What “both halves of the tape” means
One good window can be luck. Two separate windows that both survive costs is a higher bar, and still not a forecast.
The split
Take a stretch of history, for example two years of regular-session prices. Cut it in half. Score the same rule on year one and on year two. Keep the rule only if both halves are profitable after the cost assumption you already chose.
That is not magic. It is a simple way to reject a story that only worked in one regime, one news cycle, or one lucky cluster of days.
What it does not prove
Both halves green does not mean the next year will be green. You still searched among many combinations. You still have a small sample on jumpy names. You still have no third year sitting unused as a true holdout unless you reserved one.
It is better than showing only the full window, because a big year can hide a dead year. It is not the same as out-of-sample testing on data you never touched.
When a name fails
If no version of the rule clears both halves after costs, the honest product is to scratch the name. A failed name is information. A list that only shows winners is a brochure, not a research file.
Key takeaways
- Split the tape so one hot year cannot carry a dead year.
- Both halves after costs is a filter, not a guarantee.
- Names that fail the filter should be said out loud.