PDA

View Full Version : PT & PV are just the beginning... of the end


johnfromvirginia
02-10-2005, 02:33 PM
It strikes me that poker software like poker tracker and poker view may be extremely dangerous to online poker. Eventually, these things will evolve into software that just comes right out and tells you whether to fold, call or raise. The next evolution would just be a bunch of poker bots playing tight, aggressive poker without any emotion. When this happens, the online poker industry will die within months because no one can survive in that world for long, certainly not any of the sheep. If it were up to me, hand history would not be available to players. All this software designed to observe your opponents for you may eventually kill the goose that laid the golden egg.

Dynasty
02-10-2005, 05:21 PM
What % of online players even own Poker Tracker? It can't be very high.

Daliman
02-10-2005, 05:34 PM
Yes, just as the tons of readily available information on correct/good play in books and online has ruined all the games currently played.....

Dynasty
02-10-2005, 05:48 PM
You really should get that middle picture into the avatar mix. It's important because your son starts to lean into the toy basket. It really helps the flow.

johnfromvirginia
02-10-2005, 06:07 PM
Maybe you're right, but I'm not sure I explained exactly what I'm talking about. Poker Tracker still requires some effort and skill to apply--you still have to know how to play poker in order to use PT effectively. PV takes it to the next level and moves more of this info right onto the table--you still need quite a bit of skill to apply the information, but it's a set of skills that any dedicated person can learn. It's no longer limited to those who can observe what's going on around them, remember it all and apply it in the trenches like the seasoned pros of yore. If this were to evolve into "bots," which seems to me is only a matter of time. Then, you'd just turn on your computer and leave for work and when you got home, you'd walk up to the computer and see how much you'd benefited from a day of playing flawless, perfect poker while you were off doing something else. You put one or two bots at each table and the game becomes almost impossible to beat. It would be like sitting down at a table with the two very best pros you can imagine on their very best days and having them never make a technical error.

The only thing that might keep these things from spreading like wildfire is the possibility that they'd be more profitable to their creator if he/she were to use them himself rather than sell them. They could conceivably make more money using them on their own than they could make selling them.

Daliman
02-10-2005, 06:39 PM
I wanted to, but the guy who did the avatar for me (ty SuitedSixes) said that since it had a different, brighter shading to it, it gave an odd flicker effect and just looked bad. I think there wasa cropping issue too. But yes, i wish i could have had it in there.

shummie
02-10-2005, 06:44 PM
I get your point, and it's scary.

For an analogy. Suppose you were playing chess for a living (or like the posters here are playing poker). Now suppose that a deep-blue style (heck, even Chessmaster style) bot was available for you to use a la Poker Tracker. Now regular humans have no chance at making money playing chess online.

Poker is a bit more complicated to program good AI for and there haven't been that many smart guys looking into. However, the smart guys are looking into it now AND computing speeds are ramping up all the time due to Moor's law ect.

One thing that will make programming a good poker player harder than programming a good chess player is the possibility (and really the need) to "change gears" when playing. In chess, any dumbo who knows the rules knows what needs to be done to win the game. Capturing your opponents queen is just about always a good thing to do. In poker, raising your pair of queens on the river is highly situational dependent. I'm being really general here, but this is all a result of the fact that in chess you have perfect information while in poker you don't (i.e. you can't see your opponents cards).

Now, I don't mean to take the chess analogy too far. I'm just trying to say that at the time poker is a harder problem for game theorists and computer scientists to solve. However, no one has proven that poker is inately unsolvable by a computer... especially at lower limits.

If you use Poker Tracker and multi-table, a lot of time you will look at a players stats (40/0) and know EXACTLY what to do in your situation. Poker Tracker could just fold the AJo for you when said player raises preflop. If you've played 50k hands plus, you know that poker can become very repetative.

My best guess is that within 5 years there will be poker-playing PT type programs good enough to beat the fish like we are now. Within 10 years there will be poker-playing PT type programs that can beat US like we are beating the fish now. By this time, playing online will come down to who can write the best computer program.

shummie
02-10-2005, 06:52 PM
[ QUOTE ]
The only thing that might keep these things from spreading like wildfire is the possibility that they'd be more profitable to their creator if he/she were to use them himself rather than sell them. They could conceivably make more money using them on their own than they could make selling them.

[/ QUOTE ]

This makes a bit of sense. The guys who are on the cutting edge of Poker AI right now are keeping a lot to themselves and making money by working on Danniel N. XBox Poker or creating Texas Hold'em slot machines. Computer Scientists though, for the most part, like to share. Think Open Source movement. AI specialists especially like to share. The problems involved with making a good poker playing AI are applicable to a wide range of AI studies. The findings of the people working on AI is going to leak into the academic community. The second generation of people who run into these papers and algorithms on poker are going to be able to make bots from them. And by this time, not everyone who is smart enough to do this will be from MIT or be profit-motivated enough to keep there work out of wide spread use.

