Does anyone have advice on making drift assumptions when building a MC Simulator? It seems to me as the mean of the sample becomes larger or the size of the sample becomes smaller, you should not assume zero. I also suspect it may depend on how far into the future one is looking. Any comments or suggested reading (preferably not ph.d. level) would be helpful.
What are you doing here? Drift is supposed to be that thing that goes away with risk-neutral valuation.
I’m not pricing options or anything sophisticated. Just bored and curious. I understand risk-neutral valuation and why drift is not an issue, but say you wanted to simulate price moves on an unhegded long position. It seems to me that at some point drift has real implications? Maybe my question should be what types of analysis are suitable for monte carlo, and what is not.