Poker Bots - Quality information

This is not just another random 'omg bots' thread. Here is a very detailed blog about poker bots and information that can be used to make online poker more fair for the players.

http://www.mattmazur.com/category/poker-bot/

This is a nice comment thread about the blog:

http://news.ycombinator.com/item?id=657632

I found it quite an interesting read and has a lot of pertinent information. It is up-to-date as of June 15 2009 which is also nice (not some random article from years ago that is useless).

That's really interesting stuff - I just read all of his blog posts. I think the biggest worry about bots is that their existence puts off fish rather than anything else.

it really was a good read

the most interesting thing is that the bot can *profit*. You'd think it'd be really exploitable, but maybe the people there are just that bad

On June 16 2009 04:58 Oly wrote: I think the biggest worry about bots is that their existence puts off fish rather than anything else.

Agreed, it fits the 'online poker is rigged' belief perfectly. Hopefully something like this never makes it into mainstream news.

the weird part is that the bot played HU sngs
why wouldn't he just make it play FR ?

HU is a more perfectly informed game and can make more perfect decisions i guess

Its easier to figure out 1 guy rather than a whole table of guys. There aren't as many variables when playing 1 person, and the complexity and interactions go crazy the more people there are.

I see there's another Paul Graham fan on LP.

That guy is lacking business intelligence. If you read his emails it's clear he needs the attention and wants people to say "you're so smart".

He should have told pokerstars "My bot made 500$ in september 2008. Hire me as a consultant and I'll provide you with my algorithms and analysis"

On June 16 2009 09:28 royalsu wrote: That guy is lacking business intelligence. If you read his emails it's clear he needs the attention and wants people to say "you're so smart". He should have told pokerstars "My bot made 500$ in september 2008. Hire me as a consultant and I'll provide you with my algorithms and analysis"

I don't think that's the case, and I can't tell if you're being serious or jesting when you say "he should have told pokerstars..." but it earns a pretty big fail either way imo

anyway I spent an hour or so reading his poker bot blog and some of his other blogs and it was pretty interesting, actually (lol if you had AOL back in the day, with progz and appz, his Meridian blog entry was fairly nostalgic and enjoyable)

I am being serious night. He has a legitimate asset that he's giving away for free. Instead of waiting for a reply from them (he mentioned it was a week or so since the last email), he proceeded to give away free information to them. He already got his funds seized and I don't think pokerstars can even (or wants to) prosecute him in court. He now has the upper hand imo.

Ex-thieves, ex-hackers have started lucrative businesses consulting from the very people they steal from.

Basically I think what happened is he spent a shitload of hours on this project and didn't want it to go to waste. He wanted recognition so he emailed stars, then blogged about it so that everyone knew what he was doing.

But if he really wanted to help, he never should have blogged about it so that other cheating bot creators would know the strengths and weaknesses of his bot. He would have quietly marketed it to stars, partypoker, fulltilt, etc..

wow, what a good read. I wonder how much does a poker bot affect the whole community.

lol... if you can't beat a bot, quit poker

On June 16 2009 10:24 royalsu wrote: Basically I think what happened is he spent a shitload of hours on this project and didn't want it to go to waste. He wanted recognition so he emailed stars, then blogged about it so that everyone knew what he was doing. But if he really wanted to help, he never should have blogged about it so that other cheating bot creators would know the strengths and weaknesses of his bot. He would have quietly marketed it to stars, partypoker, fulltilt, etc..

On another note.. sounds like pokerstars have the right person in charge of bot detection.

read it all, seems to go heavily inline with some of the poker sites out there. cake network for example, there almost was a few people who played hours on end without chat making the same standard robot like plays

On June 16 2009 07:07 Zalfor wrote: HU is a more perfectly informed game and can make more perfect decisions i guess

Well i dont think so, hu is more about adaptability/exploit weakness and i have no doubt players like Fayth's caliber would totally destroy any bots hu as we speak . Bots would probably excell in sngs where a solid push/fold strategy is primordial, not to speak about perfect ICM calculation on the spot and massive multi-tabling. Poker is not like chess and i doubt it can be solved in the next 50 years. And even then there will still be live poker.

That was a good read, thanks for linking it. I can very much identify with his motives/drive for this type of thing, so it was especially interesting to me.

I play PLO... any bot in PLO would get completely ripped, HU,SNG,FR or no matter what.

Low stakes SNG's and LHE HU is the best for bots. Also i suspect a game like FR razz a bot can play pretty close to perfection.

