The model performs as designed. One outcome validates nothing on its own — but the sample size is building.