M  E Z^k, k <= 10: recursion == integer-partition formula: True; kappa_n = int x^(n-2) dK (n <= 10): True; coefficient of every free moment m3..m8 affine in (m1, m2): True; the only other term is 126 m3^2 in E Z^10: True; Poisson check max error 5.3e-46  (0.5s)
M  number of monomials of E Z^k, k = 0..10: [1, 0, 1, 1, 2, 2, 4, 4, 7, 8, 12]; E Z^10 constant term = 945 (= 9!! = 945)
N1 2gap     : accepted by Arb 0/142, by Sturm 0/142  (expected none) OK
N1 gap/2    : accepted by Arb 142/142, by Sturm 142/142  (expected all) OK
N2          : accepted by Arb 0/142, by Sturm 0/142  (expected none) OK
N3 2slack   : accepted by Arb 0/142, by Sturm 0/142  (expected none) OK
N3 slack/2  : accepted by Arb 142/142, by Sturm 142/142  (expected all) OK
     majorant gap min(phi - |z|): min 4.596e-07, median 1.251e-06; dual slack: min 1.376e-06, median 1.976e-06
N5  chord coefficient replaced by its negative (126 c10 (L+U) -> -126 c10 (L+U)): dual still holds in 12/142 cells (sensitivity only)
N4  worst cell 127: max F = 0.8791944631 (= max vertex value: True); with C = max F - 1e-9: Bernstein accepts False, Taylor accepts False (expected False, False)
N4  worst Taylor cell 127: Taylor bound 0.8799740542; with C = that - 1e-9: Bernstein accepts True, Taylor accepts False (expected True, False: the Taylor bound is not tight)
N6  top edge lowered by 1e-6: cover broken in 127/127 cells (expected all)
N7  A - 1/1000 >= 4/(3C): False (expected False)
RESULT: ALL NEGATIVE TESTS BEHAVE AS EXPECTED (2s)
