Two Plus Two Publishing LLC Two Plus Two Publishing LLC
 

Go Back   Two Plus Two Poker Forums > >

Notices

Poker Theory General poker theory

Reply
 
Thread Tools Display Modes
Old 11-12-2013, 07:48 PM   #76
ZenFish
journeyman
 
ZenFish's Avatar
 
Join Date: Dec 2012
Location: GTOpia
Posts: 300
Re: Pokersnowie question

Quote:
Originally Posted by mme View Post
collusion is ruled out.

i am too lazy to dig it out, but this was the gist of the reply to some weakish prove attempt on why GTO can not exist in multi player situations and thus sno**e could always be beat in theory (discussed in this subforum not long ago)
Derailing a bit, but to clarify what I meant:

My understanding of the problem is that two or more more players can sometimes increase their EV's by changing their strategies together, even if neither one could have achieved that by changing his strategy alone. This is called an "alliance".

An alliance is not collusion (as in information sharing), each player operates on his own, but can anticipate certain actions from his opponents, because it makes sense for them to do so. Like everybody checking to showdown when a short-stack is all-in at a tournament FT, because everybody gain when he is eliminated, and they maximize the chance of that happening by not forcing out each other before showdown.

You can always find a Nash equilibrium (where one player can not increase his EV by deviating alone), but the forming of alliances can lead to some players gaining by deviating from it.

So the question is: Is "GTO" well-defined for multiway pots? If it isn't, how can you claim that your algorithm can get there? And if it is, how can you approximate it well in a 6-max game, when even HU NL is nowhere near solved? Smart people with math degrees are encouraged to comment.

Last edited by ZenFish; 11-12-2013 at 08:01 PM.
ZenFish is offline   Reply With Quote
Old 11-12-2013, 07:58 PM   #77
knircky
veteran
 
Join Date: Jan 2010
Posts: 2,097
Re: Pokersnowie question

Quote:
Originally Posted by ZenFish View Post
Do you have a link?
So you think it is not GTO, but you have not even downloaded it? They have a website and you can download the program and it will show you the entire strategy. They made it completely open.

It would be nice if you guys instead of hating on it actually look into it, understand what it does so we could actually talk about it.
knircky is offline   Reply With Quote
Old 11-12-2013, 08:03 PM   #78
ZenFish
journeyman
 
ZenFish's Avatar
 
Join Date: Dec 2012
Location: GTOpia
Posts: 300
Re: Pokersnowie question

Quote:
Originally Posted by knircky View Post
So you think it is not GTO, but you have not even downloaded it? They have a website and you can download the program and it will show you the entire strategy. They made it completely open.

It would be nice if you guys instead of hating on it actually look into it, understand what it does so we could actually talk about it.
How does downloading the program and running hands through it teach me how it works? And why should it be my job to check that it's actually the GTO strategy? Not hating, but would like to see some science behind it, as opposed to pseudoscience. When somebody claims "X = Y", they should prove it, not leave it to the reader to disprove it.

So, if you know the program, please tell how we can get to the strategy contained therein, because I am curious (sceptical, but curious). How should I go about to verify that it's GTO? You seem to believe that it is. Why?

Last edited by ZenFish; 11-12-2013 at 08:10 PM.
ZenFish is offline   Reply With Quote
Old 11-12-2013, 08:09 PM   #79
RustyBrooks
Carpal \'Tunnel
 
RustyBrooks's Avatar
 
Join Date: Feb 2006
Location: Austin, TX
Posts: 23,201
Re: Pokersnowie question

I've downoaded the program. I ran some hands through it. I did not see any way for it to show me "the entire strategy"

Actually to be perfectly clear, I didn't even really download it. There is a "free trial" link and when you go there, it starts downloading the program on it's own.
RustyBrooks is offline   Reply With Quote
Old 11-12-2013, 08:12 PM   #80
Burnss
Pooh-Bah
 
