Does anyone know how diagnostics percentiles are calculated? I have a Sharpe of 1.2879 which claims to be 100%, but I know it is lower than many other Sharpes out there. So it can’t be using the actual distribution of diagnostics.
I found this post. My percentiles agree on the basis of some of those numbers but not on others.
Not that it matters a great deal, although this could be giving false hope to folks out there whose models are not as good as they appear on the diagnostics (and I appreciate this is only validation diagnostics and nothing to do with live).