Loismustdie 0 Posted July 30, 2005 Share Posted July 30, 2005 the point of the 500 hands is that i watch you showdown 500 hands to setup a distribution of the hands you play. And it has nothing to do with "you always raise on these or never raise on those" it's called it probability distribution because it gives a probability for each hand possible based on how you act, and though we may change up gears somewhat we don't change all that much in the long run. Also, learn some statistics before you discount 500 hands as an accurate sample size. All I have to do is do things differently than I did in 5oo hands to beat your AI- I wouldn't have to change to much either. Alittle luck comes into play and your AI is bent over the kitchen sink. Link to post Share on other sites
nrs02004 0 Posted July 30, 2005 Share Posted July 30, 2005 every hand it watches is incorporated into the AIs database. So you could probably do pretty well against it if you knew you were playing an AI, but it is just gonna look like a regular player. I think that while you usually vary up your game a bit, limit holdem at a full table is much more about folding when you're beat and getting more money in the pot when you're ahead than trying to confuse you're opponent (not that you don't try to confuse your opponent, but it's really not the top concern), so you won't vary up your game as much as you think... Link to post Share on other sites
elkang 0 Posted July 30, 2005 Share Posted July 30, 2005 It's true it doesn't take much to win. Simple basic play can win money online at many levels. But, publically available bots are closer than I thought (LA Times registration required). Following a player assisting model - there's no defense except those employed by online Chess and Scrabble players.Chess and Scrabble are both far more solvable than poker and played only heads up.A bot that can play well *heads up* isn't difficult at all. Linking a story about a new bot that barely beat a 10 year old bot (Poki) that humans (including me) can just obliterate over the long term is really fairly meaningless.This is PR for Laak, period.In many ways it is easier to design a bot to play standard ring games than heads up play. I could not say what would be more difficult - as they require different sets of rules. At the very least a bot can potentially increase the skill level and win rate of 90% of the players out there. Probably a lot more likely if the bot can take over the human who will undoubtably tilt. In any case, my point about chess and scrabble is in regards to bot policing which is still many steps away -- we need to have a significant bot presence first.Of course, really all this talk is meaningless without 10K of hand histories. An interesting distraction for me, but if anyone has the time to try it out - I'd be glad to see the results.I agree that the article is a PR piece for Laak, because he is Phil Laak and Jenn Tilly's bf - and the media seem to love covering him since he has good representation. Being such an article (like all mainstream poker articles are) doesn't affect the growing bot scene, only helps expose the progress of bot development to more of the public. It is closer than I had thought (if they are giving out $100K prizes and getting press coverage) but that may just mean that I misjudged things. My time line for delivery of profitable (auto)bots now is under 3 years. I had put it at 5 before. But, what the hell do I know. Link to post Share on other sites
alf13 0 Posted July 30, 2005 Share Posted July 30, 2005 I am surprised that a "good" bot doesn't exist already.....To be succesful...it would have to do this....A) be able to create a database of players styles and stats to be able to adjust to each individual player accordingly.B) Be able to recognize players from it's memory bank instantly and adjust to them. C) Play a perfect tight aggressive style.I think if this bot existed (it is very complex...it would have to problem solve and almost think for itself in adjusting to others play) it could be set loose on a site and you would have a cash machine.I do believe that nowhere in the near future will a bot "KILL" a game....but a bot like I described that can make 25k a year playing 5/10 limit? Why not? Link to post Share on other sites
RonnyMemphis 0 Posted July 30, 2005 Share Posted July 30, 2005 I lost my arm in an unfortunate accident when I was in the service jumping out of helicopters. I was medically discharged and sent home. I went to the local VA, because I was promised a new state of the art prosthetic limb.It was tough getting used to, but it had some interesting assistive software that made getting used to the limb a little less difficult. It was about this time, while looking for work that I discovered online poker.I signed up, not knowing a thing about bankrolls or even poker and made my first $50.00 deposit. It wasn't long before, out of curiosity, I attempted to play with my prosthetic arm to see if I could do it. I was baffled at first when a number of times, when I held only two cards and hadn't even seen the flop and was about to call, the arm seemed to malfunction and ended up folding the hand.A number of these "misclicks" began to occur more frequently, and I got in touch with my VA. The technician informed me that there was no visible defect with the limb, and as far as he knew, everything was working properly.I got back home and attempted to give it another shot. The same thing kept happening. I noticed sometimes it would call when I wanted it to fold. But more and more, I realized as I studied poker, that the times it mis-clicked were helping me. The arm was calling bluffs. The arm was making smart plays. Me and the arm moved up the limits like wildfire. The arm plays perfect poker, and I've learned to give in to it's decisions.So to answer your question; is there anyone who has created software sophistocated enough to wipe out online poker from the top down? I'd say the answer is a resounding yes.Watch out for me at the 100/200 tables friends. Link to post Share on other sites