Burnss's Avatar
 
Join Date: Feb 2012
Location: #TeamPingu
Posts: 4,546
Re: Pokersnowie question

RustyBrooks wins thread

Quote:
Originally Posted by knircky View Post
Feel free to prove that it is not GTO. Given the limitation they provided.
Mindblowing how people can think like this. Power of marketing i guess.
Burnss is offline   Reply With Quote
Old 11-12-2013, 09:46 PM   #81
knircky
veteran
 
Join Date: Jan 2010
Posts: 2,097
Re: Pokersnowie question

Quote:
Originally Posted by ZenFish View Post
How does downloading the program and running hands through it teach me how it works? And why should it be my job to check that it's actually the GTO strategy? Not hating, but would like to see some science behind it, as opposed to pseudoscience. When somebody claims "X = Y", they should prove it, not leave it to the reader to disprove it.

So, if you know the program, please tell how we can get to the strategy contained therein, because I am curious (sceptical, but curious). How should I go about to verify that it's GTO? You seem to believe that it is. Why?
You can use the scenario builder.

It provides you with a pre flop chart for its range against any possible situation, then post flop it shows all possible moves and %ages.

It's extremely complicated obviously, but the entire decision three is there and quite nicely presented.

They are not saying they are done by the way they are still calculating it and they have the bit size limitations of course.

But from a practical perspective this is the first time for me to actually see what GTO looks like. I have never seen anything where I can check what a GTO strategy looks like.

I was pretty quickly able to learn HU this way.

Last edited by knircky; 11-12-2013 at 09:54 PM.
knircky is offline   Reply With Quote
Old 11-12-2013, 09:48 PM   #82
knircky
veteran
 
Join Date: Jan 2010
Posts: 2,097
Re: Pokersnowie question

Quote:
Originally Posted by RustyBrooks View Post
I've downoaded the program. I ran some hands through it. I did not see any way for it to show me "the entire strategy"

Actually to be perfectly clear, I didn't even really download it. There is a "free trial" link and when you go there, it starts downloading the program on it's own.
Im sorry clicking buttons is so hard. Maybe try again ?

I'm sorry but you are all acting like crybabies when it's right inf root of you.
knircky is offline   Reply With Quote
Old 11-12-2013, 10:00 PM   #83
RustyBrooks
Carpal \'Tunnel
 
RustyBrooks's Avatar
 
Join Date: Feb 2006
Location: Austin, TX
Posts: 23,201
Re: Pokersnowie question

Quote:
But from a practical perspective this is the first time for me to actually see what GTO looks like. I have never seen anything where I can check what a GTO strategy looks like.
Why do you keep saying stuff like this?

I'll look at the scenario builder.
RustyBrooks is offline   Reply With Quote
Old 11-12-2013, 11:40 PM   #84
knircky
veteran
 
Join Date: Jan 2010
Posts: 2,097
Re: Pokersnowie question

Quote:
Originally Posted by RustyBrooks View Post

I'll look at the scenario builder.

I'll be interesting to hear where u think this thing is flawed. I have not been able to find much. I remeber there was one spot where my coach clearly disagreed with snowie when it reraised a TP on paired board.

I did not get the preflop-limp HU range and thought it must be exploitable, hence I opened the thread about it. But know I think i get actually why its balanced (enough).

What confuses me about snowie is that is often recommends less EV moves over higher EV moves.
knircky is offline   Reply With Quote
Old 11-13-2013, 02:22 AM   #85
SchDonk
old hand
 
SchDonk's Avatar
 
Join Date: Jul 2013
Location: metadating
Posts: 1,610
Re: Pokersnowie question

Quote:
Originally Posted by knircky View Post
What confuses me about snowie is that is often recommends less EV moves over higher EV moves.
Doh, because it's not exploitive?
Jesus ****ing christ... is this really so hard to understand?
SchDonk is offline   Reply With Quote
Old 11-13-2013, 07:16 AM   #86
Burnss
Pooh-Bah
 
