Brains vs. AI poker rematch coming to Rivers Casino - Page 18 - Poker News

Two Plus Two Forums Poker News & Discussion News, Views, and Gossip

Brains vs. AI poker rematch coming to Rivers Casino

Post Reply Subscribe

...

Page 18 of 53

First

8 13 14 15 16 17 18 19 20 21 22 23 28 38

Last

Page 18 of 53

First

8 13 14 15 16 17 18 19 20 21 22 23 28 38

Last

01-16-2017 , 08:20 PM

#426

samooth

veteran

Join Date: May 2009 Posts: 3,350

raise 250 reraise 875 call
check check
check bet 2450 (1.4xPOT) call
check jam for 2.5xPOT (fold)

mirror hand:

raise 274 reraise 1918 (fold)

Last edited by samooth; 01-16-2017 at 08:35 PM.

Quote

01-16-2017 , 08:22 PM

#427

GoBackToGo

enthusiast

Join Date: Aug 2010 Posts: 82

did the programmers say anything about what they are doing to the bot at night (lol)?
tbh, if they looked at data and found that limping was not making the bot enough money and then stopped the bot from limping, that would kind of defeat the purpose of this competition?

Quote

01-16-2017 , 08:38 PM

#428

ezdonkey

veteran

Join Date: Jun 2006 Posts: 2,985

Quote:

Originally Posted by samooth

97 vs 88 went super well for the humans ^^

raise 274 reraise 1096 call
bet 584 call
bet 3288 call
check bet 12330 call

cheet: "these bots have no respect"

Bot instachecked river.

Has anyone paid attention to any timing patterns? I was thinking if bot acts instantly when it has a trivial decision?

Quote

01-16-2017 , 08:38 PM

#429

.Alex.

Pooh-Bah

Join Date: Oct 2004 Posts: 5,672

The programmers have been super sketchy for both of these challenges. Wouldn't trust them one bit.

Quote

01-16-2017 , 08:41 PM

#430

THAY3R

Carpal \'Tunnel

Join Date: Jun 2004 Posts: 36,550

I believe every turn the bot calculates river decisions as well which is why the turns take so long and the river is instant

Quote

01-16-2017 , 08:55 PM

#431

ezdonkey

veteran

Join Date: Jun 2006 Posts: 2,985

Hand s2jason1 #205 or something, T63-Jcc-Ao runout. cheet checks down A8o and bets river ~1/2, bot something like 5,2k.
Seemed like a spot where bot might want to polarize down to KQ for value and Q3(offsuit, don't remember if suited turned bdfd) and K3(same criteria) for bluffs. (Or maybe some AT as well, and I'm guessing it leads 66/33 on turn some amount?)

Are the hand histories available from somewhere? I would like to know what was the exact number ratio how many bluffs bot was allowed to have with that sizing.

EDIT: Saw from Dong's stream bot had K3o. Bot wins vs cheet with 5,2k - Dong x/r only to 870 and bot calls with A8o.

Quote

01-16-2017 , 08:57 PM

#432

ezdonkey

veteran

Join Date: Jun 2006 Posts: 2,985

Quote:

Originally Posted by THAY3R

I believe every turn the bot calculates river decisions as well which is why the turns take so long and the river is instant

Instant OOP? I've seen bot do some serious river tanking, but idk if it was only IP and needed to know villain's action before wasting any computing time.

Quote

01-16-2017 , 08:58 PM

#433

samooth

veteran

Join Date: May 2009 Posts: 3,350

raise 274 reraise 1370 call
check bet 1013 call
check bet 1763 call
check bet 5223 (63%pot) fold

mirror hand:

raise 250 call
check bet 250 call
check bet 500 call
check bet 2000 (fold) (dodge)

Quote

01-16-2017 , 09:18 PM

#434

samooth

veteran

Join Date: May 2009 Posts: 3,350

raise 250 3bet 1125 4bet 3375 5bet 7560 fold

raise 205 reraise 875 call
check check
bet 1522 call
check bet 4794 (POT) call

Quote

01-16-2017 , 09:21 PM

#435

LOLCh1pPorn

Pooh-Bah

Join Date: Nov 2012 Posts: 4,580

Thanks a lot for the hands! This is really interesting

Quote

01-16-2017 , 09:23 PM

#436

samooth

veteran

Join Date: May 2009 Posts: 3,350

raise 250 reraise 1125 call
check check
check bet 1125 call
check marmalade call

mirror hand:

raise 350 reraise 1050 call
check bet 777 call
check bet 2265 call
check bet 14322 (leave 1586 behind) fold

Last edited by samooth; 01-16-2017 at 09:47 PM. Reason: spelling; marmalade = jam = all in

Quote

01-16-2017 , 09:27 PM

#437

mrno1324

veteran

Join Date: Feb 2016 Posts: 3,055

wtf is marmelade

Quote

01-16-2017 , 09:29 PM

#438

samooth

veteran

Join Date: May 2009 Posts: 3,350

raise 250 3bet 875 4bet 2625 call
check check
bet 4567 call
jam for 13k call (one for the good guys)

mirror hand:

raise 274 reraise 822 call
check check
bet 1233 call
bet 3082 (fold) (turn volume down)

Last edited by samooth; 01-16-2017 at 09:52 PM.

