Poker Bots Invade Online Gambling

[frEQuency]

Just like in the rest of the world you can't forget the multi-pronged approach for defrauding a group of unsuspecting users. There is also use of "fluffers"who act as players being similar to the bots being spoken mostly of and they sit in the game and also watch player tendencies based on hand history along with it's hard-coded program functionality.

These will use the information they process to make incremental bets at prime positions in order to maximize pot value and create pot odds giving more reason to add more money for the unsuspecting. The person raising is not making the money yet is the one drawing attention....do the math. These groups are the ones getting away with it still and not being mentioned in quotes of losses and seizures and I would have to guess account for most of the wrongful exchange of chips.

 

[atreides]

I am also interested in joining this study group... I think this will be a great exercise if we end up building something

 

[Deleted member 2387]

So an interesting problem is how do you write code to spot bots ?

Maybe if online gaming made it possible to trash talk to your opponents. Seriously, how many cookie cutter responses can a bot designer make before they start repeating?

 

[Scooot]

Yes, good point. I don't think we have enough time and resource to talk about "poker bot" in general. Just pick one specific game and get to know it well. Maybe you can make a suggestion of a game and type of bot that are not, shall I say, "extensively researched" and opportunities to explore are worthy of our investment.

Just like new people here come saying "I want to be a quant", I guess it's wise to have experienced members to point to an area in quant finance where we should read more about.

I'd say HU PLO.

If anyone would like to talk about building some software/program for analyzing PLO please shoot me a PM. I have some dream "gadgets" I wish I had to help advance my game that just aren't available on the net. These are not designed for in game use...just homework to perfect a strategy and find errors in my opponents play.

 

[CutiePi]

Jumping in a little late in the thread but I would definitely be interested in joining/contributing to this project as well; I currently work at a hedge fund but played live cash in Vegas for a year (as well as tournaments). What's interesting about poker is not the mathematical decision making component, but rather the behavioral pattern recognition. This is what is so difficult to model in bots; the math is easy - however I'm willing to argue that a strictly mathematically oriented player (i.e. looks for +EV situations and pushes volume) is probably barely profitable, on average. Assuming someone is playing to maximize profit (not always the case, whether the player is conscious of this or not), the math alone I think will really limit maximizing your win rate.

http://www.fifthidea.com/artificial...ficial-intelligence-the-human-mind-and-poker/

 

[APalley]

Jumping in a little late in the thread but I would definitely be interested in joining/contributing to this project as well; I currently work at a hedge fund but played live cash in Vegas for a year (as well as tournaments). What's interesting about poker is not the mathematical decision making component, but rather the behavioral pattern recognition. This is what is so difficult to model in bots; the math is easy - however I'm willing to argue that a strictly mathematically oriented player (i.e. looks for +EV situations and pushes volume) is probably barely profitable, on average. Assuming someone is playing to maximize profit (not always the case, whether the player is conscious of this or not), the math alone I think will really limit maximizing your win rate.

I believe Chris Ferguson would be a good counter example.

 

[CutiePi]

Have you watched him play? Yes, he is one of the more "textbook" math players, however he (as just about any pro) understands the need to be unpredictable and shift gears. I.e. play the player and not just the cards and math.

Morever, there's an entirely separate argument looming in the background here - poker is an ever evolving industry, just like anything else. Online poker is probably the most overt testiment to how rapidly the game has evolved. If you look at a player's chart from 5 years ago in most cases you'll see a decreasing trendline... as the games have gotten tougher as players have improved. It's exactly analogous to the evolution of the stock markets (remember when coupons were ripped off a sheet of paper and run down Wall Street...) and the diminishing arbitrage due to efficiency.

I probably more clearly could have stated that as players in general improve, the profit due to purely mathematical decision making decreases.

Most of the name players today began playing pre-internet poker; I'm not undermining the success they've had but poker 5 years ago v. poker today has changed dramatically. Another subtle nuance is live v. online play - they are two completely different games. To say nothing of tournament v. cash (of which I believe Ferguson is predominantly of the former).

I would also be surprised if Chris Ferguson himself would agree with the statement that he is a purely mathematical player.

 

[Andy Nguyen]

Quick survey of the landscape: who are the makers of add-on intelligence software to help players analyze and improve their play? From what @Scooot said, it does sound like you have to build it yourself if you want something more sophisticated.

 

[Scooot]

PokerTracker and Holdem Manager are two systems you can purchase that helps you analyze your opponents play as well as your own.

They both datamine hand histories (HHs) which are basic standardized text files created after each hand is played...sort of like a report of the hand. They then can help you analyze a whole slew of different parts of the game and track your progress.

I use HEM as it works better with PLO but both are very good and will improve your game dramatically.

