Quote:
Originally Posted by Rikers
wp
you can model that decision at the end of your trials. So if you have n ->100 runs measure the expected payoff per unit of risk for both decisions at the end of those runs and take the better one. Obv. at one point (higher n) the 50% of 600 will surpass the 200 one.
I would just calculate the confidence interval...
Quote:
we have formulas that takes into account your bankrolls risk..*
My first thought E(x)/Var. Or simply calculate the Sortino ratio E(x)/downside deviation of your bankroll.
(my example was kind of simplistic since it voided the bankroll problem)
I think that if you are trying to assess "adequate" you have to be thinking a little bit of utility. If you drop utility from consideration, you simply calculate EV, pick the larger answer and are done.
There is no obvious reason to want to minimize variance unless you have utilitarian aspirations or are solving for minimax.
Quote:
*if you are simply stating that 100,000 is big money vs the 200/600 prizes I stated that the utility theory is disregarded....
I assumed that you were glancing out of the corner of your eye at risk of ruin (a utilitarian notion).
The 100,000 precluded that concern.