A health warning: as with so many screens that worked for ages in backtest, this is an approach that stopped adding value not long after the suggestion was made. Of course: MI meets Murphy. And Mungo's curse.

It's true I noticed that it was a better than average industry over the long haul, among a few others, quite a while ago. But the suggested MI screen stopped outperforming around end 2019 right when I was posting some backtests, and started underperforming strongly, at least for a while, around mid 2022. Didn't keep up with SPY in 2021, 2022, 2023, or 2024. Though the 2020-2024 CAGR was 12%, which is hardly a wipeout at least. I don't know what happened in 2025.

An interesting way to try to figure out if it might recover when market dynamics change from their current setting, or should be relegated to the scrap heap of overmined history: check the median sales growth rates, ROE, and P/S and P/E valuation multiples of the picks in the last couple of years that it worked. Compare that to current figures. If the group's typical business success metrics are still good (the first 2 things), but the valuations are better (last 2), then it's a good omen that it might start working again when people stop buying nothing but gold and LLMs.

Jim