Burnss's Avatar
 
Join Date: Feb 2012
Location: #TeamPingu
Posts: 4,546
Re: Pokersnowie question

^^

Quote:
Originally Posted by knircky View Post
There are a bunch of folks here talking about GTO that don't even understand what it is.
lol'd
Burnss is offline   Reply With Quote
Old 11-13-2013, 07:56 AM   #87
ZenFish
journeyman
 
ZenFish's Avatar
 
Join Date: Dec 2012
Location: GTOpia
Posts: 300
Re: Pokersnowie question

Quote:
Originally Posted by knircky View Post
I'll be interesting to hear where u think this thing is flawed. I have not been able to find much.
We have no way of knowing whether or not it's flawed, because 1) nobody knows what the GTO strategy looks like, and 2) the program has not been tested to measure how far away from GTO is is (which can be done, even if we don't know the GTO strategy).

For starters, they could put PokerSnowie into the Annual Poker Competition to see how it performs against other bots. They probably haven't done so, because they don't think they're doing any better than other research groups, which are very far from the GTO solution to NLHE. And this would hurt their sales, especially if they were badly beaten.

Secondly, they could test PokerSnowie with algorithms like the one discussed here: https://forumserver.twoplustwo.com/sh...5&postcount=62 to measure how far from the true GTO strategy its approximation is.

Believing PokerSnowie has a good approximation to GTO just because you see some percentages and their marketing teting tells you it's good percentages is naive.
ZenFish is offline   Reply With Quote
Old 11-13-2013, 08:24 AM   #88
mme
old hand
 
mme's Avatar
 
Join Date: May 2009
Location: true anti-CFR-HUDs
Posts: 1,668
Re: Pokersnowie question

Quote:
Originally Posted by ZenFish View Post
For starters, they could put PokerSnowie into the Annual Poker Competition to see how it performs against other bots. They probably haven't done so, because they don't think they're doing any better than other research groups, which are very far from the GTO solution to NLHE. And this would hurt their sales, especially if they were badly beaten.
from what i take from the podcast, they have a limitation in the scenario builder on how many hands you can analyze (time based as i would guess). with a bit of help from the community it shouldn't be too hard to bypass this and pit it against some known bot.

edit:
thinking again ..because this just dropped into my head while typing. isn't this the way to bypass any protection mechanisms: dedicate some of your snowie time to a pool of players (virtual or not)? a bit of load balancing here and there and you get something approaching full coverage in return.

Last edited by mme; 11-13-2013 at 08:36 AM.
mme is offline   Reply With Quote
Old 11-13-2013, 11:08 AM   #89
ExaMeter
Pooh-Bah
 
ExaMeter's Avatar
 
Join Date: Jul 2006
Posts: 3,986
Quote:
Originally Posted by SchDonk View Post
Doh, because it's not exploitive?
Jesus ****ing christ... is this really so hard to understand?
Im not going to double check this in my copy of mop but I'm very sure that chen and ankenman showed that there are no loss leaders in a gto solution of poker just how there are no loss leaders in a gto solution to any game.
Never playing a dominated strategy, that means playing one with lower ev than the alternative strategy, is simply an inherent feature of a gto solution.

sent from phone, lol auto correct
ExaMeter is offline   Reply With Quote
Old 11-13-2013, 03:55 PM   #90
plexiq
old hand
 
Join Date: Apr 2007
Location: Vienna
Posts: 1,639
Re: Pokersnowie question

^ You are right that a Nash strategy never includes line choices with a lower EV than the alternatives when calculated against the other Nash half-strategy.

("Dominated" has an even stronger meaning though. If a choice is dominated, it means that it has lower EV than the alternatives against every possible opponents strategy.)

