Performance is measured by looking back in time. The obvious technical issue is how far back to look. So far as I know, there is absolutely no clear answer.
For example, consider Classical Specter/FF5 playing the 16 day averages from March 2020 through yesterday.
The Long average natural log return is .98 which is a little less than tripling your money. That performance is underestimated for the 3x Bulls because SPY is included.
However if we look at performance starting in May, the story is different:
Here we see the strategies to be worthless. My explanation has been that the upward trend in the market, makes all the positive states with an F good to play with the possible exception of xFF. The positive F states are xCF, xFC, and xFF. With xCF and xFC playing well, xFF by definition has to play relatively poorly.
Have to admit it is a little distracting to publish the signals, because there are extra things to worry about such as the quality of my comments and the reader’s reaction.
The bottom line is it is a lot of work to make easy money.