The only reason, though, that this is necessary is due to public perception.
A bunch of bots battling it out at an online poker table isn't by itself a concern for the online poker site - they get paid either way - the concern arises because the customers, the public, won't play if they perceive that bots are being used.
(Bots allow collusion and stealing from other players - not stealing from the house)
Interesting article - I've seen enough good poker players to know that people can do exceptional things like this - I don't find the story that unbelievable.
(What I find more unbelievable is that someone with that much skill didn't have bots going on somewhere....)
"The Goldman Sachs trader was observed issuing upwards of 60,000 false bids on 1000 different securities each, entirely within a 1 second timeframe, and was able to lift his hands from the console and hold a conversation with our SEC representative while working his many 'tables'..."
These artificial limitations do not exist in the real world. Let a computer do the hard stuff so that you have time to enjoy your 4.5 million dollar condo!
I had this happen back in my pokerstars days, a captcha pops up with a limited amount of time to complete it. However I was playing "only" 16 tables at the time.
I've been on 2+2 for the better part of a decade and can tell you the online poker community has far more pressing concerns than 'bots' that hope to eek out a few dollars an hour playing a large amount of tables. [1]
It's certainly an impressive feat to be able to play 120 tables at a time, but naturally the more tables you play the less attention you are able to give to what's going on. As online poker sites give you a time bank of around 30 seconds, if you are playing more than 20 or so tables you don't really have time to consider more than your hole cards, the board, and the effective chip stack (whichever person has the lower amount of chips, applicable in "no limit" games). To play 120 tables and make a profit is only possible at the absolute lowest levels, when you're making decisions in less than a few seconds, and only considering the cards in your own hand. Even then, I can't imagine doing this successfully.
A common criticism in the poker community about people who play a lot of tables (anywhere from 6-12 tables is pretty standard for most professional online players, less than 6 shows a pretty remarkable discipline actually, excluding those who play heads up poker in which 2-4 tables is more common) is that these people don't have time to consider their decisions, and therefore won't improve. If they have a set decision they do in every type of situation, they won't get any better by failing to explore the value of each of their options. Further, the nature of online poker is such that the games have been getting increasingly tough each year, and failing to improve will likely turn you into a break even player.
The problem is that almost every competitive online poker player signs up for what is called "rakeback" from sites like http://rakemonkey.com and http://raketherake.com. In short, these types of sites have deals with the poker sites that allow players to get a percentage (usually 27%-30%) of their rake back (the few dollars from each pot that the site takes as profit). So playing 16 tables at a time becomes very appealing to the guy trying to grind out a living playing the $1/$2 no limit games, as he can easily make $5,000 from rakeback alone.
My theory is that as a player's poker career progresses (and he ages) he becomes more risk averse, and so rakeback becomes that much more appealing. By the time I turned 18 I was playing some of the larger stakes games online as the money meant little to me (purely out of ignorance to the value of a dollar) so I wasn't afraid to try to climb the limits as far as I could. As little as 18 months later I had dropped down limits after a number of high stakes shots ended badly, and afterwards while grinding the mid-stakes games, I loved my rakeback.
It's worth noting that the 120-table guy plays SnG (sit-n-go) tournaments, which is generally considered the most mindless form of online poker in the sense that hardly any action takes place after the flop, so it is much easier to multi table.
So if I understand you correctly it can pay to just sit in on a table and auto-fold? From what little detail they gave in the OP it sounded like the guy must fold everything but primo.
Ehhh... that won't pay (don't you thin it would've been too easy?).
The problem is like the other person said - there are two blinds that travel around so in every cycle you get to pay 1.5 minimum bet (thus ensuring that you loose money if you don't play). As for the rakeback... you only qualify for return of the rake that you "contributed" by winning a pot. So for example, you win a pot of 1000$ and the casino takes 100$ in rake from that. After rakeback you receive 30$ of that original rake.
So many (most?) players rely on playing a break-even game and doing profit on rakeback. I read somewhere that pro's only win 7 big blinds per 100 hands (thats long term ofc). Thats little, very little.
Rake is something that people often underestimate - its the same mechanism casinos use with green pocket on roulette etc... Its used to give casino an edge. Rake ensures that if you keep playing long enough you will run out of money - now matter how well you play.
Good post, but there are a few slight inaccuracies:
> you only qualify for return of the rake that you "contributed" by winning a pot. So for example, you win a pot of 1000$ and the casino takes 100$ in rake from that. After rakeback you receive 30$ of that original rake.
1. Sites typically cap their rake at $3 per pot, so $100 is very unrealistic.
2. You qualify for return of the rake that you contributed by playing a pot, not by winning it. So if it folds to you in the blinds and you shove all-in and the big blind calls, that's two players who have put money in the pot. If they take $3 rake from that pot, you each get $1.50 contributed rake.
3. Not all sites use contributed rake. Some sites just equally divide the rake by all the players who received a hand, without regard to how much money they put in the pot.