Quote:
Originally Posted by knircky View Post
What confuses me about snowie is that is often recommends less EV moves over higher EV moves.
I don't know what exactly snowie displays as EV, maybe someone can clear this up. Is it the EV when other players are assumed to play with the snowie solution? In that case it should only ever play the highest EV option, and any exception to this indicates that snowies strategy is exploitable.
plexiq is offline   Reply With Quote
Old 11-13-2013, 05:57 PM   #91
Hans Gusen
centurion
 
Join Date: Nov 2013
Posts: 131
Re: Pokersnowie question

Quote:
Originally Posted by ZenFish View Post
Secondly, they could test PokerSnowie with algorithms like the one discussed here: https://forumserver.twoplustwo.com/sh...5&postcount=62 to measure how far from the true GTO strategy its approximation is.
That algorithm only works for limit afaik.
Hans Gusen is offline   Reply With Quote
Old 11-13-2013, 06:51 PM   #92
Hans Gusen
centurion
 
Join Date: Nov 2013
Posts: 131
Re: Pokersnowie question

The pokersnowie guys actually admit at some point that pokersnowie is not perfect. But that does not stop them from promoting their product with phrases like "perfect No-Limit Hold'em gameplay" as often as possible.

Last edited by Hans Gusen; 11-13-2013 at 06:58 PM.
Hans Gusen is offline   Reply With Quote
Old 11-15-2013, 01:38 AM   #93
knircky
veteran
 
Join Date: Jan 2010
Posts: 2,097
Re: Pokersnowie question

Quote:
Originally Posted by plexiq View Post


I don't know what exactly snowie displays as EV, maybe someone can clear this up. Is it the EV when other players are assumed to play with the snowie solution? In that case it should only ever play the highest EV option, and any exception to this indicates that snowies strategy is exploitable.
Yea I don't know either. They display EV in bb. And basically always take the choice of highest EV. If two options are the same they balance by playing both at certain %ages mostly.

But obv. it is focused on ranges and being able to exploit. It is very clear thats the purpose of snowie (and GTO of course).

However what happens is that snowie displays your bet size and 1/2 pot, pot and 2pot (or all in) and shows you the EV for every move.

It also recommends a bet size (which means it puts this hand your looking at into a certain range vs that of the other bet size)

Usually the bet size with the highest EV is the optimal bet size according to snowie, but sometimes it is not. That confuses me.

the makes also point out that snowie has some limitations which make it play wrong. they also say its not perfect and are using this method of finding GTO since they can't solve it scientifically. But this is the closest they can get to GTO.
knircky is offline   Reply With Quote
Old 11-15-2013, 02:31 AM   #94
tultfill
veteran
 
tultfill's Avatar
 
Join Date: Jan 2009
Location: internet
Posts: 2,928
Re: Pokersnowie question

Quote:
Originally Posted by knircky View Post
the makes also point out that snowie has some limitations which make it play wrong. they also say its not perfect
No, they are saying it is perfect. When you read interviews and other things they admit it's not perfect. They market it as perfect.

Quote:
and are using this method of finding GTO since they can't solve it scientifically. But this is the closest they can get to GTO.
What do we really know about them or their motives for trying to sell it now? Maybe they reached a conclusion that the program wouldn't get better playing itself over and over again and they realized they wouldn't be able to solve it better so they just try to sell it. They did put down a lot of resources into it, they need a payoff. Meanwhile the program is clumsy and far away from being a 6max solve, possibly so far off people will lose a lot of money trying to emulate the style, no matter what some "NL500 ZOOM beast" from albania or whatever have been paid to say.
tultfill is offline   Reply With Quote
Old 11-15-2013, 02:42 AM   #95
tultfill
veteran
 
tultfill's Avatar
 
Join Date: Jan 2009
Location: internet
Posts: 2,928
Re: Pokersnowie question

I don't mean to be condescending, I'm just too skeptical to trust companies like this. They have a lot to gain and until shown otherwise, nothing to lose.
tultfill is offline   Reply With Quote
Old 11-15-2013, 06:39 AM   #96
knircky
veteran
 
Join Date: Jan 2010
Posts: 2,097
Re: Pokersnowie question

Quote:
PokerSnowies's weaknesses

Is PokerSnowie Perfect?

