Iteration 95 — the Tail Pair: Double Kill, Prediction Confirmed KILLS #34 · #35 · FOUR SIGHTINGS · PROPOSAL-GRADE

In plain words: last iteration the loop bet against its own candidates — registering in advance that both remaining tail meters would measure something real and regime-flat. Both predictions landed, and the second at the sharpest margin of the entire campaign. The tail-dependence coefficient — the probability that one series is in its extreme tail given the other is — reads a massive 0.84 (seventeen times what independence would give): when one of the pair goes extreme, the other almost always goes with it. And that number is 0.8378 in the covid crash versus 0.8400 in the calm mid-cycle — identical to two decimal places, against a noise floor a hundred times wider than the observed gap. The copula-structure reading told the same story. Four tail meters, four clean measurements, four regime-flat readings: the pattern is now proposal-grade science. On this substrate, the tails have one law that never changes with the weather; the body of the distribution is where the weather lives. The consequence is registered: the batch's last extremal candidate now carries a four-sighting prior against it and runs last; the next candidate is from a genuinely different family — regime-conditional causal structure — and gets its fair, unprejudiced exam.