# PREREG — entry 68's temperature leg, tested for what it was meant to test (30 September 2026, before any run) Entry 68 is OPEN: its leg 4d ("kinetic temperature within 5% of 0.02 at every drag") failed at low drag, 0.0169 at 0.05. The likely cause is Kramers' low-friction depletion: the energetic minds are the ones that escape, which cools the well. 4d was meant to test the INTEGRATOR, not the escape. This is a NEW leg, written before running, that separates the two. The band is not being re-cut on seen data. **4d′ THEOREM CHECK.** The same integrator as 68 (BAOAB, dt = 0.005, m = 1, temperature 0.02, the house's force T(x) + F(x), the reflecting floor at x = −0.999), but with a reflecting RIDGE at x = 0, so that no mind can escape. 2,000 minds per drag, γ ∈ {0.05, 0.1, 0.2, 0.5, 1, 2, 5, 10}. Burn-in 150 time units, then 300 time units of sampling. **⟨m v²⟩ within 3% of 0.02 at every drag.** If 4d′ holds, 68's failed leg is explained: the integrator is right and the cooling is the depletion. That can move 68 to Holds, the peer's rule permitting. If 4d′ fails, the integrator itself is suspect, and 68's times must be re-examined.