The other big key benefit is that they imploy Heads up Displays (HUDS) that overlay on each poker game you're playing online. What you get is a pop up box under each opponents name with their information under it. What it would look like to someone not familiar with it would just be a box with a bunch of numbers on it:

25/20/5/2.5
80/65/6
1200/1.5

For example. What those numbers mean to me can tell me a lot about the players game

From the top left to the bottom right:
25 = % of hands this player plays
20 = % of hands this player raises preflop
5 = % of hands this player 3bets preflop (for NLHE it would be higher)
2.5 = aggression factor (fraction of aggressive actions vs passive actions....bets + raises / checks + calls + folds)
next line
80 = % of flops this player cbets (continuation bets...raises preflop then bets the flop)
65% = % of hands this player folds to cbets (calls a preflop raise and folds to a flop bet)
6 = % of hands this player check/raises a cbet
netxt line
1200 = number of hands of data we have on the player (very important)
1.5 = players winrate of those hands in terms of big blinds per 100 hands or ptbb



You can customize these stat boxes with whatever stats you want...there is a TON of standardized ones to choose from and you can also create your own if the software doesn't have what you're looking for.

 

[CutiePi]

Also might be worth pointing out, I don't think of Hold'em Manager (there's various other poker apps like Table Ninja which simply allows you to assign shortcut keys, very useful if multi-tabling) or other poker software as "bots" - to me (and most players I know) the term "bot" carries a negative connotation, usually implying an artificial intelligence program designed to make auto-responses and thereby increase volume.

Sites don't allow for multi-tabling auto responses i.e. the point is that there needs to be a live player forcing decisions behind every mouse click. (I believe Scooot also touched on this earlier in the thread, regarding checking for bots by sending a chat message).

I also am very much of the opinion that while you will have a few bad apples, for the most part the sites are very secure and regularly patrol to prevent these types of programs. They are insane money making machines and it's only in their interest to continue to do so.

Would also agree with endorsing Hold'em Manager, it's incredibly valuable to not only be able to build a statistical model of your opponent (the HUD) during live play, but Manager also is very useful for reviewing your sessions and looking for leaks in your game (negative EV situations) to help improve your win rate.

 

[Andy Nguyen]

There was a discussion on Quantnet about how some trading firms favor poker players as potential recruits. SIG is known as a poker friendly firm.
http://www.quantnet.com/forum/threads/wall-street-is-hiring-gamblers.4025/

 

[Scooot]

Andy,

SIG is def a firm that seeks out poker players. Another Leggo professional is a SIG trader out of Dublin. Also Bill Chen and some other pros work there as well. Bill Chen has probably THE most comprehensive poker book on the market....Mathematics of Poker. (at least...print books...some ebooks are VERY good but cost thousands of dollars).

Also, you'll notice in that thread you linked that Darse Billings name pops up. The same Darse Billings from the article linked in the OP of this thread.

Scott

 

[David Rappaport]

Scott,

I have the mathematics of poker.
Can you talk more about the e-books?

Thanks.

 

[Scooot]

Tri Nguyen (SlowHabit) has put out some nice books. Google around for that information.

Tri and Cole South (CTS) put out a book awhile ago called "Let there be range" which is amazing.

Andrew Seidman (Balugawhale) has put out some ebooks as well I believe.

Tom Chambers (LFTV or LearnedFromTV) has put out some good PLO content. Don't know the name of it, he emailed me some of it while he was working on it...very good stuff.

Reid Young (Shootaa) put out book called "Blue Book" which I believe is suppose to be very good as well. He is from Leggopoker so full disclosure but I see absolutely no income from book sales...I just know about it through the site.


For NLHE I'd recommend checking out Cole and Tri's book or looking at Reids book if Cole's is too expensive...no idea how much Reids is?

For PLO I'd recommend Tom's books. Even his rough drafts were amazingly comprehensive. I *think* I'm allowed for share that information but if anyone wants to look at it, PM me and I'll ask him if I can pass it on.

All of these books are going to be $300 or more. And I'd say the better ones will run $1k - $4k depending. Definitely don't purchase any of these if you are anything less than an advanced player already. These are really only designed for a very small, advanced player set that has a deep understanding of poker theory and concepts.


Scott

 

[Spurious]

Just wanting to chime in, I might be interested in joining a group, depending on the goal.
I obviously dont want to run a scheme that puts bots across various sites to profit. I would see it as an opportunity to increase my skills.

As Scooot, I am an active member at 2+2. Scooot might recognize me from my screen name there. I've played Small Stakes Cash games in NL Holdem and PL Omaha 6max and some Heads Up (1on1).
I am nowhere near on the same level as Scooot and havent really played the game in over a year, but I still got a fundamental knowledge of the game and I think I could definitely contribute.

