+ Register

[Bally6354]

http://www.albionresearch.com/kelly/default.php

cheers

[Bally6354]

This calculator is ok.

But remember.......

Underbetting is far preferred to overbetting. So perhaps use a 1/2 Kelly and see how the variance goes.

There is some good info at the bottom of the calculator.

[Bally6354]

I was just thinking about the Kelly criterion in relation to roulette and does it not just show what a lethal combination progressions and the house edge really are. The Risk of ruin getting larger with every losing step.

It's probably not something a lot of us give any thought to when designing methods of play. But just a basic understanding of some of this stuff does make you look at things from a different perspective.

cheers

[Priyanka]

I thought if we use Kelly criterion, it will just advice not to place any bets on roulette. Is it not true?

[Bayes]

Yanks,


That's correct. If you assume that roulette is a negative expectation game, then Kelly will recommend no bet. Alternatively, you can take a purely empirical approach, so make no assumptions (forget that you know about the negative expectation), then calculate the probability and expectation from your data, and calculate your edge from that.