In playing around backtesting screens on Portfolio 123, I've found that using one of their pre-defined ranking systems to do the final selection of stocks can improve performance significantly.

With this in mind, I modified lizgdal's YEY26 screen definition to select 50 stocks rather than 10. Then, on the screen's 'Settings' page I added the 'Core: Value' ranking system, and set 'Max # Stocks' to 10. I also changed 'PIT Method - Prelim' (preliminary financial data) to 'Use', because this booted the backtest results a bit.

I backtested this for one year (09/01/2025 - 09/01/2026) with rebalance every 4 weeks. This generated a return of 62.69%.

Backtesting from 01/01/2021 - 09/01/2026 (as far back as my subscription allows) produced the following results:

Ann Ret:    39.39%
Max DD:    -25.25%
Sharpe:      1.26
Sortino:     1.75
Std Dev:    24.34%
CorBench:    0.6    (correlation with S&P500)
R-Squared:   0.36
Beta:        0.96
Alpha:      20.74

To be clear, I tried a bunch of ranking systems, and some reduced returns significantly.

As an aside, as a mechanical investor, I'm usually only dimly aware of the names of the companies my screens pick. I certainly don't keep track of what industries they might be in.

--Peter