The past wind maximum gave no clear added prediction gain for offshore Atlantic wind changes in 2003-2023

We tested whether the highest wind recorded so far in a storm improves next-day wind-change estimates beyond current state, recent wind change, age and motion. We used a fixed IBTrACS v04r01 North Atlantic snapshot. Fits used earlier seasons only. The 2003-2023 test has 1,655 selected offshore windows from 133 storms. Adding the past maximum minus current wind increased mean squared error by 0.20%. The conditional 95% season-block interval for error reduction was -0.97% to +0.43%. RMSE rose from 15.878 to 15.893 kt. All six planned sensitivity intervals included zero. This result limits one simple history term in a small retrospective benchmark. It does not rule out storm memory or ocean effects. Future sample selection, revised winds, missing environmental controls, model form and 21 test seasons limit inference. The package includes source fields, hashes, code, all predictions, uncertainty draws, a figure and numerical checks.

tropical cyclone · North Atlantic · wind history · chronological prediction · IBTrACS · null result · storm recovery