My game recommendation would be some sort of Sit n Go. PokerStars recently stopped spreading a form of Sit n Goes by the name "Double or Nothing", which was basically: if you finished top 5 you got 2x your buyin or nothing if you finished 6-10th. They stopped spreading it because it had a lot of collusion (dont know about botting). This would have been an easy game to work with, because the strategy behind it is not too difficult.

 

[Scooot]

Of course I recognize Spurious. And don't let him kid you, I imagine that anyone posting heavy strat content in the strat forums of 2p2 are probably in the top 1% of poker players in the world in terms of skill.

I'm not interested in SNG analysis. In my opinion, there is much less value in those games vs cash games and the amount of knowledge in that form of poker is extremely high because its a "simpler" version of poker. Thus making it more efficient. That's just my two cents though, I'm sure I can people who disagree with this.

Unfortuntaley it feels like this thread is losing interest. But if any of you programmers would like a fun and interesting project to tackle I have quite a few ideas already drawn up for data analysis and would love to talk shop at least and bounce some ideas around if nothing else. Like I said before...there's a whole host of things PLO players could be using to improve their games and exploit their opponents inefficiencies in strategy but its such a complex game that it's hard to think about accurately.

EDIT: I guess I need to make this clear before I start getting PMs on 2p2. I'm not looking to build an actual bot that can sync up to poker sites and play autonomously, I'd like to build some software that can aid in analyzing strategy.

 

[Spurious]

Of course I recognize Spurious. And don't let him kid you, I imagine that anyone posting heavy strat content in the strat forums of 2p2 are probably in the top 1% of poker players in the world in terms of skill.

I'm not interested in SNG analysis. In my opinion, there is much less value in those games vs cash games and the amount of knowledge in that form of poker is extremely high because its a "simpler" version of poker. Thus making it more efficient. That's just my two cents though, I'm sure I can people who disagree with this.

Unfortuntaley it feels like this thread is losing interest. But if any of you programmers would like a fun and interesting project to tackle I have quite a few ideas already drawn up for data analysis and would love to talk shop at least and bounce some ideas around if nothing else. Like I said before...there's a whole host of things PLO players could be using to improve their games and exploit their opponents inefficiencies in strategy but its such a complex game that it's hard to think about accurately.

EDIT: I guess I need to make this clear before I start getting PMs on 2p2. I'm not looking to build an actual bot that can sync up to poker sites and play autonomously, I'd like to build some software that can aid in analyzing strategy.

Yeah, seems like the thread is losing interest, unfortunately.

I would be more than willing to work on a project regarding cash games. I've never really played any form of tournament poker, so my "expertise" is definitely in cash games.

 

[alain]

I think a lot of people here are doing midterms at this time. I'm sure interest will pick up after that. I will probably PM you guys individually.

 

[Andy Nguyen]

If we set up a working group, it will most likely be for NYC-based folks. Let start it small and keep expectation low first.

Here is a frank question and I expect frank comments.

It seems to me the folks who play poker professionally don't make much money when time invested is taken into account. We are not talking about players like Scotty Nguyen but thousands and thousands who make 5, 6 figure per year.
Compare that to many kids here with a bachelor degree, do MFE and make that much after school.

I have seen many applicants this year from people who have made 100-150K/year playing poker. I do not want to imagine the number of hours they spent online playing, studying stats and what not.

Is it a case of "grass is greener" when they look to a different line of career? How many of these people will crash and burn out after a few years with nothing (saving, degree, career) to show for it?

 

[CutiePi]

I'm glad this is all "frankness"

Grass is greener - it's all relative... and arguably the same case for both professions.

A happy, healthy, productive life will in most liklihood comprise balance between work, friends, family, etc., right? Some poker players crash and burn, some Wall Streeters do too. Some poker players die broke and peniless because they lost it all on the horses... Wall Street... you know where this is going.

RE: money - again, all relative. Making 100K as a Quant in NY happens because of the cost of living. When I lived in Las Vegas several cost of living calculators put me at roughly 1/2 of NYC - I would estimate it was more 1/3. And you're leaving out arguably the most valuable component - as a poker player you are your own boss and never have to deal with a single a$$hole on Wall Street... priceless IMHO.

How many hours do you invest in your job? This is key - poker players are not simply ambling around, occasionally playing (ok, well, some are) but the time invested in stats, reviewing hands, etc. is simply career advancement. How many Wall Streeters spend 3 years on the CFA exam (which I'm willing to argue is simply a money making scheme anyhow). I'd be curious to hear the 3 players on here break down a rough estimate of their week; when I was playing full time (and live is VERY different from online, of which the other players are) I was roughly 25 hours of play and 15 review. I don't consider talking through a hand at a BBQ with a glass of wine "part of my job" - because I genuinely enjoy it.

Anyhow, I'll try to keep this civilized - it's all relative. A great lifestyle depends on the person and how they balance it; you'll find miserable people in any field.