Quote

01-16-2017 , 09:37 PM

#439

prahsk87

grinder

Join Date: Mar 2009 Posts: 524

Quote:

Originally Posted by Cooozy

You're calling into question the ethics of the players without any reasonable suspicion/proof, which foolish and out of line.

Certain plays are -ev in a vacuum, but can be +ev in the big picture. You could make a -ev play to gain info that is more valuable in the longrun than the ev you lost to get it. The A2o hand could easily be an example of that.

If they were going to play bad on purpose, they would make it more subtle, not draw attention to it by making plays any random dude on a forum could easily demonstrate are bad.

But I'm curious, why would 3betting A2 to 34bb be better than 11-12bb? You lose way more if you get 4bet. Although I could see a benefit being that there's a better chance you can get your monsters in pre for 200bb since a 4bet from opponents will often be shoves instead of around 30bb against 11-12bb 3bet.

Quote

01-16-2017 , 09:45 PM

#440

LittleGoliath

Pooh-Bah

Join Date: Nov 2011 Posts: 3,921

Quote:

Originally Posted by prahsk87

They pbb want to see their reaction to bigger sizes (eg if it's overfolding/defending same etc)

Quote

01-16-2017 , 09:48 PM

#441

ezdonkey

veteran

Join Date: Jun 2006 Posts: 2,985

Quote:

Originally Posted by samooth

raise 250 3bet 875 4bet 2625 call
check check
bet 4567 call
jam for 13k call (one for the good guys)

Stnd cooler with top pair.

Sometimes these hands are so baffling. Bot calls here and is still up on humans. Given the graphics that hand could've been a screenshot from 2002 where some drooler calls TP $50/$100 real maney.

Quote

01-16-2017 , 09:50 PM

#442

WCGRider

Carpal \'Tunnel

Join Date: Apr 2007 Posts: 7,028

Quote:

Originally Posted by ezdonkey

The first couple of days of this challenge were total carnage. Humans dont have a hud or stats, they are seriously just in there playing blind. Pretty big disadvantage out of the gate.

As the days go on the humans have more edge because they can analyze sample.

Quote

01-16-2017 , 09:57 PM

#443

Zombiefan

adept

Join Date: Sep 2013 Posts: 1,175

Is the bot playing much different from the last match?

Thanks to all the ppl posting hh.

Quote

01-16-2017 , 10:06 PM

#444

samooth

veteran

Join Date: May 2009 Posts: 3,350

raise 250 reraise 875 call
bet 1085 call
bet 3410 call
bet 3973 call

Quote

01-16-2017 , 10:07 PM

#445

basheesh

stranger

Join Date: Mar 2016 Posts: 11

Quote:

Originally Posted by WCGRider

How is it a disadvantage though? It's not like the bot has their stats either. If anything, I'd say getting all the hand histories at the end of the day is an advantage to the humans, compared to what playing the bot "in the wild" would be like, where they wouldn't have such complete information.

Quote

01-16-2017 , 10:10 PM

#446

WCGRider

Carpal \'Tunnel

Join Date: Apr 2007 Posts: 7,028

Quote:

Originally Posted by basheesh

Because humans use huds and analyze stats. The bot is trying to play a GTO type of strategy. If humans dont have huds or stats (as they would always have playing online always) then they cant quantify the computers strategy and frequencies.

Quote

01-16-2017 , 10:19 PM

#447

samooth

veteran

Join Date: May 2009 Posts: 3,350

raise 250 call
bet 310 raise 1150 3bet 1990 4bet 5350 (fold)

i couldn't catch the mirror hand, but it went cheet opens with pocket 55 bot calls, flop check bet 345 and bot xraises huge to like 2.8k, cheet 3bet clicks it back, bot folds #clicking wars

Quote

01-16-2017 , 10:24 PM

#448

basheesh

stranger

Join Date: Mar 2016 Posts: 11

Quote:

Originally Posted by WCGRider

The current setup is closer to what a live poker setup would be like though. Having a HUD is already a form of computer-augmented play. I'm not necessarily against the idea of the humans having a HUD, but if the goal is to see whether current AIs are better than humans, I would say that no HUD is closer to actually testing that.

Quote

01-16-2017 , 10:37 PM

#449

Clayton

Carpal \'Tunnel

Join Date: Oct 2004 Posts: 32,228

Jason and Dong combine for a 70k win, Jimmy and Dan take the evening off due to Dan having an illness

Brains now at -50k

Quote

01-16-2017 , 10:47 PM

#450

Rich Checkmaker

banned

Join Date: Nov 2014 Posts: 2,441

Quote:

Originally Posted by WCGRider

not all humans

and you can certainly feel and approximate those stats if you cant necessarily quantify it.

even if they had a hud the first couple of days theyd still be playing blind right? how many hands does it take for your hud stats to matter? and how many hands without a HUD before your feel kicks in?

Quote

Page 18 of 53

First

8 13 14 15 16 17 18 19 20 21 22 23 28 38

Last

Post Reply Subscribe

...

Page 18 of 53

First

8 13 14 15 16 17 18 19 20 21 22 23 28 38

Last