0
$\begingroup$

I have a Python model for football matches, and I backtested it against some data. I have a dataset of hypothetical profits / losses from the model, which I've standardised by dividing by the stake. I'm interested in getting the variance of my returns. Just simply getting a standard deviation in Excel produced a percentage > 30. I'm thinking it doesn't make sense getting the variance this way, since you can expect some returns to be quite big, when betting say on 9:1 works out. I'm also thinking of getting the variance from the Binomial Poissin Distribution. Basically my aim is to test a few hypotheses about the model, but I can't seem to get a decent estimate for variance.

Thanks

$\endgroup$

1 Answer 1

0
$\begingroup$

You can get the variance of an individual bet from the binomial distribution. If the estimated chance of success if p, then variance is p*(1-p) (the maximum will be reached at p=1/2, making the variance 1/4)

Now, for big samples, variables tend to behaive like normally distributed ones, so you can estimate the variance of the sum of all bets as the sum of variances of each individual bet. In practice, you can always be sure that for n bets, if will stay below n/4 for the reason explained above.

Now take the square root and you have standard deviation: sqrt(n)/2 whjich may be a more interesting way to put it into the right units

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.