Microcaps (again)

Seeing the results in the 2021 thread got me thinking. In my own backtesting, I don’t see really outstanding alpha until I am narrowing the universe to very small and illiquid companies (mktcap btwn 50-250. Does this sound about right?

For example, I have not found any worthwhile alpha in the last 5 years for anything I’ve tried on the R2000. Curious what others think. I’d love to see how Yuval’s system does on a 50 stock backtest on the R2000 over the last 5 years.

OK, I tried it. 50 stocks equally weighted from the Russell 2000 with no universe rules.

The first image is using the ranking system I’m currently using. It has been backtested on the period in question, though not on the Russell 2000 in particular. But Russell 2000 stocks are in my universe, so it’s by no means out-of-sample.

The second image is using a ranking system I last revised in March 2017, so except for the first two months of 2017, it’s out of sample. I chose this ranking system because of the date it was last revised, not for any other reason.

Regarding your first statement, of the twelve stocks with the largest weights in my current portfolio, only six have market caps between 50 and 250. The other six are larger. See the third image. You don’t have to limit yourself to very small and illiquid companies in order to have good results.




Thanks Yuval. This is good data. Would you mind posting the equity curve as well?

Yuval,
Did you use variable slippage? Also what PIT method was applied for the backtests?

As requested, here’s the equity curve, log scale so that you can see the divergences from the benchmark more clearly, along with the basics of the trading system.

As requested, here’s the equity curve for the second system (the one that uses a ranking system created in early 2017), log scale so that you can see the divergences from the benchmark more clearly, along with the basics of the trading system.



Thanks again, Yuval. I have started to make some progress cobbling together some of your public ranking systems together. Much appreciated.