A note on Wolf's prime-gap distribution at extended range: reproduction of finite-size drift parameters and a negative result on a putative ω(g) correction

A short note documenting (a) a negative result on a previously claimed sqrt(omega(g)) correction term to Wolf's heuristic for consecutive prime gaps, and (b) a reproduction of Wolf's published finite-size drift parameters a(x), s(x) on a sieve dataset extended to N = 3e14. The earlier preprints (v1-v3, Zenodo records 19796305, 19958063, 19975065) reported a residual structure of the form a*sqrt(omega(g)) + b*log log N when the empirical histogram tau_g(N) was compared against a baseline C_g * Li_2(N) * exp(-g*C_g/ln N). That baseline differs from Wolf's own formula (Phys. Rev. E 89:022922, 2014, Eq. 7) by an extra factor of C_g in the exponent. Refitting against the correct Wolf 2014 baseline collapses the sqrt(omega(g)) amplitude by 78-95% and absorbs the log log N coefficient into Wolf's already-published drift. Out-of-sample R^2 on hold-out N > 1e13 is -18.3 (worse than the weighted mean), confirming the residual was an artifact of the mis-specified baseline. The note also reports Wolf 2014 Section III drift parameters a(x), s(x) on the extended range, cross-checked against Cohen's tabulation (Experimental Mathematics 34, 2024) for t in {30, 36} with relative error < 5e-5. The three earlier records v1-v3 should be considered superseded by this final note. The repository at https://github.com/dkrse/math-final-wolf provides the full pipeline, sieve histograms (28 files, N up to 3e14), and the analysis notebooks reproducing every figure and table.