To create a system that plays a certain game perfectly, the game has to be solved. Solving a game means creating a huge database with the correct action for any given game state. This database must contain every possible state of the game. For no-limit Hold'em, a game state is the exact situation, the given cards, the actions that happened in the game, the chip amounts at the table, the exact bet sizes etc. Without doing any calculations it is obvious that a game that complex cannot be solved with the current computer power and probably not in the next few hundred years.

PokerSnowie uses neural networks and estimates the right action for a given poker situation. Neural networks are able to interpolate well, so if a network has trained for two situations and is exposed to an unknown third one, it usually comes up with a good result, under the condition that the new situation is similar to the ones used for training.
This is the first paragraph of the weaknesses of snowie page.

They have dedicated an entire site to the topic of what makes it imperfect. Yet you accuse them of claiming it is perfect.
knircky is offline   Reply With Quote
Old 11-15-2013, 06:41 AM   #97
knircky
veteran
 
Join Date: Jan 2010
Posts: 2,097
Re: Pokersnowie question

I still have not seen anything useful in here what makes bad or exposed any weakness of it.
knircky is offline   Reply With Quote
Old 11-15-2013, 06:56 AM   #98
Polar Beard
old hand
 
Polar Beard's Avatar
 
Join Date: Oct 2010
Location: In the cesspit, ldo!
Posts: 1,235
Re: Pokersnowie question

Quote:
Originally Posted by AsUs91 View Post
why?
I am not familiar with pokersnowie, but after reading this thread, I elected to dowload their 10 days free trial. I was very very surprised by some of the advice it gave me. I am not well versed in GTO, but is this really supposed to be a fold?


Last edited by Polar Beard; 11-15-2013 at 07:00 AM. Reason: BTW, I found that me flatting here was probably wrong, yet the software likes my cold call
Polar Beard is offline   Reply With Quote
Old 11-15-2013, 07:00 AM   #99
SchDonk
old hand
 
SchDonk's Avatar
 
Join Date: Jul 2013
Location: metadating
Posts: 1,610
Re: Pokersnowie question

lol

ps: And htf does PokerSnowie know the holecards of the other guy? Only if it went to showdown obv. But it's misleading for the decision-making process.
I guess you were just lucky there that villain didn't have TT/99/22/KK+ which would be the huge majority of his cbetting range.

[] poker is solved

Last edited by SchDonk; 11-15-2013 at 07:07 AM.
SchDonk is offline   Reply With Quote
Old 11-15-2013, 07:22 AM   #100
Polar Beard
old hand
 
Polar Beard's Avatar
 
Join Date: Oct 2010
Location: In the cesspit, ldo!
Posts: 1,235
Re: Pokersnowie question

Quote:
Originally Posted by SchDonk View Post
lol

ps: And htf does PokerSnowie know the holecards of the other guy? Only if it went to showdown obv. But it's misleading for the decision-making process.
I guess you were just lucky there that villain didn't have TT/99/22/KK+ which would be the huge majority of his cbetting range.
Well, if villain before me opens 11% of his hands in this position (QJ+, 66+), it represents 150 combos. The range you put villain on represents 30 combos. On this flop, villain should cbet almost all his range. So I beleive that I am ahead 4/5 times. I just doubt that snap folding to a half pot cbet from the player with the initiative is optimal, but because I am not familiar with the GTO concept, I am asking

Last edited by Polar Beard; 11-15-2013 at 07:49 AM. Reason: Ho and yah, hand went to showdown, shortsatcker shoved and c-better snap folded :)
Polar Beard 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 04:30 PM.


Powered by vBulletin®
Copyright ©2000 - 2017, Jelsoft Enterprises Ltd.
Copyright © 2008-2017, Two Plus Two Interactive
 
 
Poker Players - Streaming Live Online