Hello. A review of the forum posts in the past didn't yield anything.

Is there a way to test for normality? Let's say I'm looking for stocks with the most normal/Guassian distribution of returns over, say, the last 250 bars. Has anyone tried anything like this? 

Many thanks, plus apologies.

There are quite a few different possible tests for normality in the statistical sense. Do you have a particular test in mind?

On a very related note, stock data is not normal in the statistical sense. Short term behavior which might make the data seem normal and meet an arbitrary measure of normality will not actually make it normal.

I suspect you already know this if you are asking the question, but I wanted to clarify this to both you anybody else who might read the topic later just to make sure.

Thanks, Bruce. I am familiar with how stock data in particular is typically both nonormal as a characteristic and also the normality assumption / sample size relationsip. I'm also generally aware of how often Gaussian models are described as inadequate for describing significance in market returns. I think lognormal would work too, but I am not after that yet.

I'm trying something atypical by seeking a way to measure normality, even in non-lognormal ways. I think possibly some tickers actually might have Gaussian distributions even when the general returns for the universe typically are not. I don't know what the reason for this is, but I have some thoughts, and I do think that reason is causal. Volatility as a factor is more normally distributed than return prices.

My hope in forcing an normal-distribution evaluation on all tickers is that it would make those with that hidden characteristic filter to the top where I could do some additional testing on them. 

Before trying anything, a question: is it possible to use whatever normality test for evaluating volatility itself and not returns? 

I am definitely not going to be able to use an ordered list to calculated normality in any reasonably short way using the current PCF Language.

Thanks, Bruce. I understand and I know the idea was not feasible. Thanks for taking a look. I really appreciate it.

You're welcome.