The difference between game theory and decision theory is that in game theory you need to worry about not just what is rational for you to do, you also have to consider what is rational for the other players to do.
When you play online poker, you are placing complete trust in your analysis of the trustworthiness of the ‘house’. The house could cheat you, if it desires, and probably not get caught. But you analyze that cheating would be an irrational thing for the house to do an a large scale—because large scale cheating would get caught and they would lose customers.
So, what is your analysis of what they would do to a small number of their customers who violate their rules by using real-time machine assistance? They can’t take those people to court. People who, if they were allowed to get away with it, would destroy the online poker business. Would it be rational for the Poker houses to try to cheat the rule-breakers?
Would they actually do that?
I used to count cards at blackjack. And when I did it in Reno, at a certain stage a new dealer would be brought to the table (outside the normal shift schedule). And from that point on, I would lose money. If I watched closely, I could see them dealing seconds.
The difference between game theory and decision theory is that in game theory you need to worry about not just what is rational for you to do, you also have to consider what is rational for the other players to do.
More like: in decision theory, there aren’t necessarily other actors to consider.
So, what is your analysis of what they would do to a small number of their customers
who violate their rules by using real-time machine assistance?
That’s not a violation of the rules. Assistant programs such as discussed above are used by essentially all serious players and that’s fine by the sites.
People who, if they were allowed to get away with it, would destroy the online
poker business.
Actually, online poker business is growing every year.
… a small number of their customers who violate their rules by using real-time machine assistance …
That’s not a violation of the rules. Assistant programs such as discussed above are used by essentially all serious players and that’s fine by the sites.
Use of real-time machine assistance to guide your play is not against the rules? Then apparently, I don’t understand the rules. What kinds of computer assistance do they forbid?
5.6. AUTOMATIC PLAYERS (BOTS). The use of artificial intelligence including, without limitation, “robots” is strictly forbidden in connection with the Service. All actions taken in relation to the Service by a User must be executed personally by players through the user interface accessible by use of the Software.
The first long list is about programs that are allowed, btw.
But I guess I should have been been more specific that I was talking in the context of “bot-like” programs, and what I said is completely accurate in such a context.
Of course additionally e.g. programs that show your cards to your friends are forbidden. And utilizing large centralized databases is forbidden. But programs that do such things are not “bot-like”.
EDIT: Ok, actually I am wrong. They go further in banning “bot-like” programs than I described here. I knew that the multi-purpose “poker analytics suites” that people like me widely use are allowed, and that real bots are forbidden, but I was mistaken where exactly they draw the line between these types of programs, since I hadn’t really looked into such.
I used to count cards at blackjack. And when I did it in Reno, at a certain stage a new dealer would be brought to the table (outside the normal shift schedule). And from that point on, I would lose money. If I watched closely, I could see them dealing seconds.
This seems like quite an irrational strategy on the casino’s part, and a bad case of loss aversion.
Suppose the casino calculates that it can afford to cheat one customer in twenty without getting caught (or any other proportion if 1⁄20 is unrealistic). Of all the people they could use this ability on, is the card-counter really the optimal choice? He may not have come in with very much money, and is likely to catch on reasonably fast at which point he will probably just leave. It seems much better to wait for someone who is at least slightly drunk and has come in with a lot of money to gamble, he probably won’t notice and will just keep making and losing large bets. If I recall correctly the margins in card-counting are quite small and they may be able to take many times as much money from the second guy as the card-counter can take from them.
Although you obviously have much more experience in this area than me, so please correct me if I’m wrong.
One of my friends made a living for a while by card-counting, but iirc, eventually he was banned from the casinos in Atlantic City. That seems like a cleaner way of handling things than bringing in a cheating dealer.
This seems like quite an irrational strategy on the casino’s part.
Hmmm. Does it also seem irrational of them to break the leg of someone who borrows a lot of money and can’t pay it back? Wouldn’t it be more rational for them to break legs of random drunks who are less likely to recognize their assailants?
I hope my flip response doesn’t seem too hostile. Of course the casinos that cheated me weren’t doing it to make a few extra bucks. They were doing it to punish me. Of course, for blackjack card counters, simply banning the player is also a viable strategy. And it happened to me in more than one casino. But banning doesn’t work so well with online malefactors.
Not a very effective punishment, since you could just get up and leave. If they really wanted to punish you they could have taken a photo of you and distributed it to every other casino in the country.
Would it be rational for the Poker houses to try to cheat the rule-breakers?
While I can’t vouch for every single poker site out there, the chances of them trying to cheat you are almost nil. The amount of marginal income they would make compared to the risk of a tarnished reputation makes it a foolish play (note that to cheat you, they would need to not only rig the game, but have a shill in there to divert the money to).
I used to count cards at blackjack. And when I did it in Reno, at a certain stage a >new dealer would be brought to the table (outside the normal shift schedule). And >from that point on, I would lose money. If I watched closely, I could see them dealing >seconds.
Unless this was 1950, I believe your eyes fooled you. They would ban you, yes, but the idea of a Reno casino trying to win their money “back” via slight of hand is a little silly.
The difference between game theory and decision theory is that in game theory you need to worry about not just what is rational for you to do, you also have to consider what is rational for the other players to do.
That seems to be necessary somewhere along the line when implementing decision theory too. If you don’t model the other agents in the system you will (obviously) end up making worse decisions.
