Paired Wilcoxon signed-rank on per-query RR (B vs A):
- N_nonzero = 10
- W+ = 50.00, W− = 5.00
- Two-sided p = 0.021824
- One-sided p (v1.2.0 > v1.1.0) = 0.010912
McNemar on rank-1 outcome:
| v1.2.0 rank-1 | v1.2.0 not rank-1 | |
|---|---|---|
| v1.1.0 rank-1 | 20 (concordant +) | 1 (regression) |
| v1.1.0 not rank-1 | 8 (improvement) | 1 (concordant −) |
- McNemar exact (binomial), two-sided p = 0.039062