The Monte Carlo filter is a simple simulation which allows you to assess whether a subset of probes is unusual compared to a randomly selected set of the same size. It makes no assumptions about your data or the distribution of the current quantitation values.
A Monte Carlo simulation simply selects matched random sets of probes of the same size as your target set from your currently selected probe list. It then compares either the mean, median or maximum values within the random set to the value from your target set. Finally after running many simulations it assignes a p-value to the simulation, which is the proportion of tests in which the randomly chosen set produced a value which was the same or higher than the value from your target set.
The Monte Carlo filter isn't actually a filter, in that the probe list it produces always contains exactly the same set of probes as your target list. It will add the p-value for the overall simulation to each of the probes in this list.