So, what is your analysis of what they would do to a small number of their customers who violate their rules by using real-time machine assistance? They can’t take those people to court. People who, if they were allowed to get away with it, would destroy the online poker business. Would it be rational for the Poker houses to try to cheat the rule-breakers?
I don’t make such analysis. I assume the worst case with respect to them choosing among plausible methods of reciprocation and instead focus on the logistics of whether calculator use can be detected.
The house cheating seems to be a rather benign situation. The probability that the cards being dealt this bad or worse over multiple hands given fair deals can be calculated. If there is a high probability that the house is cheating you stop.
The difference between game theory and decision theory is that in game theory you need to worry about not just what is rational for you to do, you also have to consider what is rational for the other players to do.
When you play online poker, you are placing complete trust in your analysis of the trustworthiness of the ‘house’. The house could cheat you, if it desires, and probably not get caught. But you analyze that cheating would be an irrational thing for the house to do an a large scale—because large scale cheating would get caught and they would lose customers.
So, what is your analysis of what they would do to a small number of their customers who violate their rules by using real-time machine assistance? They can’t take those people to court. People who, if they were allowed to get away with it, would destroy the online poker business. Would it be rational for the Poker houses to try to cheat the rule-breakers?
Would they actually do that?
I used to count cards at blackjack. And when I did it in Reno, at a certain stage a new dealer would be brought to the table (outside the normal shift schedule). And from that point on, I would lose money. If I watched closely, I could see them dealing seconds.
More like: in decision theory, there aren’t necessarily other actors to consider.
That’s not a violation of the rules. Assistant programs such as discussed above are used by essentially all serious players and that’s fine by the sites.
Actually, online poker business is growing every year.
Use of real-time machine assistance to guide your play is not against the rules? Then apparently, I don’t understand the rules. What kinds of computer assistance do they forbid?
(EDIT: I am mistaken in what I state in this comment. See comments below for correction.)
Essentially the deciding factor whether an assistant program is allowed is whether the program does the mouse-clicking for you.
Quoting from the Terms of Service of the biggest poker site:
5.6. AUTOMATIC PLAYERS (BOTS). The use of artificial intelligence including, without limitation, “robots” is strictly forbidden in connection with the Service. All actions taken in relation to the Service by a User must be executed personally by players through the user interface accessible by use of the Software.
That was a pretty remarkable example of selective quotation.
Immediately above the paragraph you quoted was a link to this FAQ. Note the long list of software products which they say you are forbidden to use.
The first long list is about programs that are allowed, btw.
But I guess I should have been been more specific that I was talking in the context of “bot-like” programs, and what I said is completely accurate in such a context.
Of course additionally e.g. programs that show your cards to your friends are forbidden. And utilizing large centralized databases is forbidden. But programs that do such things are not “bot-like”.
EDIT: Ok, actually I am wrong. They go further in banning “bot-like” programs than I described here. I knew that the multi-purpose “poker analytics suites” that people like me widely use are allowed, and that real bots are forbidden, but I was mistaken where exactly they draw the line between these types of programs, since I hadn’t really looked into such.
This seems like quite an irrational strategy on the casino’s part, and a bad case of loss aversion.
Suppose the casino calculates that it can afford to cheat one customer in twenty without getting caught (or any other proportion if 1⁄20 is unrealistic). Of all the people they could use this ability on, is the card-counter really the optimal choice? He may not have come in with very much money, and is likely to catch on reasonably fast at which point he will probably just leave. It seems much better to wait for someone who is at least slightly drunk and has come in with a lot of money to gamble, he probably won’t notice and will just keep making and losing large bets. If I recall correctly the margins in card-counting are quite small and they may be able to take many times as much money from the second guy as the card-counter can take from them.
Although you obviously have much more experience in this area than me, so please correct me if I’m wrong.
One of my friends made a living for a while by card-counting, but iirc, eventually he was banned from the casinos in Atlantic City. That seems like a cleaner way of handling things than bringing in a cheating dealer.
What are the rules on video recording in casinos?
Hmmm. Does it also seem irrational of them to break the leg of someone who borrows a lot of money and can’t pay it back? Wouldn’t it be more rational for them to break legs of random drunks who are less likely to recognize their assailants?
I hope my flip response doesn’t seem too hostile. Of course the casinos that cheated me weren’t doing it to make a few extra bucks. They were doing it to punish me. Of course, for blackjack card counters, simply banning the player is also a viable strategy. And it happened to me in more than one casino. But banning doesn’t work so well with online malefactors.
Not a very effective punishment, since you could just get up and leave. If they really wanted to punish you they could have taken a photo of you and distributed it to every other casino in the country.
While I can’t vouch for every single poker site out there, the chances of them trying to cheat you are almost nil. The amount of marginal income they would make compared to the risk of a tarnished reputation makes it a foolish play (note that to cheat you, they would need to not only rig the game, but have a shill in there to divert the money to).
Unless this was 1950, I believe your eyes fooled you. They would ban you, yes, but the idea of a Reno casino trying to win their money “back” via slight of hand is a little silly.
That seems to be necessary somewhere along the line when implementing decision theory too. If you don’t model the other agents in the system you will (obviously) end up making worse decisions.
I don’t make such analysis. I assume the worst case with respect to them choosing among plausible methods of reciprocation and instead focus on the logistics of whether calculator use can be detected.
The house cheating seems to be a rather benign situation. The probability that the cards being dealt this bad or worse over multiple hands given fair deals can be calculated. If there is a high probability that the house is cheating you stop.