Why most strategies need a human behind the screenIntermediate

The honest result

When we test popular ICT and smart money setups exactly as they are usually taught, most come out roughly break-even after costs. Order blocks, fair value gap retests and breaks of structure, applied mechanically on five markets, land around a profit factor of 0.8 to 1.07. Yet plenty of traders make money with these concepts. Both things are true.

What the rules capture

A rule like 'enter on the first fair value gap after a sweep' captures the skeleton of a setup. It takes every gap that fits, on every day, in every market condition. A discretionary trader looking at the same chart skips most of them: the gap is too small, the sweep was of a meaningless high, there is news in an hour, the higher timeframe is against it, the day already moved its range.

Where the edge usually is

For most discretionary traders the edge is in the filtering, not in the pattern. That filtering is often hard to put into words, which is exactly why it is hard to test. The pattern is the doorway; judgement decides which doors to walk through.

How to use a backtest anyway

1. Test the plain version to get a baseline.
2. Add the filters you actually use, one at a time: a session window, a higher-timeframe bias, a news filter, a minimum gap size.
3. Watch which filters move the result. If adding your session filter turns PF 0.95 into 1.3, you have found part of your edge and you can now trust it.
4. Whatever you cannot write as a rule is your discretion. Journal it, then compare your real trades with the mechanical version in WickCheck.

The takeaway

A break-even mechanical result does not mean a concept is useless. It means the concept alone is not the edge. Your job is to find out which part of what you do adds the edge, and a backtest is the fastest way to find out.