princeof56k 0 Posted July 30, 2005 Share Posted July 30, 2005 There is plenty of life left in online poker. What most people fail to realize is that the majority of players dont care about the odds. The most popular games at live casinos are the slots and table games (balckjack, craps, roulette). All the games are basically rigged to where players lose in the long run, but yet people still play them. Trust me, if your typical bad poker player knew he was up against 9 superbots, that wouldnt stop him one bit.However, bots could eventually (and not anytime soon) start eating into the profits of the good online players. This is not because of the bots playing at the higher limits (I think most good players could beat the bots), but because of the bots ripple effect at the lower limits. A good portion of the profit at the mid-level games comes from mediocre players beating the low-level games. This is because at the low-level games the play is so bad that these guys can beat the game. When they jump levels, they can get crushed by good players, and then they reload and continue to play at that level because they ignorantly chalk up their beating to variance. Well bots will probably be able to beat these mediocre players at the low-limit and basically cutoff some of the food supply to the sharks at the higher levels. So the players left beating the low-limit games and moving up will be a little better than what we have now with no bots. BTW, the good players will still make money as there are always bad players with a lot of cash that will just start out the higher limits. I just think they wont make as much money as they are today. Link to post Share on other sites
BetItAll33 0 Posted July 30, 2005 Share Posted July 30, 2005 the point of the 500 hands is that i watch you showdown 500 hands to setup a distribution of the hands you play. Â And it has nothing to do with "you always raise on these or never raise on those" it's called it probability distribution because it gives a probability for each hand possible based on how you act, and though we may change up gears somewhat we don't change all that much in the long run. Â Also, learn some statistics before you discount 500 hands as an accurate sample size. All I have to do is do things differently than I did in 5oo hands to beat your AI- I wouldn't have to change to much either. Alittle luck comes into play and your AI is bent over the kitchen sink.This is a pretty presumptuous statement. As was said before, you can only vary your game so much in a limit ring game. There is generally a correct decision and an incorrect decision, with a little room for creativity. The bot probably won't expect you to raise with 72 offsuit, but that won't make it profitable either. Link to post Share on other sites
BetItAll33 0 Posted July 30, 2005 Share Posted July 30, 2005 I am surprised that a "good" bot doesn't exist already.....To be succesful...it would have to do this....A) be able to create a database of players styles and stats to be able to adjust to each individual player accordingly.Why not just interface with the pokertracker database? It an access database... easy enough.B) Be able to recognize players from it's memory bank instantly and adjust to them. See part A. This can be done' date=' but it'll take a hell of a lot of code... you probably have to ignore data for players with less than 50 or so hands. Also, I'm not completely sure that this would be a requirement for a successful bot at lower limits. Players are constantly moving tables and it would make the bot exponentially more difficult to program.[/color']C) Play a perfect tight aggressive style.Obviously, this is the easy part (preflop, at least). 169 possible starting hands, 10 different table positions, a few external factors (facing a raise, etc.) and you have yourself a nice preflop matrix. This alone could almost break even at party poker $.50/1I think if this bot existed (it is very complex...it would have to problem solve and almost think for itself in adjusting to others play) it could be set loose on a site and you would have a cash machine.I think you could make a profitable bot with minimal oppenent profiling at the lowest limits.I do believe that nowhere in the near future will a bot "KILL" a game....but a bot like I described that can make 25k a year playing 5/10 limit? Why notWhy not play 15 tables of micro limits at multiple sites? Much easier to beat, and it is a bot after all. Link to post Share on other sites
dabent 0 Posted July 31, 2005 Share Posted July 31, 2005 Artificial intelligence is no match for natural stupidity Link to post Share on other sites
Paul_Castes 0 Posted July 31, 2005 Share Posted July 31, 2005 Seems like people fail to recognize the power of ADDICTION in holding poker's player base.I'd be optimistic.You said it best. Poker's BASE. While they'll remain, the thousands and thousands of weekend warriors will go away and we'll be left with poker as we knew it 15 years ago.I fully agree with you.Have a Good Luck Link to post Share on other sites
