Do Classic Stock Signals Beat The Market? We Tested Eight
Published October 19, 2026 · 6 min read · Data: 2013–2026, 164 monthly tests on about 2,200 liquid stocks
The Short Answer
No signal we tested picked winning stocks reliably. For all eight, the top-ranked 10% of stocks returned about the same as the average liquid stock, well within the range of luck. The one pattern that showed up was at the other end: the worst-ranked stocks did badly, so avoiding them helped slightly.
On This Page
How We Tested
- Stocks: US stocks priced at least $5 that trade at least $2 million a day, about 2,235 names a month on average.
- Schedule: rank the stocks on the last day of each month, buy the top 10% at the next open and hold 21 trading days. That gave 164 monthly tests from January 2013 to August 2026.
- Costs: 0.30% per month, which assumes the whole portfolio turns over every month.
- Yardstick: the same-universe average return, which was +0.49% per 21 days after costs. SPY returned +1.24%, so the average liquid stock trailed the index.
- Rules fixed in advance: we judged each signal on 2013–2019 first. A signal only counted if it did better than average there and held up in 2020–2026, which it had never been tuned on.
The Eight Signals
| Signal | Buys Stocks With |
|---|---|
| MOM_12_1 | The best return over the past 12 months, skipping the latest month |
| MOM_6_1 | The best return over the past 6 months, skipping the latest month |
| REV_1M | The worst return last month (short-term reversal) |
| LOWVOL | The lowest price volatility |
| NEAR52 | Prices closest to their 52-week high |
| TREND200 | Prices furthest above their 200-day average |
| ATTN | Recent volume lowest relative to the 60-day average (low attention) |
| COMBO | The best average rank across MOM_12_1, NEAR52 and LOWVOL |
The Results
Extra return per 21-day hold from the top 10%, compared with the average liquid stock, over all 164 months. Whiskers show the 95% range. A signal with an edge would sit clearly to the right of zero.
Excess over the average liquid stock, after a 0.30% cost. Gray bars have ranges that include zero.
| Signal | 2013–2019 | 2020–2026 | All Months | 95% Range |
|---|---|---|---|---|
| MOM_12_1 | +0.09% | +0.35% | +0.22% | -0.31% to +0.74% |
| MOM_6_1 | +0.13% | -0.18% | -0.02% | -0.51% to +0.49% |
| REV_1M | +0.05% | -0.97% | -0.45% | -1.02% to +0.15% |
| LOWVOL | +0.09% | -0.17% | -0.04% | -0.56% to +0.46% |
| NEAR52 | -0.12% | +0.17% | +0.02% | -0.36% to +0.39% |
| TREND200 | -0.15% | +0.16% | 0.00% | -0.48% to +0.52% |
| ATTN | -0.02% | -1.28% | -0.64% | -1.01% to -0.30% |
| COMBO | +0.12% | +0.01% | +0.07% | -0.42% to +0.52% |
Every range includes zero except ATTN, which was reliably worse than average. Only REV_1M passed the first stage in 2013–2019, and in 2020–2026 it reversed to -0.97%, so it failed the test. The famous momentum signals did not beat the average either.
The One Pattern: Avoid The Bottom
Splitting stocks into ten score groups shows a shape that repeats across signals. The middle of the pack earns roughly +0.7% to +1.2% per 21 days, while the single worst-ranked group earns far less.
All 164 months. The best group does not beat the middle, and the worst group lags it.
| Score Group | MOM_12_1 | LOWVOL | NEAR52 | COMBO |
|---|---|---|---|---|
| 1 (best) | +1.01% | +0.75% | +0.81% | +0.85% |
| 2 | +1.02% | +0.91% | +1.06% | +1.01% |
| 3 | +0.93% | +1.02% | +1.04% | +1.03% |
| 4 | +0.96% | +0.97% | +1.09% | +1.17% |
| 5 | +0.94% | +1.02% | +1.03% | +1.06% |
| 6 | +0.90% | +0.99% | +0.99% | +1.04% |
| 7 | +0.93% | +1.01% | +0.93% | +0.84% |
| 8 | +0.81% | +0.87% | +0.84% | +0.65% |
| 9 | +0.69% | +0.82% | +0.52% | +0.56% |
| 10 (worst) | -0.30% | -0.49% | -0.43% | -0.34% |
In plain terms, stocks that are volatile, far below their highs and have weak momentum lag the rest. That suggests a screen that skips the worst group instead of hunting for the best. Dropping the worst group added only about +0.1% to +0.14% per month, roughly 1.5% a year. For the combined score it was +0.13% per month with a range of +0.03% to +0.22%.
Red bars are years when the worst-ranked group lagged. It lagged in 10 of 14 years, and 2020 reversed sharply.
What This Means For Investors
- A simple ranking rule on public price data is unlikely to beat the market after costs. Any site that says otherwise should show a test like this one.
- Avoiding speculative, high-volatility stocks matters more than finding the next winner. That fits our other result: buying top-gainer lists lost money.
- For most people, a diversified fund is a reasonable benchmark. See What Is An ETF?
Limits Of This Test
How We Tested This, And What It Can't Tell You
- Survivorship. The universe is today's listings, so failed companies are mostly missing. We measure against the same universe, which partly cancels this but does not remove it.
- Eight signals. With eight tests, an isolated strong result can be luck, so we only trusted results that held in both periods.
- Before tax, simple costs. Real costs and taxes depend on the account and the order size.
- Not advice. This is research on past data, not a recommendation to buy or sell any stock.
Related Reading
- Do Top Gainer And Loser Lists Make Money?
- Does Shorting A Stock After A 20% Drop Work?
- What Is An ETF?
This is research on past data and not a recommendation to buy, sell or short any security. Past results do not predict future returns.
Frequently Asked Questions
Does momentum investing beat the market?
Not in our test. Buying the top 10% of liquid US stocks by 12-month momentum earned +0.22% per month more than the average liquid stock, with a 95% range of -0.31% to +0.74%, so we cannot tell it apart from luck.
Which stock signals did you test?
Twelve- and six-month momentum, one-month reversal, low volatility, nearness to the 52-week high, distance above the 200-day average, low recent volume, and a combined score of momentum, 52-week high and low volatility.
What did work?
A small effect at the other end. The worst-ranked 10% of stocks lagged the rest, and skipping them added about 0.1% to 0.14% per month, roughly 1.5% a year. We treat this as a hypothesis that still needs forward testing.
Is this investment advice?
No. It is research on historical data and not a recommendation to buy or sell any security.