Coherence and growth, October 11, 2026: N = 4,000 per group, K = 1.90 (the threshold), eps = 0.1 Six fresh starts (starts 7 to 12) of the long run at the threshold: the same model, settling, gap (rms 1e-07 radians, put back every 5 time units) and length (6,000 time units), run by nudge_test.py's own code. Set before the run: measured start by start over t = 2,000 to 6,000. 1. Same time: the correlation between coherence and growth rate in windows of 200 time units. 2. Later: the correlation between coherence and the growth rate 200 to 500 time units later. A link if the average over the starts is more than 3 error bars from 0. Error bars: from the starts. Check: start 1, run by this script for 50 time units, repeats the long run exactly. 1. Same time (windows of 200): +0.2763 (0.1044); starts +0.43, -0.25, +0.50, +0.50, +0.22, +0.26; below 0 in 1 of 6: no clear link 2. Later (200 to 500 later): -0.2061 (0.0186); starts -0.22, -0.23, -0.13, -0.16, -0.27, -0.23; below 0 in 6 of 6: more coherence now, slower growth later Beside Test 16: the fresh starts' long-run rate, t = 3,000 to 6,000: +0.0145 (0.0007); Test 16: +0.0147 (0.0006); difference -0.0001 (0.0009): agree Shown, not read as tests: the fresh starts, then the earlier run (where the patterns were noticed) Same time, correlation by window length (time units) 25 50 100 200 500 1,000 (+0) 1,000 (+250) 1,000 (+500) 1,000 (+750) fresh +0.2835 (0.0558) +0.3061 (0.0676) +0.3011 (0.0775) +0.2763 (0.1044) +0.1742 (0.1258) +0.3160 (0.1392) -0.0702 (0.2750) +0.2975 (0.1639) +0.3788 (0.1687) earlier +0.2992 (0.0560) +0.3128 (0.0623) +0.3156 (0.0678) +0.3162 (0.0722) +0.2439 (0.0950) -0.4207 (0.2357) +0.0696 (0.2879) +0.4934 (0.2661) +0.3355 (0.2938) Coherence against the growth rate L time units later (negative L: the growth comes first) -500 -300 -200 -100 -50 +0 +50 +100 +200 +300 +500 fresh +0.105 +0.156 +0.229 +0.182 +0.169 +0.265 +0.112 -0.037 -0.195 -0.198 -0.214 earlier +0.146 +0.190 +0.259 +0.221 +0.193 +0.271 +0.115 -0.007 -0.179 -0.224 -0.308 Coherence |z| over t = 2,000 to 6,000 (average of A and B over the starts): fresh 0.0251, earlier 0.0270 The two measures for the earlier run: same time +0.3162 (0.0722); later -0.2476 (0.0353)