EVILGENIUS 0 Posted July 31, 2005 Share Posted July 31, 2005 I actually prefer live poker over online(i know the advantages of online) plus i live in fl where in B&M we can only play 1/2 or straight 2. I have learned to trust my reads lately(used to call even when I was almost sure I was beat with a good hand- 2 pair/trips when straight possiblity or flush) and except for huge pots I will lay down based on my reads(only when very sure). About the same time i started this my profits went noticably up. (well duh). Online it takes away from the feel for me and is less enjoyable not being able to use my reading abilities.Online poker will always be around in some form, but live poker if at a reputable location will never go away. And I'd rather hone my skills there than online because of a more certain future of being able to continue those skills(plus its more fun to crush your opponents where you can see them cry).I play for the competition more than the money...I like the money though... Link to post Share on other sites
alf13 0 Posted August 1, 2005 Share Posted August 1, 2005 Sadly it wasn't really that unforseen. Bush had a report sitting on his desk saying that an attack was imminent. His laziness and stupidity got in the way of something being done.Luckily the brilliant American public went and re-elected himToo bad that NONE of that is true. :roll: Link to post Share on other sites
blueodum 0 Posted August 1, 2005 Share Posted August 1, 2005 "How do you think that would affect your win rate?"Bluff more. Then when the computer figures out that you are bluffing more, bluff less. In order for the program to make the correct decision, it has to compute its odds of winning a hand at every stage. If it doesn't know what the human opponent has, because that human opponent behaves unpredictably, it will have a large margin of error in its computation. On close decisions, it will therefore make mistakes. If it makes more mistakes than the human opponent over the long term, it will lose. Link to post Share on other sites
blueodum 0 Posted August 1, 2005 Share Posted August 1, 2005 "Maybe I'm oversimplifying things, but as a programmer myself, I don't think it would be difficult to write a bot to beat limit games up to 2/4. After that, it gets trickier."I agree with this. Games at that level are relatively easy to beat because they are filled will inexperienced and unsophisticated players. The point is, a decent player can ALREADY crush these games - he/she doesn't need a bot. Even if 20% of online players were using bots (and we are years away from the possibility of this), there would still be plenty of weak players for the human sharks to feed on. Link to post Share on other sites
blueodum 0 Posted August 1, 2005 Share Posted August 1, 2005 "the point of the 500 hands is that i watch you showdown 500 hands to setup a distribution of the hands you play." I've never shown down more than 150 hands at any table during any session, ever. Rarely does that number exceed 50. So you are going to sit there and catalogue every hand someone plays online, just on the offchance that you'll get a chance to sit at a table with them and have your bot play them?Sounds impractical to me.Furthermore, I play differently based on how my opponents are playing. The mere fact that your bot is there will alter my decisions. I don't think you'll be able to get a handle of the way I play, unless you (you the human I mean) study the way I play over many weeks.Obviously this is completely impractical, since there are millions of players out there and any 9 of them could sit down with you at the table and leave any time they want. Link to post Share on other sites
Mojo_Yugen 0 Posted August 1, 2005 Share Posted August 1, 2005 " So you are going to sit there and catalogue every hand someone plays online, just on the offchance that you'll  get a chance to sit at a table with them and have your bot play them?...Obviously this is completely impractical, since there are millions of players out there and any 9 of them could sit down with you at the table and leave any time they want.Actually there are already subscribtion services that do this for you already. Your 'bot doesn't need to collect the data, just access it. Link to post Share on other sites
Loismustdie 0 Posted August 1, 2005 Share Posted August 1, 2005 the point of the 500 hands is that i watch you showdown 500 hands to setup a distribution of the hands you play. Â And it has nothing to do with "you always raise on these or never raise on those" it's called it probability distribution because it gives a probability for each hand possible based on how you act, and though we may change up gears somewhat we don't change all that much in the long run. Â Also, learn some statistics before you discount 500 hands as an accurate sample size. All I have to do is do things differently than I did in 5oo hands to beat your AI- I wouldn't have to change to much either. Alittle luck comes into play and your AI is bent over the kitchen sink.This is a pretty presumptuous statement. As was said before, you can only vary your game so much in a limit ring game. There is generally a correct decision and an incorrect decision, with a little room for creativity. The bot probably won't expect you to raise with 72 offsuit, but that won't make it profitable either. I generally don't play limit. I'm talking no limit- good luck putting me on a hand. I would whoop your bots ass. As far as limit goes, I generally don't play heads up. So, in a 10 seated game with one bot having to read 9 humans? Impossible. How is the bot going to know that I just spanked it and am a little euphoric and feel like playing god knows what? And the guy next to me is going through a divorce- he's just angry, and he's gonna ram and jam everything, it doesn't matter what he has. Do you see what I am saying? A robot cannot comprehend emotion, and like it or not it plays into poker on some level. Could it sit there on play only pocketpairs tens or better and certain suited connectors? Sure, and I am sure that it would be profitable if only because you could plug it in 24/7, and it doesn't need a break. If it wins 52% of the time you are still making money. Would I be afraid to play it? HELL,NO. I would just assume I am playing a rock and go to work. Link to post Share on other sites