I totally agree with everything royalsu says.

agree w/ edzwoo

i agree with dogmeat

On June 16 2009 11:59 Sennpu wrote: lol... if you can't beat a bot, quit poker

On June 16 2009 12:33 Achoo wrote: Show nested quote + On June 16 2009 07:07 Zalfor wrote: HU is a more perfectly informed game and can make more perfect decisions i guess Well i dont think so, hu is more about adaptability/exploit weakness and i have no doubt players like Fayth's caliber would totally destroy any bots hu as we speak . Bots would probably excell in sngs where a solid push/fold strategy is primordial, not to speak about perfect ICM calculation on the spot and massive multi-tabling. Poker is not like chess and i doubt it can be solved in the next 50 years. And even then there will still be live poker.

someone who knows better can correct me if i'm wrong, but i'm pretty sure HU is more amenable to game theory analysis than other formats, even though it might seem a little counterintuitive. most people (correctly) put much more focus on exploiting their opponents' tendancies rather than trying to make themselves unexploitable, so they approach HU as a very feel-based game, but the point of the bot's approach is to play unexploitably and wait for its opponents to make mistakes.

On June 16 2009 19:44 NeillyJQ wrote: Show nested quote + On June 16 2009 11:59 Sennpu wrote: lol... if you can't beat a bot, quit poker

also ignore people who say "if you can't beat a bot, quit poker" because they really have no idea what the fuck they're talking about.

the polaris bot has fared pretty well against some of the top LHE HU specialists in the world. since LHE is much simpler on the game theory tree than NLHE, it might be true that there's no bots that could currently beat a good NLHU player, but it's just a matter of time before they figure out how to balance all the variables in NLHE to make a GTO-bot for that game.

On June 19 2009 20:52 bigredhoss wrote: Show nested quote + On June 16 2009 19:44 NeillyJQ wrote:   On June 16 2009 11:59 Sennpu wrote: lol... if you can't beat a bot, quit poker also ignore people who say "if you can't beat a bot, quit poker" because they really have no idea what the fuck they're talking about. the polaris bot has fared pretty well against some of the top LHE HU specialists in the world. since LHE is much simpler on the game theory tree than NLHE, it might be true that there's no bots that could currently beat a good NLHU player, but it's just a matter of time before they figure out how to balance all the variables in NLHE to make a GTO-bot for that game.

No, SUPERcomputers still have trouble fighting with grand masters in chess, a game of complete information, making a bot to defeat Poker is much more harder.

On June 20 2009 15:33 Baal wrote: Show nested quote + On June 19 2009 20:52 bigredhoss wrote:   On June 16 2009 19:44 NeillyJQ wrote:   On June 16 2009 11:59 Sennpu wrote: lol... if you can't beat a bot, quit poker also ignore people who say "if you can't beat a bot, quit poker" because they really have no idea what the fuck they're talking about. the polaris bot has fared pretty well against some of the top LHE HU specialists in the world. since LHE is much simpler on the game theory tree than NLHE, it might be true that there's no bots that could currently beat a good NLHU player, but it's just a matter of time before they figure out how to balance all the variables in NLHE to make a GTO-bot for that game. No, SUPERcomputers still have trouble fighting with grand masters in chess, a game of complete information, making a bot to defeat Poker is much more harder.

Not sure what you're saying no to. It's a fact that HU LHE has pretty much been solved by the alberta computer poker research group, and if the polaris bot was a person it would be one of the top 10 or so HULHE players in the world.

wtf r u on about. chess programs crush chess champions

http://en.wikipedia.org/wiki/Blunder_(chess)#Deep_Fritz_vs._Vladimir_Kramnik

Baal nowadays you can run chess programs on a PC that will beat most any grandmaster.

Also, complete information doesn't say everything. Simple incomplete information games are being destroyed by bots (LHE?). Go, a complete information game, has computer bots struggling against the average amateur club player. Luck also complicates things, but nowadays the best human backgammon players use backgammon bots as coaching tools, deriding themselves when they pick a different option than the computer. There's a lot of things that can make a game hard or easy for computers.

At the moment, NLHE is quite hard for computers, but it won't be so forever.

computation time (and therefore PC-Hardware) is irrelevant for current NLHE Bots.
They work with pretty simple Rulesystem like "if hand is better then top pair and players_at_table <7 then bet/raise" or "if holecards are 27o and stage is preflop then check/fold"
an amiga500 could solve this in realtime

If you would try to "solve" the game correctly you would have to solve a pretty complex euqitationsystem with tons of variables. This actually IS curently solveable by strong computers - but only for a headsup limit game. A fullring nolimit game simply has too much variables for current computers.

