L∞ estimate off the strongly major arc, eq. (8.18). Let 8.7×1036≤x≤e3100 be an integer, K=1000, N0=4×1014, T0=3.29×109. For every α with ∥α∥R/Z>T0/(3.6πx),
∣Sη0,x/K♯(x/K,α)∣⋅∣DN0/3(α)∣≤0.078xKN0/3.
This is the pointwise input to the minor-arc bound; combined with Proposition 8.4 by Hölder it bounds the minor-arc contribution to (8.11) by 0.078×7.09<0.56 times Kx2(N0/3)3, which is what the main term of Proposition 8.3 beats. The paper proves it by symmetry reducing to T0/(3πx)≤α≤21 and splitting into regimes:
weakly major, 3.6πxT0<α≤4πxKT0: Proposition 7.2 plus Lemma 3.3, ∣η^0(αx/K)∣≤7.2Klog2/T0≤0.077;
α≥20/N0: the crude bounds ∣S∣≤1.01x/K (Lemma 4.3, (5.11)) and ∣DN0/3(α)∣≤1/(2α);
(8.19) 3πxKT0≤α<N020: reduce to (8.20) ∣Sη0,x/K♯(x/K,α)∣≤0.078x, write 4α=q1+O∗(1/q2) with 80N0−1≤q≤KT0πx, and apply Theorem 1.3 — (1.12) for (x/K)2/3≤q, (1.9) for (x/K)1/3≤q≤(x/K)2/3, (1.10) for q≤(x/K)1/3 — each giving a bound below 0.078x/K once logx≤3100 and T0=3.29×109 are inserted.
Preamble
import Definitions.Def_TaoFivePrimes_ArcSplit
open MeasureTheory
T. Tao, Every odd number greater than 1 is the sum of at most five primes, Math. Comp. 83 (2014), 997-1038, https://arxiv.org/abs/1201.6656, Section 8, eq. (8.18) and its proof, pp. 41-43; Theorem 1.3, eqs. (1.9), (1.10), (1.12); Proposition 7.2; Lemmas 3.3, 4.3