# Negative-D twists: forced central zeros are simple (Arb), Nov 10 2026, ~4:30 PM PT

Notes only; nothing deployed. Continues q8-negative-twists.md.

## Result  [project — Arb ball enclosure of L'(1/2), conditional on PARI coefficients beyond n=2000]
Sign −1 bases: 2.2304.8t5.b.a and 2.42025.8t5.a.a. For fundamental D<0 coprime to the conductor the twist keeps root number −1 (SST Prop. 8.7 [known]; numerically confirmed with FE bits −201…−204 for every row below once enough coefficients are supplied). The completed derivative integral (same validated theta-integral as twist-simplicity.py, nu flipped, Q multiplied by |D|, 256-bit Arb) gives an enclosure of L'(1/2) that excludes 0, so the forced central zero is **simple** in all 12 rows:
- 2.2304.8t5.b.a: D = −7, −11, −19, −31, −43, −59 (L' ≈ 2.64, 3.17, 3.82, 5.14, 4.89, 5.98)
- 2.42025.8t5.a.a: D = −7, −11, −19, −31, −43, −59 (L' ≈ 7.23, 4.18, 8.41, 4.63, 4.68, 6.64)
The last four rows were the previously unreliable ones; they are now closed by supplying 43k–135k coefficients to the integral and ~138k–425k to the PARI FE check (FE −87 to −204 bits; q8-negative-twists-more-run.log; PARI floats agree with the Arb ball to 15+ digits).
Caveat (honest): the Arb run audits the coefficient list against the independently audited list only for n≤2000; beyond that the a_n are PARI lfunartin output (Frobenius audit to p≤20000 covers the unramified part). Not an RH claim; not a universal simplicity statement.
The sign +1 base 2.2304.8t5.a.a has no forced zero; its six negative twists have root number +1 and nonzero central values (q8-negative-twists.md).
Files: q8_negative_twist_simplicity.py, q8-negative-twist-simplicity-results.json, run logs q8-negative-twist-simplicity-run*.log.
