This is an interesting subject and I've wondered about other screens, particularly momentum.
Starting in 2018 many momentum screens really started doing poorly.
They have picked up some of late, but still compared to the "tuning period", results are not good.

For whatever reason, and probably wrongly, I'm not as willing to wave that off to over-tuning as quickly as some others.
After all some of the screens are rather simple so it's hard to conclude there was excessive over-tuning.
YLDEARNYEAR would be an example of that; there's hardly anything to the screen.

Rather, it seems to me that there's just something that has changed in the macro environment that has thrown the screen out of favor.
And therefore, the screen performance could improve if the conditions became favorable again.

As a momentum screen example, consider VG-Horse per this post http://www.datahelper.com/mi/search.phtml?nofool=y...
The screen has some constraints on institutional holdings and float, and looks for growth in EPS and Sales.
It also looks for low price and high free cash flow and has a typical momentum-based final sort.

At the time of the post above, the top 10 returned 34% CAGR from 19970902 through 20150918.
Since then, 20150918 through 20230310, top 10 only returned 6% - a clear failure. Did sales and EPS growth go out of style?
From pandemic lows, it looks somewhat better, top 10 returned a more respectable 17%, still half what it had been.
Sticking with it throughout 19970902 to 20230310 gives a CAGR of 25%, still nothing to complain about.
(Screen link: gtr1.net: GTR1 Backtester - VG-Horse_V2
The incd.s in original screen had to be removed to extend backtest date through 2023).

I think momentum, at least as defined in many SiPro screens, has suffered for some years now, and the recent focus on mega-caps hasn't helped either.
Might momentum ever return?

Mark