So this is a stumbling block, but not a huge one.

- Jason

johnfromvirginia
02-10-2005, 10:45 PM
Exactly. And well put. My fear is that it won't take 5 or 10 years. Instead, there may be a grad student at MIT working on it right now who could have a program on line within a year. It's not really a very big step from PT to a "bot" program. The only real protection against this is for the sites to stop releasing hand history.

DesertCat
02-10-2005, 11:40 PM
[ QUOTE ]
The only real protection against this is for the sites to stop releasing hand history.

[/ QUOTE ]

I think you greatly underestimate the difficulty of building a competent poker bot. But assume one can and will be built. Further assume it will decimate the ranks of poker players, making it online poker a -EV for all non bot participants. When the sites start losing players, do you still think they will aid their own destruction by continuing to provide access to hand histories?

mostsmooth
02-10-2005, 11:52 PM
wait a second, are you guys saying there arent bots yet? i thought the net was swarming with them already?
/images/graemlins/cool.gif

johnfromvirginia
02-11-2005, 01:43 AM
[ QUOTE ]
do you still think they will aid their own destruction by continuing to provide access to hand histories?

[/ QUOTE ]

Eventually they probably would, but the question is when. I think once the average shmuck quits playing because he can never win, he's not going to come back because the sites announce that they have figured out why and corrected the problem. You can shear a sheep many times, but skin him only once... or something like that... A lot of people just have gambling in their blood and they're just going to keep coming back no matter what happens to them, but this great wave of popularity in on line poker may be more delicate than it appears. People will probably not continue to lose their asses every single time they log in and keep coming back.

And for the record, I think you underestimate how simple a technical poker bot would be for some people to create. Once you incorporate pot odds and implied odds into the numbers that PT and PV are already providing and figure out how make the bot call, raise or fold, you've already got an AI that is completely free of emotion and will beat most players.

tolbiny
02-11-2005, 03:12 AM
"The only real protection against this is for the sites to stop releasing hand history"

Any competent programmer could just take the action and convert it into his own HH file. You won't get all the showdowns but for the most part you will have all of the fold preflop% and postflop aggression stuff.

As for your fears- i don't doubt that there are already multiple bots making cash for their owners. But these won't destroy the games, people show time and again how much money they are willing to loose with little chance of winning at all.
Bot detection will keep up (for the most part) with the bot progression, but there will always be several that keep ahead of the curve. But are you really worried that your 5/10 shorthanded and 3/6 games are going to get crushed by bots when they could be cleaning up the 10-20, 15-30, and 20-40 games?

JoeTable
02-11-2005, 03:26 AM
[ QUOTE ]
It strikes me that poker software like poker tracker and poker view may be extremely dangerous to online poker. Eventually, these things will evolve into software that just comes right out and tells you whether to fold, call or raise. The next evolution would just be a bunch of poker bots playing tight, aggressive poker without any emotion. When this happens, the online poker industry will die within months because no one can survive in that world for long, certainly not any of the sheep. If it were up to me, hand history would not be available to players. All this software designed to observe your opponents for you may eventually kill the goose that laid the golden egg.

[/ QUOTE ]

There already is pokerbot software out there. It's been banned by most sites.

Pinga
02-11-2005, 07:56 AM
Interesting. From WinHoldEm web site:

--------
Party Poker (IGM) Advisory
DO NOT use WinHoldEm at any - IGM - skin site without using WINPP. The IGM software is scanning your system for WinHoldEm and they will suspend your account. The IGM software is taking screen shots of your entire display. The screen shot is stored in the temp folder where the files are installed. Party Poker will steal your winnings and deposits if they believe you are a WinHoldEm user even if you have not violated their license agreement. We apologize for this inconvenience. A two computer anti-detection solution is available in conjunction with WINPP.
---------

Can anyone verify this claim? I am hoping this is just bull. FWIW I monitored the app for a short while and it did not produce anything in the temp folder.

Would anyone like PP to have a screenshot of their CC, taxes, or other personal information that may be open on the screen?

How about if you downloaded WinHoldEm out of curiousity but didn't use it to play poker. Lose your cash because they don't like a binary on your hard drive?

This could be a hot topic. I need to review the TOS.

