Kelly equation takes as arguments overall advantage, lose probability and payout ratio.
The fact that we don't know them exactly means we work approximately.
I insist Don.
Sure, but what Don is trying to point out is that you nee the probabilities of winning/losing/pushing each possible amount, not just the probabilities of a net win/loss/push. In blackjack with insurance and surrender, DAS and SPL3 you can lose up to 8.5 times your initial bet and you can win up to 8 times your original bet, all in 1/2 bet increments. Many other casino games have multiple payouts beyond just winning or losing. For games like this, the Kelly formula is extremely difficult to solve and we're forced to use an approximation like EV / Variance.
You can keep raising your bet as much as you like. Stopping at any particular TC is simply because a) in shoe games, if you wait to make a max bet at any higher value, the frequency is so slow that you're just wasting your time, and b) if you permit yourself to bet strict Kelly, with no upper limit, you're eventually going to get thrown out of every casino in the country.
Don
No, just the opposite. As true counts go much higher, variance actually decreases because of the increasing number of 20-20 pushes. Just look at any of the chapter 10 charts to see this. I'm assuming that you simply mean the variance of results of a hand at TC >= +6. If, on the other hand (no pun intended!), you're speaking about your own monetary variance from increasing your bets any higher, beyond TC of +6, that's an entirely different matter. Obviously, the more you bet, the greater the variance. I'm assuming you didn't mean that, but you never know!
Don
When I divide EV I find in the calculator http://www.bjstrat.net/cgi-bin/cdca.cgi
with the variance 1.31 you told me I get extremely low percentage of the bankroll. Am I doing something wrong? (8deck shoe)
Don, the problem with your literal interpretation is that OP, or anyone for that matter will never win 2.5% more hands than they lose. I believe it’s possible to win more hands than lose only at very high counts exceeding TC5. Ties are ties and aren’t entered into this thought process excercise.
Bookmarks