The Base Rate You're Ignoring in Every Trade

One of the most reliable ways to be fooled by data is to ignore how common something is before you weigh the evidence for it. This is base rate neglect, and it is behind a large share of misplaced confidence in pattern-based trading. The uncomfortable headline: a signal that is right eighty percent of the time can still be wrong the majority of the times it fires.

What a base rate is

The base rate is how often something happens at all, before you look at any signal. Market tops are rare; the overwhelming majority of days are not tops. That prior matters enormously, because a signal does not operate on the rare event in isolation, it operates on every day, most of which are ordinary. A pattern's accuracy when a top is truly forming tells you nothing on its own until you also know how often the pattern fires on the far more numerous days when nothing is forming.

The arithmetic that surprises people

Suppose genuine tops occur on 5% of days. Suppose your signal is good: it correctly fires on 80% of real tops. Suppose it also fires on 15% of ordinary days, which sounds like a low false-alarm rate. Now count the alarms across a long stretch of days. The true tops are few, so the correct fires are few. The ordinary days are many, so even a 15% false-alarm rate produces a large number of false fires. When you add them up, most of the signals are false, not because the signal is weak but because the ordinary days it also triggers on vastly outnumber the rare event it is trying to catch. The signal can be genuinely informative and still be wrong most times it triggers. Both statements are true at once, and the reason they feel contradictory is that intuition quietly drops the base rate.

Where this bites traders

The pattern is everywhere in market narratives. A setup is described as preceding a reversal a high percentage of the time, and the claim is accepted without the two numbers that would actually decide it: how often reversals happen at all, and how often the setup appears when no reversal follows. The same neglect hides inside the high-win-rate strategy that ignores the size of its rare loss, and inside the confident read of a rare chart formation. In each case a single accuracy figure is doing persuasive work that only a base rate and a false-alarm rate together can honestly support.

The fix

The correction is not a better signal, it is the habit of always asking for the denominator. Before trusting that a pattern predicts an event, ask how common the event is to begin with, and how often the pattern fires on everything that is not the event. A hit rate on the rare case is meaningless until it is weighed against those two things. Most trading stories quote the flattering first number and quietly omit the denominator, which is exactly where the truth about a signal lives. Judging a condition against the base it should be measured on, rather than the story told about it, is the discipline behind our case study.