Loismustdie 0 Posted August 1, 2005 Share Posted August 1, 2005 Sadly it wasn't really that unforseen. Bush had a report sitting on his desk saying that an attack was imminent. His laziness and stupidity got in the way of something being done.Luckily the brilliant American public went and re-elected himToo bad that NONE of that is true. :roll: Actually that is true. The administration and C.I.A. had known Osama was a threat for awhile, and there were specifics. They also were well aware of what he is capable of- hell, they trained him, and then basically cut him lose on society, but I can tell from your posts you are a smart guy, so you already knew this. Why perfectly sane intelligent people continue to look the other way is beyond me. :shock: Link to post Share on other sites
BetItAll33 0 Posted August 1, 2005 Share Posted August 1, 2005 "How do you think that would affect your win rate?"Bluff more. Then when the computer figures out that you are bluffing more, bluff less. In order for the program to make the correct decision, it has to compute its odds of winning a hand at every stage. If it doesn't know what the human opponent has, because that human opponent behaves unpredictably, it will have a large margin of error in its computation. On close decisions, it will therefore make mistakes. If it makes more mistakes than the human opponent over the long term, it will lose.More often than not these bluffs would be chip spewing. Remember, this is limit hold 'em. If you're raising the turn with middle pair, you're probably just getting called down and costing yourself an extra bet. Clearly bluffing is a necessary part of being a good limit player, but it's not that important in the big picture, especially at low limits. Limit ring games at these levels don't require a change of pace, and more often than not, your raise won't induce a mistake, it will BE the mistake. If raising is correct, then you're not bluffing. Link to post Share on other sites
BetItAll33 0 Posted August 1, 2005 Share Posted August 1, 2005 the point of the 500 hands is that i watch you showdown 500 hands to setup a distribution of the hands you play. Â And it has nothing to do with "you always raise on these or never raise on those" it's called it probability distribution because it gives a probability for each hand possible based on how you act, and though we may change up gears somewhat we don't change all that much in the long run. Â Also, learn some statistics before you discount 500 hands as an accurate sample size. All I have to do is do things differently than I did in 5oo hands to beat your AI- I wouldn't have to change to much either. Alittle luck comes into play and your AI is bent over the kitchen sink.This is a pretty presumptuous statement. As was said before, you can only vary your game so much in a limit ring game. There is generally a correct decision and an incorrect decision, with a little room for creativity. The bot probably won't expect you to raise with 72 offsuit, but that won't make it profitable either. I generally don't play limit. I'm talking no limit- good luck putting me on a hand. I would whoop your bots ass.This is the main reason why you don't understand what I'm telling you. Limit is a very mechanical game. As far as limit goes, I generally don't play heads up. So, in a 10 seated game with one bot having to read 9 humans? Impossible. How is the bot going to know that I just spanked it and am a little euphoric and feel like playing god knows what? And the guy next to me is going through a divorce- he's just angry, and he's gonna ram and jam everything, it doesn't matter what he has. Do you see what I am saying? A robot cannot comprehend emotion, and like it or not it plays into poker on some level.A bot is infinitely better at recording peoples actions than a human would be. You think that recording and analyzing the actions of 9 people would be impossible for a computer? I hope you're kidding. Your emotions don't matter. If you play hands that aren't profitable in limit hold 'em, you will not profit. If you "just spanked it" and decide to play 74 offsuit, you will be entering the pot with the worst of it against any bot. If some angry divorcee wants to cram the pot with nothing, he might get the bot to slow down, but he's probably going to show down the worst hand. Don't you see, these are exactly the players that the bot feeds on? Could it sit there on play only pocketpairs tens or better and certain suited connectors? Sure, and I am sure that it would be profitable if only because you could plug it in 24/7, and it doesn't need a break. If it wins 52% of the time you are still making money. Would I be afraid to play it? HELL,NO. I would just assume I am playing a rock and go to work.No, it wouldn't just play pocket pairs 10 or better. It would play a preflop strategy probably based on Sklansky's starting requirements or something similar. Suited connectors are very marginal hands in limit. I think that your argument might hold a little more water in no limit... but limit - not so much. Play about 10,000 limit hands and get back to me. Link to post Share on other sites
