gives humans 98% confidence of being winners. Nice "tie".
200zoomgrinder is offline   Reply With Quote
Old 05-08-2015, 01:24 PM   #1297
bip!
Slow Pony
 
bip!'s Avatar
 
Join Date: Oct 2012
Location: not on urban dictionary...
Posts: 13,669
Re: WCGRider, Dong Kim, Jason Les and Bjorn Li to play against a new HU bot

It would be very easy for them to calculate stdev from the results
bip! is offline   Reply With Quote
Old 05-08-2015, 01:28 PM   #1298
what is check
journeyman
 
what is check's Avatar
 
Join Date: Jan 2011
Posts: 372
Re: WCGRider, Dong Kim, Jason Les and Bjorn Li to play against a new HU bot

wow. what an absolutely insane statement to make.
what is check is offline   Reply With Quote
Old 05-08-2015, 01:28 PM   #1299
bip!
Slow Pony
 
bip!'s Avatar
 
Join Date: Oct 2012
Location: not on urban dictionary...
Posts: 13,669
Re: WCGRider, Dong Kim, Jason Les and Bjorn Li to play against a new HU bot

bip! is offline   Reply With Quote
Old 05-08-2015, 01:38 PM   #1300
cheet
veteran
 
cheet's Avatar
 
Join Date: Jan 2006
Posts: 2,498
Re: WCGRider, Dong Kim, Jason Les and Bjorn Li to play against a new HU bot

http://pittsburgh.cbslocal.com/2015/...rivers-casino/

A longer segment will be on the CBS morning show tomorrow at (I believe) 9:30 am est
cheet is offline   Reply With Quote

Reply
      

Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off


Forum Jump


All times are GMT -4. The time now is 03:51 AM.


Powered by vBulletin®
Copyright ©2000 - 2021, Jelsoft Enterprises Ltd.
Copyright © 2008-2020, Two Plus Two Interactive