Almost all the foxes are long and xFM which is the non buy signal does not suggest an alternate case.
Trend Following has been long since 10/27. That most recent position took a hit for the last few days of October and simultaneously some issues with my logic appeared. I think I’ve got a better handle on this now.
Originally, I was using a fixed median. With the rewrite, I was calculating a median for the entire population which worked OK but is theoretically not kosher, because decisions are made for days in the past based on today’s median. Still, that is more like a statistical misdemeanor than a felony.
The results here are based on signals over a 50 day period versus a median of the previous 200 days. That resolves the median objection and still produces apparently excellent results.
I pulled my article for SA before the editorial review and hope to submit a new one for next week.
