Difference between cran version and master

[DavideACucci]

If you run the following example

library(simts)
library(avar)

dt = 0.01
N = 30*60/dt # 1h of data at 100 Hz

sigma2_true = 4.70596e-06
gamma2_true = 2.401e-10

x_model = WN(sigma2 = sigma2_true)+RW(gamma2 = gamma2_true)

x = gen_gts(N, x_model)

#plot(x)

av = avar(x)
plot(av)

est = avlr(av, wn = 1:6, rw = tail(1:length(av$levels), 6))

plot(est, decomp=TRUE)

You get this with cran version, which seems correct:

while with github master sometimes you get this, which seems strange (top scales and confidence intervals)