# r2_is n=5 nsamples=20000000 seed=99 imax=20; mean C = 1.60010 (2-2/n = 1.60000); walks absorbed before imax: 20000000
# columns: label i  r=E N_i e^{i/n}  SE  relSE  e^{-y^2/2 or -i^2/(2n^3)}  E N_i e^{i/n+i^2/(2n^3)}  heuristic (1-i/n^2)^n e^{i/n}
y=0.25 3 1.262999 2.24e-04 1.77e-04 0.969233 1.309295 0.961590
y=0.50 6 1.324173 2.92e-04 2.20e-04 0.882497 1.529266 0.841824
y=1.00 11 0.757107 5.14e-04 6.79e-04 0.606531 1.228444 0.497036
y=1.50 17 0.000000 0.00e+00 nan 0.324652 0.000000 0.100543
x=0.5 3 1.262999 2.24e-04 1.77e-04 0.964640 1.309295 0.961590
x=1.0 5 1.308212 2.64e-04 2.02e-04 0.904837 1.445798 0.890727
x=2.0 10 1.006889 5.12e-04 5.09e-04 0.670320 1.502102 0.574573
x=4.0 20 0.000000 0.00e+00 nan 0.201897 0.000000 0.017471
# sum_{i<=imax} E N_i = 4.000000 (n-1 = 4);  mean excursion length sum i E N_i/(n-1) = 4.9999 (n = 5; rel dev -1.02e-05)
# sum_{i>=1} |E N_i - e^{-i/n}| (i<=imax, plus e^{-i/n} tail 8.27e-02) = 1.9909  [upward-biased by MC noise]
