Poker bots, what is the danger?

This is a repost of my response to Poker Butts, who wrote a very good post, that takes a different view point than mine.


Since we have been very post happy recently, I have decided to repost here to reopen the debate on this issue, rather than let it stay buried.


"Well, the most important thing is that it is an restricted domain, "micro-world", kind of application."


Well thought out post, which I disagree with about, the danger of bots. If bots are winning money in the "micro-world" then there are less players moving up in limits with money they won in an easier enviroment. Which means the ones that move up are better players.


Any money that is sucked out of the poker economy matters. Higher rake or bots, it still is less money and the remaining is tougher money.


ProPoker is rumored to have bots, RGP and here at 2+2 internet forum have posts which suggest this. I played a little .50-$1 this week at Propoker, just so I could see if the posts were true. Now, they may not have bots, but they have the tightest $1 limit games I have ever seen. Players with $200-$600 BR's, who only see 22% of the flops, and who never chat. There are live players there kidding about the bots they are playing with, all week long. No one else says a word.


I win about a BB/hr at $10-20 and broke even in the PP $20-40 for about six months. I lost every play this week and the only players at the table, I felt, I had an edge over was the live players. I lost about $150 playing .50-$1, I never had a chance.


Now, if there are players that are beating this game, I don't want to play with them either. When these players move up to the next level they are 3-4 levels below what I play now. WTF am I going to do when the best of them move up to MY level? This isn't even take into account that someone improves on the bot programs.


bots, IMHO, are a problem.


MS Sunshine

While I agree with your trickle-down economics,

it's still debatable if a bot could even beat

a micro limit..Particularly in a tight game

like you descibed at ProPoker, which is kind of strange for such a new site.


There is no denying the toughness of some of these smaller limit games recently.


I recently lost $150 this week at Poker Stars

playing $1-$2. Toughest game I've ever played in at this limit. Then again I also lost over $100 on paradise $1-$2 this week. (I consider myself to be very successful playing online, winning every limit up to $15-$30, having never played the $20-$40).


Even if they cannot beat the game (bots), countermeasure's should be taken because at the least they are only small losers to break even at the micro-limits..The games are getting tight enough as it is.


I like the idea of having a table host that would occassionally pop up a text message such as

"How are you today?" or another random question. (All users should be asked what their preferred language is) to all suspected bots.


Still, players shouldn't start being paranoid just yet, because what will happen is every losing player will start complaining that they are losing because they are against bots, which certainly isn't the case because any good player would beat a table full of bots in my opinion. Online poker has a bad enough rep as it is.

Ms Sunshine,


What are you thinking? You actually believe the reason you lost was because you played against a Bot? Lord help us all!!! Although I never posted in the threads regarding your culpability in the "Prop Scandel", I now believe you had no clue then or now. You post above has completely exonerated you from any possibility of intentional wrondoing.


You played a "whole week" at .50/1.00 and lost a $150. Using this massive amount of data coupled with the fact that you beat mid-limit games over a period of six months your conclusion is that it must have been bots who beat you. Have you never run bad for a week before? Perhaps this is a case for the X files! Rather than bots perhaps it was little green martians taking over our planet at .50/1.00 limits. This is about as likely IMO.


I tried to give you credit for posting tongue in cheek but have reread it twice and can only believe you are serious. My suggestion is you move up to $100/$300 limits ASAP so as to avoid both "The Attack Of The Bots" as well as those superior low limit players you fear. Yes Ms Sunshine run for your life "The redcoats are coming, the redcoats are coming!" Perhaps it is the men in the white coats you should fear more.


"They're coming to take me away, Ha HA, HO HO, HEE HEE, to the funny farm........"


Jimbo

Let me get this straight.

We should be afraid of the bot in seat 10 that is the ultimate poker player.

---He never tilts,

---He always plays his hands perfectly,


Or wait that is not possible is it....


Any good player knows that in order to win in poker in the long run he must adjust to his opponents more than just play his own game


Try bluffin a calling station with the seem move u used on ole timid Tim


last I checked Artificial Intelligence was a movie not reality


Show me a computer program that can adjust its play the way he would have to in order to win and I will give you my BR...


You need to spend the time u worry about bots on better things


Like a whether to check raise me

I dare u

It was the other factors that made it seem odd. As I'm responding to this post I'm checking other poker site's .50-$1 games.


PokerStars 1 game 8 players 54% $3-$35 BRs

WSEX($1-2) 1 game 7 players 67% $30-$109

Party 1 game 9 players $7-$54 $9avg pot


Now, those sound like low limit games.


ProPoker 2 games 15 players 20%(both games) and 3 players under $100 BRs and 8 players over $250 BRs.


Now, no chat in 2 .5-$1 games, isn't that odd?


As to the quote "Prop Scandel", it was the Props and I who were the victims, so to my "culpability" in this is hard for me to follow, but thanks for not calling me on it.


I don't reraise UTG raisers, even if they could be on tilt, with QT. So, I worry about what I want about internet poker. Thank you very much.


Have a nice day.


MS Sunshine

Why can't bots be programmed to play tight-agressive? Is a programming coding rule that they must play weak-tight?


