Part V of the 6N twin-prime project. Part IV closed the long-gap (210) residual of the conditional twin-gap singular series with a single-centre spatial-compression penalty, but left the short-gap (42) residual, and posed the genuine two-centre correlation between the factorisations of N and N+d as the open problem. Here we measure that correlation directly and obtain a quantitative law. Conditioning a twin centre N on divisibility by a prime q, we measure the shield S(d,q) = P(N+d twin | q|N) / P(N+d twin) on the 23,988,173 twin centres of S10. The shield factors cleanly into two measured pieces: a modular-shift gain (because q|N forces N+d ≡ d mod q, the right centre's q-survival jumps from its real baseline rate to certainty when d is q-safe) and a cross-prime hedge (the same conditioning shifts the right centre's residue distribution at the other primes, mainly q'=5, changing their survival). The product S(d,q) = gain × hedge reproduces the measured shield for both 6ΔN = 42 and 210, across all admitting primes, to within 1.5%, stable between S9 and S10. Two natural guesses are overturned. First, the 42 shield is a q=5 effect, not q=7: since N+7 ≡ N+2 (mod 5), having 5|N places the right centre at the safe residue 2, while the prime 7 that divides the physical gap is nearly inert. Second, the hedge is below 1 for 42 and above 1 for 210, which is why the single-centre theory of Part IV under- and over-shot respectively. Both factors require the measured (not idealised uniform) modular distribution of twin centres, which carries the Part I enrichment. This is the right-centre conditional survival law. We do not bridge it back to the omega-stratified gap-preference distribution r(d|omega) of Part IV — and with it the final fate of the 42 residual — which we keep as the open problem. No claim is made about the infinitude of twin primes or any prime k-tuple conjecture.
Ruqing Chen (2026) studied this question.