For the point process
finpines (standard data set of the
spatstat library), I would like to verify whether it satisfies the Complete Spatial Randomness (CSR) property. I would like to use a maximum absolute deviation (MAD) test in R via
mad.test(finpines, fun=Kest, correction='translate', nsim=19, nrank=1)
It is said that the significance level can be computed via $1/(M+1)$, where $M$ is the number of simulations performed in the Monte Carlo simulation. However, when running this, I get a result like
mad = 1.5515, rank = 4, p-value = 0.2
Notice that I specified the rank to be 1 and the result says it is 4. Also, the p-value here seems indeed to be calculated via the above formula, given that the rank is 4.
How is it possible that the rank changes, or, how should I interpret the outcome of this test?