OSTOE — open-source theory of everything, claims registry

Falsified "Twin primes after a point live only on primorial flanks" — falsified by direct count (3 of 14,871)

id: twins-confined-primorial-flanks · filed: 2026-10-17 · depends: primorial-flank-dead-two-primes · falsifier: survived as a falsification — count below 2×10⁶ settles it

Mechanism (as proposed)

[S] Twin primes past some point are confined to the two flanks of base boundaries — the only twin survivors are the twin-candidate pairs the +1/+2 escape manufactures at each completed primorial base; every other twin dies out.

What was tried

Direct enumeration, same morning: twin pairs below 2,000,000 = 14,871. Exactly 3 are primorial flanks: (5,7) at 6, (29,31) at 30, (2309,2311) at 2310. Twins per inter-primorial interval grow (10 between 30–210, 55 between 210–2310, 398 between 2310–30030, 4,168 between 30030–510510) — the "desert between bases" the prediction requires is instead the densest territory on the strip. Nearest twin to 30030 is (30011,30013), 17 away; nearest to 510510 is 41 away. Twins do not home in on base boundaries.

What survives

[F] Primorial-flank candidate pairs DO occur at every base boundary (forced by the +1 escape) and survive at ~42% against a ~1% candidate-pair baseline — the small-wheel immunity is real and strong, but it is a point guarantee (two addresses per base), not a confinement. Base flanks are twin-rich points, not twin-only territory. The refutation is mechanical (count), not interpretive.

Downstream

Nothing downstream depends on the confinement; it was a stand-alone prediction. The surviving weaker version (flank immunity at a point) is filed in primorial-flank-dead-two-primes. The neighborhood-scale generalization was separately falsified the same morning (twin-rich-near-primorials).