BetItAll33 0 Posted August 1, 2005 Share Posted August 1, 2005 Games at that level are relatively easy to beat because they are filled will inexperienced and unsophisticated players. The point is, a decent player can ALREADY crush these games - he/she doesn't need a bot. Â Yes, but the bot can put in 12 hours days while I'm on vacation. Â I enjoy poker as much as the rest of you, but if you just want to hand me the money instead, that sounds better. Link to post Share on other sites
BetItAll33 0 Posted August 1, 2005 Share Posted August 1, 2005 "the point of the 500 hands is that i watch you showdown 500 hands to setup a distribution of the hands you play."  I've never shown down more than 150 hands at any table during any session, ever. Rarely does that number exceed 50. So you are going to sit there and catalogue every hand someone plays online, just on the offchance that you'll  get a chance to sit at a table with them and have your bot play them?Sounds impractical to me.Sounds like you're not a limit player.  After 50-100 hands I have a read on every player at the table using pokertracker.  It sounds like your defense against being predictable is to switch tables.  If that's the case, then how can you ever get a read on anyone yourself?  I don't see what's impractical about logging every hand.  It's a computer for god's sake.  It's not like it's going to get tired.Furthermore, I play differently based on how my opponents are playing. The mere fact that your bot is there will alter my decisions. I don't think you'll be able to get a handle of the way I play, unless you (you the human I mean) study the way I play over many weeks.If you're getting "creative" at $1 limit poker, you're probably losing money.  That's just the way it is.  If you're going to be "tricky" and play small suited connectors, you will be getting the worst of it most of the time.  Of course since you won't be sticking around for more than 50 hands, you might get lucky and win the one pot you play against the bot.  Don't spend it all in one place.Obviously this is completely impractical, since there are millions of players out there and any 9 of them could sit down with you at the table and leave any time they want.The 9 people that sit down with me are the 9 people at my table.  Who cares about the millions?  Tracking these 9 makes perfect sense.  Link to post Share on other sites
Canada 0 Posted August 1, 2005 Share Posted August 1, 2005 No....I enjoy playing online....but it is not the same as a live game.Online Poker= Sex with a blow up doll.Live Poker= Sex with a real womanAlf enjoys playing online poker.Online Poker= Sex with a blow up doll.=> Alf enjoys sex with a blow up doll.I always wondered who bought those things :? Link to post Share on other sites
blueodum 0 Posted August 1, 2005 Share Posted August 1, 2005 I think you're missing the big picture here. At most normal low limit tables, there are a mix of players. Against the good players (presumably including your bot) I'm going to bluff at opportune times to make you think I have a big hand. I know when and, more importantly, WHO to do it against and who not to. Experience has taught me this. I guarantee you that if I've been at your table for less than 30 minutes, you will have no clue that I don't have a power hand. Therefore your bot will fold a better hand.If your bot thinks I'm bluffing a lot, it will begin to call-down more with inferior hands. Thereby losing money to me.I don't bluff the calling stations or the maniacs, I value bet them to death. My play is governed by who is at my table. I don't ever play against a table full of skilled players because I can't make significant cash even if I outplay them. Link to post Share on other sites
BetItAll33 0 Posted August 1, 2005 Share Posted August 1, 2005 I think you're missing the big picture here. At most normal low limit tables, there are a mix of players. Against the good players (presumably including your bot) I'm going to bluff at opportune times to make you think I have a big hand. I know when and, more importantly, WHO to do it against and who not to. Experience has taught me this. I guarantee you that if I've been at your table for less than 30 minutes, you will have no clue that I don't have a power hand. Therefore your bot will fold a better hand.If your bot thinks I'm bluffing a lot, it will begin to call-down more with inferior hands. Thereby losing money to me.I don't bluff the calling stations or the maniacs, I value bet them to death. My play is governed by who is at my table. I don't ever play against a table full of skilled players because I can't make significant cash even if I outplay them.Here is the "big picture" as it pertains to the original topic:1. You will not know who the bots are. You cannot single them out.2. There are not that many opportunities to bluff significant pots and win without showdown in low limit hold 'em. More often than not, you will be spewing chips directly into the bot or other opponents just like the rest of the terrible players.3. Everything you've said is strategy that all good players are using. You're not special. Good players adapt. A good bot would too.I guarantee you that if I've been at your table for less than 30 minutes, you will have no clue that I don't have a power hand. Therefore your bot will fold a better hand.So you're going to raise marginal hands to get this bot to fold at a ten handed table. That should work great for you. What do you think the result of these raises will be after say, 100k hands? Link to post Share on other sites
Recommended Posts
Create an account or sign in to comment
You need to be a member in order to leave a comment
Create an account
Sign up for a new account in our community. It's easy!
Register a new accountSign in
Already have an account? Sign in here.
Sign In Now