Subject: Re: Timing: recent highs
Given data has continued to fill in since then, do you think 99, 114, or some other number is best? I would guess that any number roughly in this range is within the error bars, but thought I'd ask.

I wrote some python code about 6 years back to explore questions like this. You can find it here:
https://github.com/jcmiii/Dyin...

I've also got some spreadsheets you can download there.

There are actually two parameters: M (looking for a new M-day high)and N (in the last N days). This program found that SPY returns are maximized for M in the range M=138-142, and N=214. If the algorithm is restricted to M=N my recollection is that 114 was best, but just by a hair. But in 6 years the optimal parameter values could have changed. And of course they change significantly when considering a different index (e.g. IWM, NAS).

John