I received the email below from a software engineer who has experience working in the online poker industry. He asked that I not include his name but I can attest for the fact that he’s a solid engineer who’s skills I respect.
I found his take very interesting because unlike most people who claim that rigging a game would be trivial, this person has been very involved in the industry and knows a little something about the problems someone might encounter if they actually had to implement a rigged system.
As I’ve said previously it is very easy to claim that rigging a game is easy however it’s an entirely different thing when you have to sit down and figure out solutions to questions such as the one this software engineer raises. In fact, he raises only one of many, many such issues that one would need to face if they were to actually implement such a scheme.
In reply to your most recent post, “Online Poker Isn’t Rigged … Again!” What most people don’t realize is that there isn’t a proper heuristic to determine whether the players involved in a hand should be dealt using a rigged deck. Let’s ignore how the poker site is going to rig the cards for a minute and instead focus on how to determine whether the current hand should be rigged.
With Google, they have some complicated heuristic algorithm that determines why one page scores better than the other one. With Anti-Virus Software, they use a heuristic algorithm that determines whether a program contains elements similar to a virus. But how does one do the same with determining whether a poker hand should be rigged? Should it be the number of people involved in the hand or the amount of money they have in their account? Maybe by frequency of deposits to the site or whether the player recently doubled his stack? There are a lot of factors for online poker sites to use so if they were going to actually rig the hands they would want the most optimal heuristic they could find to maximize their cheating profits.
But let’s say that the owner of the site is a greedy bastard who doesn’t really care who gets dealt a rigged hand. “AS LONG AS I GET AN EXTRA $10K IN RAKE A DAY I DON’T CARE!” Then he now needs the players in the hand to follow along with the way the rigged hand is setup. So in a situation such as Aces vs Kings, most of the time all the money will go in preflop. But maybe because both players in the hand have deepstacks and one of the players knows the other player’s tendencies so well he is able to get away from Kings preflop for 300 big blinds.
Even if the site owner decides to rig Aces vs Kings hands that’s not that much more money per day that the site is getting in rake from the player base to justify the risk. The owner then has to decide whether to increase the frequency of AA vs KK situations or start rigging other hands and situations. Now the rigged deck AI must be able to evaluate whether two players involved in a hand will get their money into the middle. How is the AI suppose to determine whether the players will generate the maximum amount of rake in that hand and how will the AI know how the hand action will go such as whether a player will raise/call/fold a flop? Now the owner needs the AI to determine whether or not these players have any “history” versus each other and tell the AI to act accordingly. But then once an AI can detect those type of tendencies one may as well NOT rig the game and concentrate on a bot that will be able to play NLH completely under the radar.
So for me as a programmer, I can decide to go through all that trouble to make some extra money while risking my reputation or I can just increase the max rake on the site by 100%. Through the powers of laziness I have just increased my daily gross without having to lift a finger.
I think you are making it sound more difficult than it actually may be. It only takes a few hundred hands to determine a players tendencies. Using this information it’s possible to manipulate the game to achieve a more favourable outcome on avg. Also, rigging something like AA vs. KK is too obvious. Making it so that strong players (who start off with a equity edge against the fish and play more top pair hands) lose just a little bit more so that the fish are happy to play and the strong players still barely break even instead of completely own the game is a more likely scenario. All you have to do is use the information from a few hundred hands and place players into standard player tendency categories. Then you give the fish a hand that will draw good and beat the strong player a little more often than it would, shifting a bit of the strong players profits to the weak players, making the weak lose a bit less but happy to play and the strong players barely profitable but still playing. I’ve played just over 50k hands of FL .25/.50 at different sites, 12k at ultimate bet. My won money at showdown at UB has been stuck at 49.5% after the first couple days I ‘ve played there and it doesn’t seem to be changing. Compare this to 54-57% for all other sites. UB is just unusual and I’m staying away. I recommend others stay away as well.
Great post, I agree. Just searched on Google about this and found your blog… Check mine if you want, I log all my Poker Session there.
How difficult could it be for a site to rank players by their winnings? Think they now very well who is a losing player and who is a winning player. How difficult could it be to deal a second from time to time when the card that is supposed to come of doesn’t help the losing player?