Function bootstrapPairedTest

Bootstrap paired test comparing per-query reciprocal-rank scores.

Algorithm:

1. Compute observed delta = mean(candidate) - mean(baseline).
2. Resample paired differences with replacement n=iterations times.
3. p-value = fraction of bootstrap deltas ≤ 0 (one-tailed, candidate better).
4. 95% CI = [2.5th, 97.5th] percentile of the bootstrap distribution of deltas.