Systematics vs. new physics origin of DESI BAO outliers

Determine whether the outlier measurements in the DESI Year-1 BAO dataset—specifically the points identified as P = 1 (z = 0.295, BGS, D_V/r_d), P = 3 (z = 0.51, LRG1, D_H/r_d), and P = 4 (z = 0.71, LRG2, D_M/r_d)—arise from systematic errors in the DESI observations and analysis or instead indicate new late-time cosmological dynamics beyond the ΛCDM model.

Background

The paper applies the Internal Robustness (iR) Bayesian model comparison approach to the DESI Year-1 BAO measurements to test for internal consistency and identify potential outliers. The analysis highlights three specific data points (P = 1, 3, 4) as potential outliers that notably affect robustness metrics.

Excluding these outliers yields parameter estimates for w0 and wa that are consistent with ΛCDM, suggesting that the anomalous points may be driving apparent deviations toward dynamical dark energy in the full dataset. The authors also test robustness against an alternative dark energy parameterization (Barboza–Alcaniz) and find the same outliers persist.

Given extensive DESI systematics checks reported elsewhere, the source of these anomalies remains ambiguous, motivating a focused investigation into whether they stem from residual systematics or genuine new physics at late times.