How do we test a trading myth?
We write each belief down as an exact rule before we test it. We compare it with a fair yardstick on decades of data. Then a rule fixed in advance gives the verdict.
Why do we write the rule first?
If we chose the rule after seeing the results, we could keep changing it until it looked good.
So each rule is fixed and dated before any result exists. We use the belief's usual settings, not settings tuned to the data.
What do we compare it with?
Each belief is compared with what you would get anyway, and that yardstick is fixed before the test too.
For a timing rule, the yardstick is simply holding the market, with cash earning the Treasury bill rate while the rule is out. For a calendar effect, it is usually ordinary days, weeks or years.
How do we know it isn't luck?
Every result comes with a likely range, and a result whose range includes zero could easily be luck.
We test many beliefs, so a few would look good by chance alone. Each article also gives a wider range that allows for all the calendar beliefs we test. Every test we run is counted, even the ones that found nothing.
Why do we show the last 20 years?
A pattern that held in the 1930s may not hold today, so every long-history result also shows the last 20 years beside it.
Some beliefs held for decades and then shrank. Others never held at all.
Where does the data come from?
The calendar tests use daily US stock market and Treasury bill returns from Kenneth R. French's data library, going back to July 1926.
Returns include dividends and are before taxes. Code computes every number on our pages, so nobody types one in.
What do the verdicts mean?
The verdict comes from a rule we fixed before the test, never from a judgment after seeing the numbers.
- MATHThe belief held up. It beat its fair comparison and passed every check we set in advance.
- MYTHThe belief did no better than its fair comparison, or did worse. It never means you should do the opposite.
- HALF-TRUTHPart of the belief held up and part of it did not. The article shows which part is which.
- FADEDIt worked before it became famous, then stopped working.
- UNTESTABLE AS STATEDThe belief has no exact definition, so we show how far apart its common versions land.
Is this investment advice?
No, these pages are educational and describe the past, which does not predict the future.
A MYTH verdict means a rule did not beat its fair comparison. It never means you should do the opposite.