Past movement speed gave no clear 24-hour wind-change forecast gain for selected Atlantic storms in 2001-2025

We tested whether past movement speed improves the next 24-hour wind-change forecast for selected Atlantic storms. We used revised NOAA HURDAT2 records at 10-22 N and 20-60 W, with origin winds of 35-110 knots and tropical status throughout each 48-hour window. Models used earlier years only. In 1,121 test windows from 120 storms in 2001-2025, adding past speed reduced mean squared error by 0.84%. The conditional 95% three-year block interval was -3.58% to +4.37%. Equal weight per storm gave a 1.53% error increase. A positive pooled speed coefficient did not establish a stable forecast gain. Future-status selection, revised records, dependent windows, a simple model and no ocean or atmospheric state measurements limit the result. This is not a test of the causal effect of ocean cooling. Code, aggregate evidence, source hashes, a numerical audit and a figure are included. Source records and per-window tables stay outside the package.

tropical cyclone · HURDAT2 · translation speed · intensity prediction · chronological validation · null result