Trading charts carry a subtle bias: markers are drawn after the fact, at prices that were not available when the signal formed. An order-flow indicator looked excellent on the chart, and a trader's eye saw a hit rate of around 95%. NOVAQ replicated the indicator exactly, with zero deviations across more than 770,000 data buckets, then measured where a trade could actually be placed.

The gap

The marker is drawn at the far side of the signal bar, while a real entry is possible only when that bar closes. On NQ the median distance was 34 ticks.

The test

Over 34,000 ways of using the signal were tested, and none survived the audit. In the main modules, 29 of 9,338 configurations were positive in and out of sample, about what chance would give. The signal marks an exhausted impulse, with a small contrarian effect below costs.

A lesson about controls

An early test gave an exciting p-value of 0.005. The random control was too easy, because it flipped more often than the real signal. A fair control removed the result. Designing the control is often the hardest part of a study.

A later search across 453,600 combinations left a single fragile candidate. It did not replicate on the second market, so it went to a forward test without capital.

Limits

The fine-grained data covers about two months, so a longer history could change the regime picture.