--
Pinga

Pinga
02-11-2005, 08:46 AM
Talking to myself, as usual. Here's the skinny:
----------
7. Consent to Use of Anti-Cheating Software. Our software package used to access and use the Service contains certain features designed to detect use of automated programs that enable artificial (non-human) intelligence to play on our site. Use of such "bot software programs violates our Terms and Conditions of Use, and is detrimental to other players. By installing our software and using the Service, you consent to our software performing the functions described in this section. Our software may perform any or all of the following functions in order to detect the use of illicit automated programs and ensure that we maintain a "cheat free environment for all users: (1) scan your list of active software applications while you are using the Service, (2) scan your list of active processes while you are using the service, (3) scan the files in your PartyPoker.net-related program folders to ensure that only "non-hacked versions of our software are being used. If any of the foregoing processes reveals a suspect application or process, our software may (1) scan the files associated with the suspect application or process and compile a composite mesh (i.e. a profile that characterizes the files associated with the application or process) to be matched against profiles for known illicit automated programs, and (2) capture a screenshot of your desktop, and transmit these to PartyPoker.net for review. Our software will not perform any random search of large portions of your hard drive or other files, and will not transmit any information to us or to any third party other than the information necessary to identify use of illicit automated program as described above. Our software will not alter any files or information on your computer, and will not interfere with the operation of any of your applications. You may terminate use of our software at any time by uninstalling the PartyPoker.net software package.
-----------

Semi-invasive, but understandable.

They must realize they are surfing against the tide much like the antivirus industry. Doesn't take much to hide a program. Someone could write a bot that would just sit on the ethernet between hub and PC and alter the commands passing back and forth. Or an excel macro-bot. Or...

To get more on topic:

A lot of people seem to feel that a poker bot would be inneffective. I have no proof of this, but I disagree. I have to imagine a well programmed book-bot would be profitable on PP .5/1 or 1/2.

I understand it would not beat the best of the pack, and may do poorly on 15/30, but most simple fold-call-bet decisions can be modeled. Basic strategy, optimal bluffing, etc. As the original poster proposed, it would be easy to build more advanced decision making by using PT database mining. Although how much can you base a statistical desision on 50 hands?

All of this would be significantly better than your average low limit player.

On a higher front, consider this. If I program a bot to make the decisions that I will make anyway and allow it to press the buttons for me, have I really done anything wrong? What if I sit and watch it? I would have pressed the buttons anyway. There is no difference.

What if I was paralyzed this was the only way I could resonably play? Would it be fair then?

I'm not trolling, this whole bot thing is truly an interesting dillemma if you give it some thought.
--
Pinga

P.S. I do not approve of poker bots or have anything to do with them. I lose all my cash fair and square.

P.P.S. Now I need to make sure I didn't download any bots when I was first researching online poker... delete delete delete.

drexah
02-11-2005, 10:10 AM
All i have to say is "..."

johnfromvirginia
02-11-2005, 07:40 PM
Thanks for the research Pinga, it provides a little comfort that at the very least the major sites are aware of the problem and are doing something about it. It looks like bots are unlikely to become the widespread problem that I thought they might. Although, as you point out, some of the best and the brightest are likely to give it a shot and it will always be a potential risk.

SomethingClever
02-14-2005, 05:31 PM
[ QUOTE ]
Exactly. And well put. My fear is that it won't take 5 or 10 years. Instead, there may be a grad student at MIT working on it right now who could have a program on line within a year. It's not really a very big step from PT to a "bot" program. The only real protection against this is for the sites to stop releasing hand history.

[/ QUOTE ]

You do realize that the sites themselves police this sort of thing, don't you?

Party has frozen accounts for using WinHoldem, a gay poker bot that could allegedly beat the micro tables.

There may always be a battle between hackers and the sites, but in the end, the sites have s[/i]hit-tons of money and a lot at stake to snuff out any bots.

jcmack13
02-15-2005, 09:57 PM
it's kind of like a missle defense shield, though, isn't it? people will just build a better missile. it wouldn't be too hard to write a program that logs the style of play of opponents and matches it against known pokerbot algorithms. while a live player is using such software at a table, it could flash a warning saying "player32 is likely a pokerbot" or something. and since the live player would know that player32 is literally playing like a robot, i would think that would open a lot of opportunities to take pots away from it. bots would likely be programmed to be tight and nut peddle, so 9 times out of ten a good flop or turn bet would cause a bot to fold. if you were aware you were against bots, you could probably play against them quite profitably.

of course, then they would just build a better missile. the logical evolution is more sophisticated and more complicated and more difficult to detect bot algorithms, which in turn would beget more sophisticated bot detection, and so on.

i guess a lot of this depends on what AI is capable of becoming. if a machine can ever become a better learner than a human mind, i guess pokerbots would be only a minute part of the fundamental questions superhuman AI raises.

slickpoppa
02-15-2005, 10:12 PM
Winholdem is just a board with a nail in it. But pretty soon people will come back with even bigger boards with even bigger nails.

Not too long ago the idea of a computer beating the best chess player in the world seemed distant. With the amount of money flowing into internet poker and the number of computer geniuses with time on their hands, it is inevitable that someone will write a program that can beat any limit game. The only question will be if the sites can develop adequate countermeasures.