Pokersnowie question - Page 12 - Poker Theory

Looking at the Pokersnowie calling range in the SB vs. a Pot BTN open the only hand in it's range to call is KTs which it flats 23% of the time. This can't be correct it should flat a wider range to be harder to read or it should not have a flatting range.

If the pot is opened on the BTN for pot with TT and Pokersnowie calls in the SB if the flop comes T K K the best play is to give up unless we improve to quads without having to pay to get there.

Quote

01-14-2014 , 01:44 AM

#277

andyg2001

veteran

Join Date: Mar 2009 Posts: 2,905

Quote:

Originally Posted by andyg2001

Except I can't get quads so it should just be a fold.

Quote

01-14-2014 , 02:24 AM

#278

jglsd1

veteran

Join Date: May 2008 Posts: 2,445

Quote:

Originally Posted by ExaMeter

Oh boy.
Can everybody finally listen to what zachvac says and let this thread die already?

sent from phone

This :-(

Quote

01-14-2014 , 05:18 PM

#279

zachvac

Carpal \'Tunnel

Join Date: Apr 2008 Posts: 13,775

Quote:

Originally Posted by Pamik

i hope holdem die soon, and than some new game become popular, where no one knows what is optimal play, and there are no softwares or training videos. it would be epic survival of the fittest. some will get broke, some will get rich very quick, but it will happen. no game will last forever, but people will always gamble.

Why do you think nlhe dying will cause another game to get popular? Poker being unbeatable also doesn't mean fish will stop playing it. Roulette is unbeatable too yet plenty of people blow money on it every day.

Quote

01-14-2014 , 05:39 PM

#280

andyg2001

veteran

Join Date: Mar 2009 Posts: 2,905

Quote:

Originally Posted by zachvac

If there are a decent number of fish playing relative to the rake, then it's not going to be unbeatable.

Quote

01-14-2014 , 08:12 PM

#281

Pamik

centurion

Join Date: Feb 2013 Posts: 114

Quote:

Originally Posted by zachvac

Poker being unbeatable also doesn't mean fish will stop playing it. Roulette is unbeatable too yet plenty of people blow money on it every day.

yeah you right, fish doesn't care if game is solved or not lol. but lets assume that snowie is perfect GTO now, so every reg who is willing to study snowie at least 2hrs a day for one year will become pretty much unbeatable to even best players. they will not be able to defend their lobbies since there would be no weaker regs. baisically end of high stakes since even fish there are serious enough about poker, that they will start using snowie very soon.

as we can see snowie team is seriously determined to reach that near perfect equilibrium (perfect for human undarstanding of the game anyway)
they might do it very soon. my rough guess is about 3 years.

Quote

01-15-2014 , 03:07 AM

#282

knircky

veteran

Join Date: Jan 2010 Posts: 2,097

Well even if snowie finds GTO no human will ever be able to memorize it. So its not like ur can take GTO cheat sheet to the table like basic stategy for BJ.

However I do believe as well that the more poker is solved the harder the games will become.

However snowie is just one tool that pushes in this direction.

Quote

01-15-2014 , 03:54 AM

#283

Ellekappae

stranger

Join Date: Oct 2011 Posts: 6

Quote:

Originally Posted by Pamik

yeah you right, fish doesn't care if game is solved or not lol. but lets assume that snowie is perfect GTO now, so every reg who is willing to study snowie at least 2hrs a day for one year will become pretty much unbeatable to even best players. they will not be able to defend their lobbies since there would be no weaker regs. baisically end of high stakes since even fish there are serious enough about poker, that they will start using snowie very soon.

as we can see snowie team is seriously determined to reach that near perfect equilibrium (perfect for human undarstanding of the game anyway)
they might do it very soon. my rough guess is about 3 years.

Fish will never going to study and improve their game. The proof is that you can find tons of material to study, but doing it would require sacrifice, and so most of them just keep playing (and maybe improving a little by playing...slowly).

