Scratch Kelly from Criterion

My first entry in the math section!

The Kelly OG Criterion!

OG for Odiousgambit, claiming coining rights for this gambling criterion! The problem for that likely is that this is actually well known and I am just re-inventing the wheel? This is for the player facing a house edge.

I find myself after a loss constantly asking myself, was it unreasonable to expect much chance for a win considering how many bets I made? Of course you go to the standard deviation math to get a handle. What occurs to me doing this, is that the more the EV approaches the value of one standard deviation, the more foolish you have to be to think you can erase the EV with a lucky session.

For example, betting "any 7" in Craps which has the double disadvantage of high HE and immediate resolution, my number crunching indicates that a session of 100 wagers puts the EV at 89% of one standard deviation. 200 wagers puts it at 150%; in other words, you would have to be luckier by a factor greater than one standard deviation to bet that much and even be able to break even. Common sense too says that doing this roughly 2 hours of betting on "any 7" you would have to be hoping for incredible luck to do any good.

Going to Blackjack with rules putting a player against a .40% HE, 560 bets [often about 8 hours] puts the EV at 8.23% of one standard deviation. 16 hours worth is 11%, 24 hours worth is 14%. Again, using common sense also as a guide, it would seem to me the danger zone is in the 10% range. In other words, as long as the EV is no more than 10% of one standard deviation, it is reasonable to think you could have a good percentage of winning sessions? Perhaps, though, it should be about 8%?

Anyway, no doubt this is old ground and somebody already can point to something.

Unfortunately, with my abilities it is possible too that I screwed up the math. I was using what I found at this WoO webpage.

edit: meh...