Note 02 5 November 2026 EquitiesFactorsSmall caps

Four tries at beating the market by picking smaller companies

We built four rule-based ways of picking US small and mid-sized stocks and tested each on years of history, failed companies included. The best one picked better stocks than the pool it chose from, but by too little to rule out luck, and none of them beat simply owning the S&P 500.

Our final version, a year

12.8%

The stocks it chose from

9.6%

The S&P 500 (SPY)

14.8%

March 2017 to September 2026, after estimated trading costs. Every figure here comes from a backtest, a simulation on past data.

1What we tried

Smaller companies get much less attention than the giants. Fewer analysts follow them, and big funds often can't buy enough of their shares to bother. If prices are slow to catch up with the facts anywhere, it should be there. So between July and November we tried to write a set of rules that would pick the better ones: no hunches, just company accounts and prices, applied the same way every time.

We had four goes at it. Each time, the rules and the bar for success were written down before we looked at a single result, so we couldn't move the goalposts afterwards. Every test kept in the companies that later went bust or were bought out, paid for the cost of trading, and was measured against two things: the pool of stocks the rules were choosing from, and SPY, the S&P 500 fund anyone can buy.

2What happened

The first version came from well-known investing books: cheap stocks, screened for quality, with some momentum on top. It lost to the pool it was picking from, 4.6% a year against 6.8%. Over this decade, cheap small companies simply didn't beat expensive ones.

The second tested 24 ideas from academic research, mostly about profitable companies and companies buying back their own shares. A few looked promising on their own. But when you try that many ideas, a few will look good by chance, and none held up once we allowed for it. We kept the three best as candidates anyway.

Then we went further back. We rebuilt the market for 2012 to 2017, years none of our tests had touched, and ran two of the candidates on them without changing a thing. One failed outright. The other did no better than the stocks it was choosing from. The third couldn't be tested, because the older data wasn't good enough.

The last version combined the strongest ideas: profitable companies, companies shrinking their share count, and stocks that had been rising. It made 12.8% a year. The stocks it chose from made 9.6%. That looks like a win, but the gap swings so much from year to year that it can't be told apart from luck, and most of it is explained by well-known effects that are cheap to get elsewhere. Meanwhile SPY made 14.8% a year, with smaller falls along the way.

3Why it fell short

The pond mattered more than the fishing. Over this decade smaller companies as a group trailed the S&P 500 by five to six points a year. No stock-picking edge we found came anywhere near making that up.

A few winners did most of the work. Take out the final version's ten best stocks and its edge over its pool disappears. One company, AppLovin, during its 2023 to 2024 run, accounted for about a third of it.

Timing it didn't help. The results came and went with market conditions, but nothing we could see at the time said when. Every rule we tested for switching in and out did worse than just sitting still.

4What we took from it

Getting the data right was harder than designing the strategies. A popular free price history turned out to hold only the companies that survived, which makes any test run on it look better than it should. At one point a single mislinked record made one of our tests show a gain of 234% a year. Once it was fixed, the result was about zero.

If you read about a strategy that beat the market on paper, three questions are worth asking. How many ideas were tried before this one? Were the companies that failed left in? And did it beat an index fund, or only the stocks it was choosing from?

None of this shows that investing in smaller companies can't work. We tested rules applied to public accounts and prices over one decade, and it was a decade the biggest companies dominated. Someone doing real research on individual businesses, or a different decade, could tell a different story.