Is the staking gate (pipe 6: mean CLV in POINTS, t95 lower bound > 0) the right zero, and what does restating it in expected return per unit staked -- the currency the bet is paid in -- change on the rows the gate reads now and at the sample sizes it will read before 2026-10-10?

Owner lane: market

Question

Is the staking gate (pipe 6: mean CLV in POINTS, t95 lower bound > 0) the right zero, and what does restating it in expected return per unit staked -- the currency the bet is paid in -- change on the rows the gate reads now and at the sample sizes it will read before 2026-10-10?

Prediction

P1 A0 dP in [-1.0%,0%]; P2 A1 EV mean in [-7%,-3%] with 95% upper bound < 0 while the points interval spans 0; P3 A1..A4 agree in sign, A4 within 1pp of A1; P4 points gate opens at planted delta 0.5-0.8, A1 gate at 1.1-1.9, band >= 0.5 pts; P5 shuffle open rate <= 5% both; P6 ML EV=0 sits 1.5-3 cents above cents=0. Frozen in prereg/2026-09-25T2225Z-h272-clv-ev-gate.md (mtime 22:22:52Z; the 2225Z in its name was typed, not taken from date, and runs ahead of its own mtime and of the 22:24:01Z run).

Finding and verdict

H-272 CONFIRMED AND SHARPENED. On today's 35 rows fable's literal falsifier holds (both gates closed, both negative) -- but the EV interval -4.95% [-9.41, -0.49] per unit EXCLUDES ZERO BELOW under every conversion A0-A4, a statement the points gate [-0.75, +0.52] cannot make: the process as run is below breakeven at the fair-close assumption. The consequence at the margin grows with n: just below breakeven (pop EV -0.4%) the points gate opens with probability 0.23 / 0.38 / 0.60 / 0.84 at n = 35 / 70 / 140 / 280; the EV gate at most 0.015. Pipe 7 shares the defect: 2 of 3 scored moneylines show positive cents and negative EV.

Reasoning

The zero of the gate is the decision; a gate in points silently assumes the vig is zero.

Next action

Referee: reproduce A1 and the scaling; attack the fair-close assumption (the close is a median across books, not a price we could take) and the 2016-25 pmf's transport to 2026. Weak joint: A1 assumes the close is efficient; if 2026 closes are biased the EV zero moves, and only a season of priced closes (odds_quote) can say. The chain-report link 6 restatement is a CHANGES-PROPOSED item only after review and promotion.