#41
|
|||
|
|||
Re: For the naysayers
It's software. Replicate it on another system, ad infinitum.
|
#42
|
|||
|
|||
Re: I think you are
Just the rumor of a bot being on PartyPoker could drive fish away. The suggestion of impropriety is enough to plant doubt into some minds.
And unlike the current alarmists crying "rigged sites", this post exists to confirm there really is a bot playing at PartyPoker. |
#43
|
|||
|
|||
Re: For the naysayers
[ QUOTE ]
Quote: -------------------------------------------------------------------------------- What do you think a poker bot can do better than the human who wrote it can do (besides play 80 hour sessions with no breaks)? -------------------------------------------------------------------------------- Or in other words - Why are you afraid of a bot? [/ QUOTE ] What can a bot (computer program) do better than the human who wrote it.....hmmmm...well let's see: - If I write a simple program to count to 1,000, it could do it about 1,000 times faster than I could That is the simplest example I can think of. What makes a good poker player? Good decisions. Based on what? - card selection? - understanding of odds? - agression? - understanding how your opponents will act/react? You at least have to concede that a computer is unmatched in the "mathematical" part of the game. It's calculations are precise and never wrong. How about reading a player. When you read a player, what exactly do you think you are doing? A Vulcan mind meld without the touching? Are you reading their mind? No. What you are doing is recalling what they have done in the past and attempting to infer what they will do in the present based on the information you have on this opponent. Do you make notes on every player? How about on every play? Can you sift through all your notes in a reasonable enough time? Do you just rely on memory? How good is your memory? A computer can make notes on every play made by every player. It can examine these notes a lot faster than you could ever possibly imagine. (It would not need to record every play made, just the relevant information). A computer can make most decisions you make...probably better. Poker playing bots are not "several years" away. They are just around the corner. (When I say "they", I speak specifically about the UofA bot that was tested against the Count). I do not believe the original poster in this thread is anywhere near as sophisticated as the AI department of the UofA and I don't believe HIS BOT will ever be a real threat to anybody (no offense intended here). The creators of the UofA bot have only to give it the ability to "learn" and "adapt" to the play of others, or in other words, to make better decisions based on the play of others. Do not for a moment believe that this cannot be done or that is even that difficult to do. The creators need only determine what it is they want/need their bot to learn. They already have the methods to make it happen. Believe me, once a computer learns something...IT IS LEARNED. This same AI department created a chess bot that could defeat a world champion. Now you can sluff that off as "chess" is an easier game to defeat using AI because it has a smaller search space of possible outcomes or some such nonsense. These guys did this 10 years ago when computers were a fraction as powerful as they are now and AI methodology was still in its infancy. When we learn something, because we are NOT COMPUTERS, we sometimes deviate from the proper actions based on what we have learned. A computer does not. I am not being alarmist or suggesting that we should be immediately wary of poker playing "bots". However, I am certainly not naive enough to think that - A world-class poker playing bot is not yet achievable (it is just around the corner) - When one is created, that it will not be able to beat the best in the world (it will) - That if someone could have an army of these bots playing online and making them a fortune that they would not (they would) I don't believe there is anything to worry about right now, or in the immediate future because the people that have the ability to create a world-class poker playing bot are scientists, not scumbags. But ask yourself this. If you could have a few hundred world class bots playing poker for you, would you? Sites should already be aware of this and be looking for ways to make this impossible. I'm not sure they could if they wanted to. (wow, this got really long...sorry about that) |
#44
|
|||
|
|||
Re: A Bot\'s Life: 2
[ QUOTE ]
crack the network packets or Windows commands of the Poker clients [/ QUOTE ] Use hooks http://msdn.microsoft.com/library/de...dn_hooks32.asp They work. Also interesting reading at http://www.codeproject.com/dll/hooks.asp?target=hook |
#45
|
|||
|
|||
Re: For the naysayers
I guess what I'm really trying to get at is real-world situations.
A bot has AA vs. 1 opponent. Flop is KK4 and gets capped. Now what can a bot do differently than you? A bot has AA. Flop is KT8 suited and gets capped. Now what can a bot do differently than you? A bot has KK. Flop is AQ4 and gets capped. Now what can a bot do differently than you? A bot has JJ. Flop is AK3 and gets capped. Now what can a bot do differently than you? Now what happens when you "learn" how the bot plays? What do you gain by knowing the bot folds in each of the above situations? What do you gain by knowing the bot calls in each of the above situations? |
#46
|
|||
|
|||
Re: For the naysayers
It's not the fact that the bot is unbeatable.
