It sounds like Geometry's using a sort of modified Kelly system.
http://en.wikipedia.org/wiki/Kelly_criterion
It can be applied to sports wagers, horse racing, and probably to casino blackjack if you're counting cards.
If you assume 1 to 1 odds on your money (most pool bets) and you are absolutely certain you have an edge, you find out how much to bet by taking your odds of winning minus your odds of losing.
Like if you're absolutely positive you're at least 10% favorite to win a match, the formula would simply be 60 - 40 (multiplied by 1-to-1 odds on the money). The result is 20. Therefore you should bet 20% of your bankroll.
Over the years gamblers have figure out the flaw in this system... even if you're correct on your estimates, an unlucky streak can send you home broke. It's fine in the long haul but you also want to pay attention to the short term. You may not get a "long haul" with some of the opponents you have successfully clocked, they may quit and never be seen again.
So guys like jdx, if they used a system like this at all, would modify it... like they might bet 15%-20% less than what the system suggests.
Geometry's estimates are conservative too. They give 10% less with a lower edge and 20% less with a big 80-90% edge. If you truly have 100% odds of winning then you should bet it all. This is all assuming you're using your disposable income, you could lose every dime and still afford to live.
The problem in pool is there is an element of luck that never would quite guarantee 100%, and between luck and the difficulty of clocking people's speeds (especially those who are willing to lay down) it'd be hard to ever get an accurate estimate of your odds of winning. Even if I was positive I had a 10% edge, I personally wouldn't bet more than 5% and I'd want the other guy to commit to several sets. It's still uncomfortably close to a coin flip.
http://en.wikipedia.org/wiki/Kelly_criterion
It can be applied to sports wagers, horse racing, and probably to casino blackjack if you're counting cards.
If you assume 1 to 1 odds on your money (most pool bets) and you are absolutely certain you have an edge, you find out how much to bet by taking your odds of winning minus your odds of losing.
Like if you're absolutely positive you're at least 10% favorite to win a match, the formula would simply be 60 - 40 (multiplied by 1-to-1 odds on the money). The result is 20. Therefore you should bet 20% of your bankroll.
Over the years gamblers have figure out the flaw in this system... even if you're correct on your estimates, an unlucky streak can send you home broke. It's fine in the long haul but you also want to pay attention to the short term. You may not get a "long haul" with some of the opponents you have successfully clocked, they may quit and never be seen again.
So guys like jdx, if they used a system like this at all, would modify it... like they might bet 15%-20% less than what the system suggests.
Geometry's estimates are conservative too. They give 10% less with a lower edge and 20% less with a big 80-90% edge. If you truly have 100% odds of winning then you should bet it all. This is all assuming you're using your disposable income, you could lose every dime and still afford to live.
The problem in pool is there is an element of luck that never would quite guarantee 100%, and between luck and the difficulty of clocking people's speeds (especially those who are willing to lay down) it'd be hard to ever get an accurate estimate of your odds of winning. Even if I was positive I had a 10% edge, I personally wouldn't bet more than 5% and I'd want the other guy to commit to several sets. It's still uncomfortably close to a coin flip.