Poker Bot piece

[TomCollins]

The best bots will adjust to how you play. Plus every bot plays differently.

[UATrewqaz]

That's very true, if you could perfectly identify a bot's poker strategy, you could devise the perfect counter-strategy.

[Pirc Defense]

[ QUOTE ]
Once the algorithms the bot follows are figured out, all we have to do is induce them to fold.

[/ QUOTE ]

Uh, mag, I don't think you've thought about this issue much. Attempting to induce a bot to fold that is playing +EV almost always would mean that you are, at least some of the time, going to be playing -EV. You're basically saying, make a really strange chess move that will confuse the computer! Well, all that happens when you do that is you've got a piece somewhere on the board where it shouldn't be, and the computer is going to take advantage of that. Same with poker. The only way your plan would work in the long run is if the programmer of the bot was stupid, but that class of bot is not under discussion.

[Wake up CALL]

[ QUOTE ]
You're basically saying, make a really strange chess move that will confuse the computer! Well, all that happens when you do that is you've got a piece somewhere on the board where it shouldn't be, and the computer is going to take advantage of that.

[/ QUOTE ]

Utter nonsense, the two are not at all correlated in this respect. If you know how to cause the bot to fold with the best hand the move will always be +EV. You are forgetting that the board resets when the hand ends, think before posting.

[Pirc Defense]

[ QUOTE ]
Utter nonsense, the two are not at all correlated in this respect. If you know how to cause the bot to fold with the best hand the move will always be +EV. You are forgetting that the board resets when the hand ends, think before posting.

[/ QUOTE ]

The bolded part was not in the original argument. What you're saying is obviously true, but also obviously trivial. Follow your own advice?

[ChessMan]

[ QUOTE ]
Once the algorithms the bot follows are figured out, all we have to do is induce them to fold. All you have to do is realize who the bots are.

I'm not sure exactly how they work, but if they are programmed to play only +EV, you just have to give them -EV, and you take the pot!

Now we just need to put together a bot database. It seems as though the best way is to use the bot itself, since it identifies who is using it at the table (to my understanding).

[/ QUOTE ]

You are right on the money. I suspect most bots are quite predictable. Once you discover a predictable opponent, you are making money off of him/her/it. I would argue that some people wouldn't care that the opponent they are making money off of is actually a bot. I don't.

Of course, I understand if someone prefers to make money from humans rather than bots.

Also, I'd like to point out that cheating is cheating. If two opponents collude, it doesn't matter whether they are both bots, both human, or one is human while the other is a bot. A bot doesn't necessarily imply a cheater. However, given that profitability would increase, a bot maker would be tempted to implement collusion.

[Wacken]

The board resets, but the table does not.
If do that particular action to make him fold a few times, he will notice and catch your bluffs. Unless it is of course poorly programmed.

[Sniper]

[ QUOTE ]
If you have an AMM (Automatic Moneymaking Machine), you surely would keep it to yourself. I know I would.


[/ QUOTE ]

History shows that there are many reasons why you would share/sell your money making idea/system. Look around you, and you don't have to look very far to see profitable systems being sold.

Just think about Poker Tracker for example. The programmer has made millions by selling this program, far more than he could have made by simply keeping it to himself and using it to only enhance his own poker playing ability.

[ QUOTE ]
It looks a bit like all the trading systems that are for sale to trade the markets: black box systems with "proven" records etc. Why sell the damn thing if you can make millions with it in the markets? Because they don't work!

[/ QUOTE ]

There are many profitable black box and open source systems actively trading the financial markets.

[ QUOTE ]
Imagine: computers ruling the financial markets. They participate, yes, but they don't RULE it.

[/ QUOTE ]

The overwhelming majority of trades on the finacial markets every day are automated system generated trades!

[Jimbo]

[ QUOTE ]
Just think about Poker Tracker for example. The programmer has made millions by selling this program, far more than he could have made by simply keeping it to himself and using it to only enhance his own poker playing ability

[/ QUOTE ]

PT is a very poor example. He could only play perhaps 4 to 8 tables st one time himself and would still make mistakes and make more as time went on. A perfect poker bot that could beat ring games at any reasonable level at all would never be more profitable if it was sold rather than used. It could play hundreds of tables 24/7, never get tired and never make mistakes. In a few weeks it would garner more profit than PT will in it's entire history. There would be no reason at all to sell this perfect bot to others.

That being said it doesn't and won't exist in our lifetimes. Pirc keeps thinking poker is like chess, it is not and is not comparable. In fact an exploitable flaw in an otherwise poker bot would be quickly discovered by the majority of it's oponnents and exploited. This flaw may be so deeply ingrained that is is not only impossible to fix but to identify in the first place.

Note that 'perfect' play literally means unexploitable play. There is a (possibly mixed) strategy that will make it so that, in the long run the best that an opponent can do against the computer player is break even regardless of what strategy the opponent has or descisions the opponent makes.

An example of a perfect poker strategy is not to play. There's no way for an opponent to stop you from breaking even. Playing on raked tables is a loosing proposition, so really, what you're looking for in poker is exploitative play - taking advantage of the opponent so that you can beat the rake.

Against players that make blatant mistakes (like playing too many hands), a solid bot is likely to do quite well since it will (over time) cash in on mistakes like that.

Since, in order to profit, the bot must be looking for exploitative strategies, it is concievable to have a meta-bot that strings the bot into trying various exploitative tactics, and counter-exploiting them, but that's a *very* tricky thing to do in a game like poker.

In practice, profitable poker is about landing fish, so any sane bot is going to leave a tough table rather than fighting with a strong opponent.