p10 prereg (4 Nov 23:35, before looking): independent check vs 20CRv3 daily-mean prmsl (nearest grid point), per ship mean of (obs - 20CR) in inHg, median over days, ships with >=30 days.
WRS transcriptions postdate 20CRv3 (2019) so these obs were likely not assimilated: treat as independent.
P1: across non-flagged ships, Pearson r between network offset and (20CR anomaly minus its fleet mean) >= 0.6.
P2: slope of 20CR-anomaly on network offset in 0.7..1.3.
P3: of the ~10 1863-shift ships, >=7 show a same-sign shift vs 20CR (before/after 1863-07-01) of >=0.03 inHg.
