The null-info diagnostic

The standalone tool: drop in an entry rule and a stop/target ratio, and see whether the entries beat the chance baseline the geometry implies.

The null-info diagnostic is the fourth validation check — null information — pulled out as a tool you can run on its own. No strategy, no account: an entry rule, a stop and a target are enough.

What you give it

  • Market— a symbol and timeframe from forex, commodities, indices, crypto or stocks. It runs over the symbol’s full price history. Futures aren’t supported here.
  • Indicators — picked from the indicator library, the same one the strategy builder uses.
  • Entry condition and direction— a condition in the same language as a strategy’s entry rules. It can use your indicators, like ema200, and each bar’s open, high, low and close. volume works on crypto and stocks only: forex, commodities and indices carry no volume data.
  • Stop and target — the stop sits a multiple of ATR from entry; the target sits R times that stop distance.

What it computes

It simulates every trade that entry would have taken over real bars, then compares two numbers: your observed hit rate — the fraction that reached target before stop — and the null hit rate, which is 1 / (1 + R), the win rate a zero-skill entry produces with that target on a driftless random walk. The gap between them is reported as a z-score.

The result reads Beats chance, Does not beat chance, or Inconclusive — the last one when there is nothing to judge, such as zero trades generated or a degenerate baseline at exactly 0% or 100%.

Why this is the sharpest cheap kill
If the entries do not beat the chance baseline for their own geometry, there is no model, threshold, or optimization downstream that can build an edge that is not there — the weakness is upstream of anything a model touches. And it takes no trust in a model to read: it is arithmetic on your own hit rate.

What it does not do

It checks direction only — whether the entries lean the right way — not formal significance. A full strategy submission’s pre-registration and placebo checks are what judge whether an edge that beats chance is also large enough and stable enough to believe.