URGis the V1 model's probability (× 10) that this setup pays ≥ 1% favorably within 2 hours of the fire. UNC is how close to 50/50 that probability is — high UNC = unclear read, low UNC = confident either way.

V1 doesn't decide buy / stop / sell. The detector fires on the TFO rule; the model only scores the already-fired setup. Any STOP/TGT lines are rough heuristics, not model output. The 1% threshold is hardcoded, not ATR-adjusted yet — V2 will normalize across tickers. Open the deep-dive view for the full model-scope breakdown.