MS Sunshine

the link below

Exactly, what about a table full of players that play using these concepts at .50-$1 games.


MS Sunshine

Hello there, MS. Let me first say that I respect your differing view on the dangers of bots. Let me also say that my view is based on the assumptions that there are very few bots out there, that those that are out there aren't very good (they might be tight, but that does not mean that they are good), yet, and even, and especially, when they maybe gets good, there still aren't going to be many of them, as it really isn't easy to write a truly good poker bot - much, much, much harder than to write a truly good chess bot, and not too many people out there can do even that, and also I made the assumption that there are no secret bots put out there by the cardrooms themselves. If these assumptions do not hold, my view might have to change.


Now over to an explanation of something that I think that you have misunderstood: microworlds. It seems to me that you took this to mean that the bots play micro-limits, but that is not so. A microworld is a term in AI for a simple, well defined, and limited "world" in which an AI program is operating, in contrast to the vast, uncontrolled, and unpredictable real world we all live in. The most famous example of a microworld is the "blocks world", which consists only of a number of blocks of different sizes and shapes and colors, placed on a "tabletop". It is of course much simpler to construct an AI program that is successfull in one of these clearly defined microworlds than in the vast, uncontrolled real world. That is also one of the criticisms of microworlds.


Now, an online poker room could be viewed as a microworld. And it does not matter if the bot plays 0.5/1 or 20/40. The poker room being a microworld should make it much easier to write a bot that operates in it. One way to make it harder (for outsiders) to write a bot for your poker room would be to make it less of a microworld, and more like the real world. Make it more unpredictable. Things like maybe regularly changing the user interface slightly. Make the user interface randomly behave slightly different at different times. We humans easily cope with this. It is hard to write a bot that can cope with this. Of course I'm sure that there are ways around this, but at least it makes it harder. We can never protect ourselves completely from theft, just lock our houses and cars the best we can. It's better than to leave the doors open.


Of course you can still be right about some people writing unsofisticated bots that are tight but can only win at micro-limits, so they leave them there, and money don't move up. But then again, I myself play 0.5/1, even though I have recently started moving up to 1/2 and hopefully soon 2/4 and higher, and I see like 20% flops and very rarely chat, so maybe those players that you saw are not bots at all. Maybe they are just poor, ambitious players like me, who also have grown tired of chatting. Maybe. Probably. I have never played at ProPoker though.

I have played in 0.5/1 games at Paradise where the flop % have dropped to like 12-13%. That is tight. That is too tight. People are not playing good poker. They are just playing tight.


Also, from what I have seen at Paradise, there really aren't that much difference in the average percentages at 1/2 and 20/40. People are more aggressive at 20/40, and actually looser postflop. Of course more people are playing better at 20/40, but the general type of game don't really seem to differ all that much. But I have never played above 2/4, so correct me if I am wrong.


And I am also not surprised that there is not much chat. Paradise's chat stink.

Thanks for the education. I did misunderstand the term microworlds. If the bots, several, are placed in the games by site operaters than most of this can be gotten around. Poker is alot easier to beat, with so many less variables, than chess.


MS Sunshine

"Poker is alot easier to beat, with so many less variables, than chess. "


Maybe the way you play.


Actually I believe expert level chess has already been programmed with the

Deeper Blue experiment. Expert level hold'em has not.


With chess it's just a matter of figuring out what is the best move based on

future moves against a single opponent..For example if I have 10 possible moves.

I simulate move #1. Then I would have to simulate my next 10 possible moves:

move #1a, #1b, etc to get the best possible outcome. This would enable me to think

many moves ahead. I'm sure there's even more to it, like implementing and

recognizing strategies.


Now think about all the complexities of hold'em.

(their are some preflop. But not too deep).

How many players, called preflop? How many bets preflop (1,2,3, capped?)..

Now the flop..Should I bet my ace high against a 589 flop against 1, 2, 4, or 8

opponents? What about KT2? (And 2-3 hundred more flop textures/hand types)Is the leader to my left or right? Is the leader

aggressive or passive. We're already getting deep and we're only on the flop.


Now the turn...We have a board of KT29, but how does the turn card relate if at all

to not only the previous flop action, but previous preflop action? Is it a flush

card? or did it pair the board?


Keep in mind that you have to have nodes branching out for each level, each flop type,

etc..By the time you reached the river you would possibly have hundred's of

thousands of decision nodes. Some of which may possible have multiple actions.


It could possible take a team of 20+ expert hold'em player/programmers a few

years to accomplish such a feat.


If anybody really doesn't grasp what I am saying, they simply haven't thought

it out much..Next time your playing think about all the variables you would

need by the time you reach the river, and how long it would take to input them.

You also could buy Turbo Hold'em and laugh at how bad it plays, but also appreciate

the time it must have taken Bob Wilson to program.


Then take Bot Author who is a self-proclaimed amateur who hasn't even spent too

much time on his bot....He is without a doubt full of it that his program

can beat 2/4 in the long run. It doesn't add up.

LOOK IN THE BIBLIOGRAPHY --LMAO

you are an idiot. read the documents fool.