It's the fact that the bot is good enough to be at least a slight winner, taking up seats that would have gone to fish. It's the fact that it is possible to create so many bots that they take enough money out of the game to drive the fish away. And of course, the fact that the bot can be replicated to create a profitable volume enterprise. |
#47
|
|||
|
|||
Re: For the naysayers
Good post/response!
I add to that: Bots don't go on tilt or mini-tilt... Bots will make online play worse then it already is.... It has NO BENEFITS to anyone other than it's abUSER. -t |
#48
|
|||
|
|||
Re: For the naysayers
Party lets you play 4 tables. Play 4X at each of their skins. Set-up multiple accounts. Use 5 computers. Play 2 different bots (or 3 or 4) using different accounts at the same table and have them share information.
It wouldn't be difficult to program a bot to beat 1/2 online just playing ABC poker. Do the math. If it won 1BB/hr * 20hr/day * 300day/yr * 25tables = $300,000/yr easy. What if it could beat 2/4, or 5/10? Geez, think outside the box. Lost Wages |
#49
|
|||
|
|||
Re: For the naysayers
[ QUOTE ]
I guess what I'm really trying to get at is real-world situations. A bot has AA vs. 1 opponent. Flop is KK4 and gets capped. Now what can a bot do differently than you? A bot has AA. Flop is KT8 suited and gets capped. Now what can a bot do differently than you? A bot has KK. Flop is AQ4 and gets capped. Now what can a bot do differently than you? A bot has JJ. Flop is AK3 and gets capped. Now what can a bot do differently than you? Now what happens when you "learn" how the bot plays? What do you gain by knowing the bot folds in each of the above situations? What do you gain by knowing the bot calls in each of the above situations? [/ QUOTE ] If you think that a bot is one-dimensional (follows a simple set of pre-defined rules), then you are absolutely right. In each of the situations above, ask yourself what you would decide to do and how you made your decision. The answer to your question is that an "expert" bot can do all of the things you can do. It can probably do it better because - It can remember every hand it ever played against an opponent - It can reference every single one of these hands when it is determining the best course of action You will not be exploiting the weaknesses of the expert bot. It will be exploiting yours! Situations where you are making the "best" decision are where the "expert" bot will make the same decision. Situations where you are NOT making the best decision are where the "expert" bot will not make the same decision. It will make the "best" decision. The "expert" bot will also be able to decide, much better, than you or I, whether "tight agressive", "loose agressive" or somehere in between is the best style for the game it is in, or select games more efficiently where it would have the best advantage. Once again, don't underestimate the potential of an expert bot or the AI community. "Expert" systems (bots) have been forecasting your weather for several years. |
#50
|
|||
|
|||
Re: For the naysayers
For the most part, I agree with OldLearner's response on the advantages a poker bot has over a human. Even Sklansky agrees that a computer program could be written at least good enough "with proper game theory tactics and randomizations so that heads-up nobody could beat it" (HPFAP, page 153). But that's not the point--poker, in my opinion, seems like a game in which a computer has a much larger advantage than in something like chess.
Anyway, I just wanted to point out that you're probably wrong about reading the packets off the network and using them to create a bot, at least for a few of the online poker sites. While the packets are encrypted, and thus a third party observer cannot read them, all the information that a given client needs to read a packet coming to it off the network is on the computer. For example, Paradise Poker uses SSL, specifically the OpenSSL implementation, when communicating over the network. This means that not only do you have access to their encryption algorithm in source form (as opposed to assembly/binary) (just want to point out that this is, of course, a good thing, as I don't, and no one rightfully should, trust Paradise Poker programers as cryptographers), but you have a pretty good idea of where to start looking in memory for the key that's currently being used. All it takes is a skilled reverse engineer and you could have the memory location of the key in at most a day (probably less time, since PP uses OpenSSL in a DLL, thus with exported symbols, which means you can use a kernel debugger and break on calls to the decryption routine). Anyway, that would at least cut down on the overhead involved in running multiple tables, and if the communications protocol isn't too advanced (I don't see why it would be), you might even be able to create cloned clients so you can have multiple log-ins on the same machine, thus running more than 3 tables. --EnderX |
|
|