== job 1840: #159's largest L vs #161's L(T_x,q); the LOOSE support == A. CONTROLS — the published diagonal and the published run census ok A: L(T_5,p=7) = 2 == #161 published diagonal 2 (D = 3, T_5 = 3) [spectrum head [(2, 2)]] ok A: L(T_7,p=11) = 1 == #161 published diagonal 1 (D = 15, T_7 = 15) [spectrum head []] ok A: L(T_11,p=13) = 2 == #161 published diagonal 2 (D = 135, T_11 = 135) [spectrum head [(2, 6)]] ok A: L(T_13,p=17) = 2 == #161 published diagonal 2 (D = 1485, T_13 = 1485) [spectrum head [(2, 72)]] ok A: L(T_17,p=19) = 2 == #161 published diagonal 2 (D = 22275, T_17 = 22275) [spectrum head [(2, 1088)]] ok A: L(T_19,p=23) = 3 == #161 published diagonal 3 (D = 378675, T_19 = 378675) [spectrum head [(2, 11746), (3, 62)]] ok A: T19/p=23 runs of 3 = 62 == published 62 (and max = 3 == published diagonal 3) [spectrum [(2, 11746), (3, 62)]] ok A: diagonal entries 7 and 8 -- mine L(T23,p=29) = 2 (below) and the route's quoted L(T29,p=31) = 4 -- match the published 2 and 4 B. T23: my census max vs #159's own recorded max L (read at source) ok B: T23 D = 7952175 == 7,952,175 and largest gap == G2(23#) = 204 T23 by 29: L_census = 2 top classes [(2, 243822)] R_loose = 2 loose support = 3 T23 by 31: L_census = 3 top classes [(2, 247526), (3, 276)] R_loose = 2 loose support = 3 T23 by 37: L_census = 2 top classes [(2, 95896)] R_loose = 2 loose support = 3 ok B: T23 by 31: my L_census 3 == #159's recorded max L 3 ok B: T23 by 37: my L_census 2 == #159's recorded max L 2 ok B: T23 by 31: my R_loose 2 == #159's recorded longest qualifying gap run 2 ok B: T23 by 37: my R_loose 2 == #159's recorded longest qualifying gap run 2 ok B: every measured L_census <= 4 at T23 (route success condition L <= L(T_x,q) <= 8) C. THE LOOSE SUPPORT over the primes #161's table covers (T23 words) ok C: T23 by 29: pure-Python R_loose 2 == numpy R_loose 2 (two implementations of one definition) ok C: T23 by 31: pure-Python R_loose 2 == numpy R_loose 2 (two implementations of one definition) ok C: T23 by 37: pure-Python R_loose 2 == numpy R_loose 2 (two implementations of one definition) ok C: max loose support over q in [29,1009] at T23 = 3 (at q = 29) — <= LMAX = 8? ok C: for every q > G2(T23)+2 = 206, R_loose = 0 (the support is exactly 1: no gap can satisfy g = q-2 and g = 0,2 is impossible for a T23 gap) [123 such primes] ok C: #159's quoted folds T31/37 (support 3+1 = 4) and T37/41 (3+1 = 4) are <= 8 with their own recorded max L = 4 and 3 D. #968's identity, checked numerically at the shared pairs ok D: T23 by 29: #159's largest L (census max) 2 <= #161's L(T23,29) = 2 (same object; equality expected) ok D: T23 by 31: #159's largest L (census max) 3 <= #161's L(T23,31) = 3 (same object; equality expected) ok D: T23 by 37: #159's largest L (census max) 2 <= #161's L(T23,37) = 2 (same object; equality expected) 23/23 checks passed; all_pass=True