HEY RETARD--THEY DID USE THE BOOK FOR INFO--MAYBE YOU NEED OT LEARN HOW TO READ..LOOK IN THE BIBLIOGRAPHY SECTION AND YOU WILL SEE MASONS BOOK. WHAT A MORON

http://www.cs.ualberta.ca/~jonathan/Grad/papp/node74.html

A few comments:


If there are bots which beat you, why woudnt they alwready play at the higher limits? Even if the bot owner had cash problems, the bots wining would soon bring anought $'s to move up in limits. (if we assume a 400 BB bankroll, 1BB/h win rate and 10h of play/day, you *double* in limits every 40 days).


As you know, the University of Alberta has develped a poker playing boot: poki. You can play against it at http://www.cs.ualberta.ca/~games/poker/

would you (or better yet, you and other good players) care to play a bit against it to test its strenght? It is surly possible (but probably difficult) to make a better bot then poki. But poki can stand as a benchmark for what a few 1000's man hour dedicated work can give you, bot wise.


Gatlif

hey idiot did you read why its there? obviously not fool. you have to read the material to understand it. i know this is beyond your capability.

I do believe I have had my first experiance with a bots game. Pokerstars 5:10 holdem, there are two games going one full with 35% players on the flop. One 19% on the flop with 6 players I go on the waitting list for the full game and take a seat while I am waitting in the tight game. Two rounds no chat from anyone than after paying both blinds every player at the table leaves in the same split second. At first I thought the site went down but after leaving the game everything was operating fine. Very strange indeed

Forget limit, play PL or NL and don't worry about bots.

Actually I believe expert level chess has already been programmed with the

Deeper Blue experiment.


Made up of Top Class Chess players around the world....I do believe, And it was Beaten!


The same can be done with poker bots.

why do bots always have to be an excuse for people that suck? it's truly amazing...

Very interesting post, and the (serious) replies as well, particularly the one that references "pokibot".


These are issues we should all be looking very carefully at re online poker.


Thanks for the post mssunshine and the resulting dialogue.


I have no respect whatsoever for the usual plethora of replies that make false assumptions about what you said, ignore the majority of points you raised, and instead do the usal "you suck, don't blame it on conspiracy theory about bots, etc etc"


I do have respect for those who can look at these issues, debate and evaluate them serously, and realize the serious potential destroyers of good internet poker, that they could be.


Whenever money is involved, and oversight is poor, there will be dozens of scams being attempted. The naive head in sand, knee jerk crowd here who thinks poker skill is THE only major factor to consider in internet poker success, have their heads in dark places.

If you think reading and employing strategies from S&M books wouldn't help, why do you even read these forums?

You cannot compare chess and poker:


Poker: game of imcomplete information. Cards are hidden and the order of the deck is "random"


Chess: game of complete information. It is all there in front of you.


End of story,

Treefrog

Let me quote the research:


Because of this, even against bad players an optimal strategy is likely to only break even. In contrast, a maximal strategy using opponent modeling (which does not assume perfect play from its opponents) would identify weaknesses and exploit them for profit


So basically, unless the "bot" can predict your plays it will have a very hard time winning. Still so scared? Think about how long it would take to compile a database of specifically your individual plays in every situation. This would be necessary for maximal play by a bot. And that can still be defeated if you mix-up your actions appropriately. The computer could compile a default "maximal" strategy based on all opponents, but that would still be flawed against any one opponent to some degree.


To summarize, computers cannot out think humans when "incomplete information" is involved..they are not intuitive by nature.


Treefrog

The Generic Opponent Modeling performed almost as well as the Specific Opponent Modeling.


From the source:

However, our implementation of specific modeling does not appear to produce a significant advantage over generic modeling. We recognize that the granularity of the context for gathering action frequencies may be so coarse that the error undermines the gain. For example, Loki does not recognize that, for some players, calling a bet is automatic given they have already called a bet previously in that round - the second action contains no information. Some other variables which may make the identification of the context stronger include the previous action, number of opponents and number of callers remaining. However, the number of cases becomes very large so it would seem necessary to somehow combine the action frequencies of ``similar" scenarios.


One thing to consider - at the time this thesis was written they were using a 150MHz computer and large Hard Drives at that time were 4GB. Today with 2GHz+ machines and 100GB+ HD's available, much more work in areas such as opponent modeling can be done.

If it is possible to have abot that can win it will destroy internet poker.

The Generic Opponent Modeling performed almost as well as the Specific Opponent Modeling.


Understood. The passage you quoted is basically saying the same as I was: #1 the generic modeling is not that great and #2 the specific modeling would not be that much of an improvment until there was a huge amount of data gathered.


At best, if a bot were at my table, it would appear as another solid player, and I would adjust accordingly. If it were predicable tight-aggressive I would kick it's ass eventually. If, on the other hand, it were better than me I would adjust and/or avoid it (like I do with other solid players). This is nothing new here.


The only thing that would really scare me is teams of colluding bots. This is only one step further along in the programming. Hopefully this should be trivial to detect though.


Thanks,

Treefrog