I guess playing GTO would be like terrifying for the variance, so only really serious player (mindset, bankroll) can afford a game like that.

Quote

01-15-2014 , 04:26 AM

#284

Babarberousse

old hand

Join Date: Jul 2013 Posts: 1,752

Quote:

Originally Posted by knircky

Well even if snowie finds GTO no human will ever be able to memorize it. So its not like ur can take GTO cheat sheet to the table like basic stategy for BJ.

If GTO was found (by snowie or any other program) it would be replicated by free programs very quickly.
Look at internet chess, even though there's no money to be made, there are a lot of people who use chess engines to cheat. It will be much much worse with poker when stronger AIs start spreading.

Quote

01-15-2014 , 08:21 AM

#285

jimgreer

stranger

Join Date: Mar 2013 Posts: 7

Quote:

Originally Posted by Babarberousse

I basically agree with this, but poker is a much harder AI problem than chess. For instance, here's a chess AI written in under 1200 characters: http://home.hccnet.nl/h.g.muller/max1.html. While it's probably not super strong, it's a hell of a lot better than a poker AI of that size would be.

Last edited by jimgreer; 01-15-2014 at 08:27 AM. Reason: shouldn't have said it was strong

Quote

01-15-2014 , 12:28 PM

#286

Babarberousse

old hand

Join Date: Jul 2013 Posts: 1,752

Quote:

Originally Posted by jimgreer

Yes, I discussed that in the chess vs poker thread. A poker AI is indeed a much bigger challenge.
However micromax is just a programming feat, I haven't played against it but its level is probably ridiculous. I would expect a poker program given a few instructions (TP = half psb, >TP = psb, no draw no pair = fold to bet,...) with a bit of strategy randomization to have a comparable poker level.

Quote

01-15-2014 , 05:50 PM

#287

mme

old hand

Join Date: May 2009 Posts: 1,668

poker AI in < 120 chars. not a very good one as i would be the first to admit. you may want to find a way to weight the possible actions more intelligently.

Code:

while True:
	button = random.choice((button_fold, button_call, button_bet))
	button.click()
	time.sleep(timeout)

Quote

01-15-2014 , 11:49 PM

#288

Babarberousse

old hand

Join Date: Jul 2013 Posts: 1,752

I think your program is fine. It should be break even at nl2 or vs hellmuth.

Quote

01-16-2014 , 02:49 AM

#289

mme

old hand

Join Date: May 2009 Posts: 1,668

the swarm intelligence of 2+2 could extend it to measure the GTO'ness of any agent presented. though not likely to be getting perfect within a reasonable time frame it may outwit the shout-o-cracy approach pretty fast.

/ducks

Quote

01-16-2014 , 05:22 AM

#290

mangler241

grinder

Join Date: Apr 2007 Posts: 534

Quote:

Originally Posted by RustyBrooks

Talking about the methodology is pointless. I have a methodology that is 100% perfect - brute force. If I search the solution space exhaustively, I can guarantee you I will find a solution. Eventually.

So methodology is not the issue. Results ARE the issue. A method that will someday produce a GTO solution is in no way interesting. Until it does actually get close.

Claiming to have a solution to HU NLHE, or even having something really close, is an extremely remarkable claim and should be treated as such. Claiming to being close to a solution for 6max NLHE is beyond that by many orders of magnitude. Also known as "lying"

I thought it was known as "marketing".

Quote

01-16-2014 , 07:30 PM

#291

fontaine

journeyman

Join Date: Apr 2008 Posts: 224

Quote:

Originally Posted by fontaine

I received an email from the Snowie team today, saying that they have updated the Snowie engine after reviewing my play - and they offered me a one year free subscription if I can still beat Snowie for >5bb/100 over 6k hands. I guess I will give it a go when I have time for it

I just completed the second challenge. Won 1482bb in 6017 hands for a win rate of 24.6 bb/100. To be fair i ran really hot the last 1k hands, the win rate after 5000 hands was only 7.5 bb/100. It was much harder this time, after Snowie learned how to combat my minbets in 3bet pots.

