Complexities will kick in, as skewness will induce heteroscedasticity. No problem when using the right model but creating chaos when left unaddressed. Time series data will raise challenges like non-stationarity. Vincent's approach…"
"To really feed your method with some realistically simulated data I suggest to generate lognormal data with a range of different sigma's and see how you compete with the traditional techniques. A good link to theory, application and generality…"
"Not worth it on these simplistic data sets to be honest. As long the results are similar there's nothing to worry because we can continue using numerical or parametrized recipes. It might all become a bit more interesting when simulating more…"
"Also - considering your remark on the Gaussian assumption - don't underestimate the strength of the central limit theorem in this particular method: you're averaging random variables in each bin, inevitably bringing you very close to a…"
"While I'm studying this post a bit deeper, after first read I noticed the cause of the (...must be too...) big discrepancy between the model-free and classic CI's: It's not the individual sample's SD and/or CI that is studied…"