I tried the strategy, and probably because the companies knew I was invested in them, they did terribly.
I tried analyzing the companies, and picking the best of them, but it was too time consuming.
The first problem might be a fluke, or it might be the problem that good deals down in the bottom fishing zone are pretty scarce in the modern era.
This is a tough screen to implement, in a variety of ways.
The second problem is common: you generally do NOT want to look at the firms in a quant screen, and most especially not NetNets.
Murphy’s law of MI: if you filter the picks, the one you filter out will be the ones that do best.
In NetNets, every single stock is going to have problems, or they wouldn’t be that cheap.
So even if you’re good at security analysis, spotting problems is not too hard, but sorting them is trickier…which ones matter?
I long ago concluded that maybe you can beat the market by a point or two, but that the real reason for investing is because you enjoy it.
Well, I do disagree with you there. The reason to invest is to make money.
You can enjoy yourself in other ways!
There isn’t really anything to be gained with low volatility either, as long as the CAGR is good measured on the day you retire.
But, if you have multiple ways to invest that are equally likely to work well, then sure, invest in the way that suits your personality and is more fun.
I find the secret is to keep it simple, with (as you note) modest goals.
I like KISS.
Here’s a screen for your humble consideration. Using the Value Line 1700 set of stocks, old school.
Screen 1:
Take the 1200 of the 1700 stocks that are closest to their 52 week highs, just to “crowd source” out things that are doing badly or at least unfashionable.
Of those, take the 50 with the highest ROE.
Screen 1:
Take the 1200 of the 1700 stocks that are closest to their 52 week highs, as before.
Of those, take the 50 with the highest 5-year growth of sales per share.
Do an SOS: sum the ranks on those two screens and pick the best sum of ranks.
Monthly, buy top 20 stocks, hold-till-drop at rank 50.
i.e., for any stock no longer ranked in the top 50, replace it with the highest sum-of-ranks stock you don’t already own.
After friction, that would have beat the S&P 500 by 11.0%/year since May 1997.
OK, big deal, you can prove anything with a carefully tuned backtest. Great backtests are a dime a dozen.
But I created that screen some time in September 2018.
Since January 2019, without modification, it has beat the S&P by 5.5%/year—not too bad, especially given the high S&P return since then.
Only 3.25 years, but that’s not nothing. At least it didn’t blow up immediately.
And, even if it turns out to be totally overtuning (in the backtest era) and luck (post discovery so far), how bad can it be?
High ROE firms tend to outperform average firms over the long run (by 5.2%/year top 20 in VL, last 20 years, no friction).
High sales growth firms tend to outperform average firms over the long term (by 7.0%/year top 20 in VL, last 20 years, no friction).
You’re unlikely to do much worse than the broad market sticking to those two categories.
So…if the screen works, you outperform.
If it doesn’t, you probably track the market over time. (or so I predict)
Re the out of sample results:
This post describes almost the same thing, from Sept 2018.
https://discussion.fool.com/isn39t-mi-nothing-but-backtests-on-a…
But the figures above are from the specific variation that I saved on my computer.
Jim