This is a wonderful example of how the human brain excels even the strongest computers of our time since most of us find it quit simple to analyse most NLHE poker situations in a split second

@mow

Only a grand achievement if you believe computers are already better than humans in most things.

After all, nowadays pocket calculators are faster than the best mental calculators.

On June 20 2009 19:17 genjix wrote: wtf r u on about. chess programs crush chess champions http://en.wikipedia.org/wiki/Blunder_(chess)#Deep_Fritz_vs._Vladimir_Kramnik

did you even read your link you gooftard, he made a blunder that cost him the game that he otherwise had in the bag lol

dont be a nit

On June 20 2009 19:03 bigredhoss wrote: Show nested quote + On June 20 2009 15:33 Baal wrote:   On June 19 2009 20:52 bigredhoss wrote:   On June 16 2009 19:44 NeillyJQ wrote:   On June 16 2009 11:59 Sennpu wrote: lol... if you can't beat a bot, quit poker also ignore people who say "if you can't beat a bot, quit poker" because they really have no idea what the fuck they're talking about. the polaris bot has fared pretty well against some of the top LHE HU specialists in the world. since LHE is much simpler on the game theory tree than NLHE, it might be true that there's no bots that could currently beat a good NLHU player, but it's just a matter of time before they figure out how to balance all the variables in NLHE to make a GTO-bot for that game. No, SUPERcomputers still have trouble fighting with grand masters in chess, a game of complete information, making a bot to defeat Poker is much more harder. Not sure what you're saying no to. It's a fact that HU LHE has pretty much been solved by the alberta computer poker research group, and if the polaris bot was a person it would be one of the top 10 or so HULHE players in the world.

LIMIT poker....

On June 21 2009 07:06 Night2o1 wrote: Show nested quote + On June 20 2009 19:17 genjix wrote: wtf r u on about. chess programs crush chess champions http://en.wikipedia.org/wiki/Blunder_(chess)#Deep_Fritz_vs._Vladimir_Kramnik did you even read your link you gooftard, he made a blunder that cost him the game that he otherwise had in the bag lol

Super computers starting showing an edge in chessmasters from 2005 to date, before that humans had clearly an edge, plus these are specialized super computers.

On June 21 2009 22:52 Baal wrote: Show nested quote + On June 20 2009 19:03 bigredhoss wrote:   On June 20 2009 15:33 Baal wrote:   On June 19 2009 20:52 bigredhoss wrote:   On June 16 2009 19:44 NeillyJQ wrote:   On June 16 2009 11:59 Sennpu wrote: lol... if you can't beat a bot, quit poker also ignore people who say "if you can't beat a bot, quit poker" because they really have no idea what the fuck they're talking about. the polaris bot has fared pretty well against some of the top LHE HU specialists in the world. since LHE is much simpler on the game theory tree than NLHE, it might be true that there's no bots that could currently beat a good NLHU player, but it's just a matter of time before they figure out how to balance all the variables in NLHE to make a GTO-bot for that game. No, SUPERcomputers still have trouble fighting with grand masters in chess, a game of complete information, making a bot to defeat Poker is much more harder. Not sure what you're saying no to. It's a fact that HU LHE has pretty much been solved by the alberta computer poker research group, and if the polaris bot was a person it would be one of the top 10 or so HULHE players in the world. LIMIT poker....

Right, I said HULHE is solved. And I said there's probably no bot that can currently beat a good HUNLHE player. So again wtf are you saying no to.

anyway though, isnt the point that these sorts of bots would only really WANT to show up in less skilled games by default since they cant really play around in high stakes low traffic games without getting noticed..

so, they'd more likely be mucking around in something much more high traffic that they can just hammer away at quietly?

bump for those who wanted an update to this guys blog

On June 16 2009 11:59 Sennpu wrote: lol... if you can't beat a bot, quit poker

A well programmed shortstacking bot would crush any of us at the tables.

On September 26 2009 08:46 Maynard! wrote: Show nested quote + On June 16 2009 11:59 Sennpu wrote: lol... if you can't beat a bot, quit poker A well programmed shortstacking bot would crush any of us at the tables.

thats a scary thought

and why i generally stick to the 50bb min tables