As far as folding to profit from rakeback, let's look at a 1/2 game. If 40% of hands see a flop, and the average postflop pot is $20, we're looking at $1 rake per postflop pot for 40% of hands, so $0.40 rake per hand. Divide that 9 ways and you have about $0.045 rake paid per hand. Get 30% back, and you're only looking at 1.5 cents per hand, or 13 cents per orbit. You pay $3 per orbit in blinds, so clearly folding is not going to work to make any money.
"Rakeback pros" (a derogatory term) make money by playing near-breakeven poker for a huge volume, let's say 250,000 hands per month (200 hours per month, 1250 hands per hour (a low estimate)). Using my numbers, that comes to $3,750 each month from rakeback.
Nowadays, 7 big blinds per 100 hands is considered a very solid winrate at midstakes games (2/4 through 5/10). This is actually described as 3.5bb/100 which is read as 3.5 "big bets" per 100 hands, which is the traditional system used for limit hold'em, where a big bet is two big blinds. I remember way back in 2003/2004 when PartyPoker was the most popular site and people were posting graphs of 10bb/100 (big bets) winrates at stakes as high as 10/20 over hundreds of thousands of hands. The fact that having a winrate like that is impossible today even at stakes as low as 1/2 shows you how much tougher the games have become.
Rake ensures that if you keep playing long enough you will run out of money - now matter how well you play.
Correction. Rake ensures that if you keep playing long enough, SOMEONE will run out of money. There are many people who can play indefinitely and make money.
In hold'em variant poker games, there are typically two blinds that rotate around the table, forced bets that are similar to antes. This prevents people from folding indefinitely.
It's a market where people spend a lot of money on search engine optimization, but the best asset to starting a rakeback site is being affiliated with high-volume pros who are willing to sign up for a few sites under your rakeback site. An insider's tip on rakeback SEO- the people who are using Google to find rakeback sites generally aren't experienced players, and won't make you as much money as you might hope. (Of course the ideal situation is that you sign up a few hundred inexperienced players and after a year or two a dozen or so have developed their game to the point where they are making you more than all your original sign ups combined)
Do you mean start one from scratch or join an existing rakeback provider as a sub-affilate?
If you're starting one from scratch, the players under you will need to be putting massive volume for you to offer competive rates. For example, FullTilt will pay you 20% of the rake paid by the players signed up under you. Compare this to the existing rakeback providers who generally payout 27% for fulltilt. You're losing money here. You'll have to move up 2 tiers to having your players generate at least $25,000 in rake per month to get to a 30% payout tier, where you're making 3%, not counting overhead and acquisition costs. If you get to $50,000 per month, you'll get another bump to 35%. How many players do you think it'll take to get to that level?
The other option is to sign up with an existing rakeback provider and leverage their platform. They already have a ton of players and are able to offer top payouts across multiple sites. You'll generally get 2-5% of your players rake here, but you only have acquisition costs and the sites generally run promotions for your players like freerolls.
The second option is what most of the training sites use and that is their primary income stream, even though they're charging $25-30/mo for access. If they're able to make the players in their downline better, the players will play longer, move up in stakes and generate more rake.
In general pros don't want to waste money acquiring a player who isn't really good, deposits $50-$200, generates $0.59 in rev share for the pro, goes broke in a week and never plays again.
Well, I don't think a bot became sentient and decided to make some money on online poker. Usually there are people writing the bots, and the poker site has to make sure they aren't using them, not that the people aren't bots.
I understood that, I just found sentences like "The player was able to verbally chat with our representative during play, while continuing to play." amusing. As if him being unable to chat might have indicated failure of the Turing test.
The conversation test may have been to establish the the player wasn't blindly following instructions from a robot. Holding a conversation with the investigators would at least make it difficult to follow surreptitious advice.
The point, though, was to prove the computers weren't playing automatically and that the human was the one actually making decisions - the computers were just helping with the manual non-strategic aspects of running that many games at once.
"In the end, rs03rs03 averages 120 tables at once when at full speed, playing about 100,000 hands over the course of about 55,800 seconds. That's about 1.8 hands per second (slightly more actions, though, since some hands are multi-decision)."
Can anyone speculate what strategy might allow him to play 1.8 hands per second on average?
First of all, he was playing tournaments with a buyin of $1 maximum- likely less. So he can get away with playing a very simple strategy.
He was certainly not making standard raises, I imagine he was playing a "push-fold" strategy from the start, meaning he either went all-in or folded on his initial turn to act. His range for this strategy probably included TT, JJ, QQ, KK, AA, AK, AQ and maybe a few more hands later on in the tournaments.
Cascaded tables... A table with a pending action pops to the front, and others go immediately behind it. Various helper programs put the mouse in the right place, set the appropriate bet size, register for new tournaments, etc. Basically all he has to do is click Check, Bet/Raise, or Fold (using the keyboard or a controller), and everything else is taken care of.
This is tangential to your question, but proper strategy in Double or Nothing tournaments is to play extremely tight. The top 5/10 players get just under double their buy-in back. Winning one hand can sometimes be enough to win the entire tournament, and I've even won one without playing a single hand. The point is that he's folding probably 90% of hands preflop.
