Quote:
Originally Posted by toltec444
1. The rigged software makes the favorite hand win less frequent than it would be expected. But note (this is very important) the favorite hand stills wins most of it, indeed almost in the right frequency.
That leads tothe conclusion that even if that software is running the better player will win money in the long run and the bad player will lose.
Note one important point, there is no sense in simply revert the equity of two hands playing HU. Theres no sense in making AA win 20% against 88. That would give no extra benefit for the poker room.
The ideal scenario for a poker room is that every hand played had 50% of equity, that would generate the maximum amount of rake possible.
Let's assume your general premise is true, that leveling out the equity leads to maximum profit by keeping the money circulating and thus all players' bankroll lasts longer, they play more hands, pay more rake. For the sake of argument we'll let that one go for a moment. I could argue this assumption isn't correct, since if no one can win you have no regulars, who pay most of the rake. I could also argue that losers losing faster and depositing more money more often might come out better. Or any other number of other permutations in this incredibly complex dynamic system of the money float on the site (which you think you understand). But as you suggest we'll say they just do it a little bit, making winners win a little less, and we'll assume your premise.
Further, let's say that to keep things not easily detectable, that the variance from expectation of a particular hand matchup type needs to be less than 1%, as you said, "almost in the right frequency". So if a particular pair-over-pair is supposed to win 81%, we'll only let it win 80.5%, so that even in a huge sample no one would be very suspicious. Moving the outcome much more than that would be a very obvious outlier statistically (exceeding 3-4SD) given a large enough sample, and the undetectable argument would fall apart. I'm giving you some leeway here, because even 0.5% shift is readily detectable if it is across the board and constant. It would be like a big sign saying "rigged".
Now, we also know that the average rake on a NL100 9-player table is about $.07/player/hand (using FT scale) and in an hour's time the average rake earned at NL100 is a little under $5 per constant player (you can calculate these yourself if you like). And let's assume that NL100 is the most common stakes played online across all cash games, or representative of a mean rake. Rake is capped as stakes go up, and FT never takes more than .33/player/hand at any stake NLHE 9-player table, since it is capped at $3.00 per hand. So we'll use .07 as our mean. So far we should be close enough to agree.
So from this foundation, please explain the math for how a half percent shift in equity across the board (all hands win 0.5% less than they should) is going to increase that rake by an amount that is meaningful to the site. Show us the business case for doing this. Keep in mind that the top two sites both probably make over $1 billion a year in rake. So to be meaningful and worth the risk of exposure, let's say that you have to increase that by at least 10%. I think anything less isn't worth considering in such a high margin business. If you as an owner are making $100 million a year now, would you risk losing it all plus jail for another $10 million? Highly doubtful, but to make it easy let's just go with at least a 10% increase necessary.
Now show me how it works. Show me the math. Show me how this small shift in equity is going to extend bankrolls and change player behavior enough to add 10% to the rake. Please explain this to me.
I'm guessing that you can't, because it is going to add up to a trivial amount of profit (if any at all) for the risk taken. That's why equity manipulation will always be detectable. To be worthwhile it will be visible.
Edit - by the way, what's your theory for how rigging tournaments (fixed fee up front) increases profits?
Last edited by spadebidder; 05-27-2009 at 06:43 PM.