I played 66/64 with 5.38 AF. 100% open in utg, btn and sb positions, and 100% 3bet bb vs sb. I defended against most 3bets and 4bets when in LP.

Quote

01-17-2014 , 04:09 AM

#292

PokerRon247

Carpal \'Tunnel

Join Date: Jul 2008 Posts: 11,192

Question for the guys helping Snowie out by playing it and giving feedback: If Snowie was instead called UltimatePokerBot and they advertised it by saying "we're building the bot to beat all bots, the bot that will eventually kill off poker, come help us in our quest", would you still be as willing to help?

Quote

01-17-2014 , 10:24 AM

#293

Babarberousse

old hand

Join Date: Jul 2013 Posts: 1,752

Quote:

Originally Posted by PokerRon247

The theoretical challenge poker is for modern AIs is more interesting to me than the money I'll ever make playing poker.
I can understand this is different for you, but you are trying to delay the inevitable. However, fortunately for you, Snowie is still far from killing your poker games (unless you play nl2).

Quote

01-17-2014 , 10:40 AM

#294

PokerRon247

Carpal \'Tunnel

Join Date: Jul 2008 Posts: 11,192

Fwiw I don't really have a opinion on the question (except a purely selfish one of wanting the game to never get any harder, but I do understand all sides). It was more the fact that if someone came onto 2p2 and said "hey guys I'm building a bot, will you help me?" then I'm sure he wouldn't have anywhere near as good responses as the Snowie team will have had after creating a legitimate and flashy looking website and using their marketing skills to entice people in with "perfect poker" etc.

Quote

01-17-2014 , 10:47 AM

#295

Power On/Off

banned

Join Date: Nov 2013 Posts: 11

Quote:

Originally Posted by Babarberousse

However, fortunately for you, Snowie is still far from killing your poker games (unless you play nl2).

That's why the montenegro bots are playin up to zoom 1k, right?

Quote

01-17-2014 , 01:12 PM

#296

Babarberousse

old hand

Join Date: Jul 2013 Posts: 1,752

Quote:

Originally Posted by PokerRon247

I get your point. The term "bot" has a very bad connotation in poker, and probably for good reasons.

Quote:

That's why the montenegro bots are playin up to zoom 1k, right?

Sources ?

Quote

01-17-2014 , 01:40 PM

#297

ExaMeter

Pooh-Bah

Join Date: Jul 2006 Posts: 3,984

Check the snowie blog. They say that they put an add into a paper in montenegro and taught these guys poker from scratch with the help of snowie. Or so they claim.

sent from phone

Quote

01-17-2014 , 01:49 PM

#298

andyg2001

veteran

Join Date: Mar 2009 Posts: 2,905

Quote:

Originally Posted by Power On/Off

That's why the montenegro bots are playin up to zoom 1k, right?

There not bots just people who trained with Snowie before it was released.

Quote

01-18-2014 , 08:11 AM

#299

Babarberousse

old hand

Join Date: Jul 2013 Posts: 1,752

Quote:

Originally Posted by ExaMeter

Check the snowie blog. They say that they put an add into a paper in montenegro and taught these guys poker from scratch with the help of snowie. Or so they claim.

sent from phone

That's what I thought.

Quote

01-19-2014 , 01:43 PM

#300

mike1270

adept

Join Date: Mar 2005 Posts: 1,078

if they have a group of montenegro guys crushing so hard after using snowie to learn poker why not release graphs and player names? if the stories are true it would really help their sales?

Quote

Page 12 of 93

First

2 7 8 9 10 11 12 13 14 15 16 17 22 32 62

Last

Post Reply Subscribe

...

Page 12 of 93

First

2 7 8 9 10 11 12 13 14 15 16 17 22 32 62

Last