I don't know what you mean by "pre-determined". I download historical data and from that calculate both trend and momentum for multiple lookback periods. Each of these result in a rating, which are then combined in an overall rating for the ETF. This is done for the ETFs I track, and then ranked to find the top 5 performers.
Back in the 90s, I susbscribed to a mutual fund database and analysis program called "Investor's FastTrack". That was before and at the very beginning of the advent of ETFs. It was my first exposure to ranking and switching algorithms, and to the research on industry/sector trend and relative strength persistence. Some of those algorithms derived their lookback parameters from backtesting, and these rolling lookback parameters were applied indefinitely. I guess that is what I was referring to as "pre-determined". Another approach was to choose and update the lookback(s) by technical/chart observation. Sounds like you may be using a hybrid of both approaches.
By the time I had wrapped my head around the concept and decided that I preferred it to the interpretation of chart constellations, oscillators, trendlines, etc., the dot com party was over and I had missed a larger part of it than I should have by moving in and out of the market WAY too often. I was fortunate to have gone through this phase while the bull dragged me and my naivety along.
So, we're definitely using the same general playbook as I see it. I was just curious how you arrived at your method. And by the way, my universe of potential trading options is a lot larger now that I am retired and have more time to indulge my tendency to "piddle". The verdict is still out on whether the extra effort will be compensated.
Last edited:

