# rfPermute Normal Behavior?

I'm using the rfPermute to calculate the pval of the feature importance. Is it normal for a feature to be significant, but those with a higher %IncMSE to be not? Or have I done something wrong?

LFS.rfP2 <- rfPermute(
C40_WKS ~ .,
data = LFS,
ntree = 1000,
na.action = na.omit,
nrep = 500,
num.cores = 6
)

plot(
rp.importance(LFS.rfP2, scale = FALSE),
alpha = 0.06
)