The single substitution
Change the regime read and leave every other input alone.
The counterfactual is deliberately narrow. The entry is the same, the stop is in the same place, the volatility translation produces the same distance, and the price does the same thing. The only difference is the classification committed before the trade — Trend in one version, Normal in the other. Everything that follows is downstream of that one field, which makes this the cleanest available demonstration of how much the read actually decides. It also makes the exercise honest: nothing has been arranged to favour either outcome.

