I did post using every start date, and have data on any sub-time period of the nasdaq100 you want to 10/93. The 325 day SMA on SPY works and massively reduces the worst drawdown; recovery also massively faster than qqq and spy. I presented 1,2, 5, 10 stock strategies.2 week and 4 week trading periods. Sharpes, drawdowns of 10% or more, mdd, ui, and using the best data you can buy short of academics', i.e. Norgate data. to 1993. And much better than just using 12 month momentum. It has made me a monster return. Sometimes owning 5 stocks is less risky than 100 or 500 according to hard data. I have tested it every which-way. Trailing stop limit orders mid-month. Trading more often or less often. Hold till drops. Weighted versions. Mixing strategies together. The stocks are liquid. Studies show .1% spreads are realistic for stocks this liquid.
And unlike gtr you can find the particular stocks owned and determine how dependent you were on a few big winners (you're not). You do cycle through winners, which is the point.
And you would have generated 2 to 4 times SPY and QQQ had we known this in 1993.
It is better than index funds on every single statistical front but one: the number of 10% or more drawdowns, but you quickly recover from them and make much more than an index fund in between.
There are better screens than pure momentum, but even that is impressive.
I paid thousands of dollars for data in the 80s when I started with mechanical investing; now with AI you can do what Wall St. used to be able to do but not you.
I see no evidence the historical returns are fake or wrong. MU is up hundreds of percent in the last year. INTC went up 100% in a month. The tests showing big returns are verifiable.