His logic is as follows:-

The Moons of Mars appear to be captured asteroids, and a rough guess can be made at their composition and expected reflectance spectra. (From memory, they are C1 chondrites, but I could easily be wrong). Quick check of Bill Arnett's 9 planetspages coming up [btw, these are worth a look by everyone from time to time, and right now, in the wake of the Pluto demotion, the header is quite amusing]

OK, the present state of knowledge is that they may be C-type (Carbonaceous Chondrites) but the exact sub-type is unknown. Their densities are too low to be solid rock. Bill reports theories that this may imply that they have some ice mixed in, but I suspect that this theory predates the discoveries of low asteroid densities in cases such as Eros. (But in its favour, Phobos appears to be the source of some sort of outgassing)

Anyway, IF we know the spectra of unaltered material that we would expect from Phobos and Deimos, then we could measure the actual spectra and determine how much martian material has been lifted into space and now contaminates the regolith of those moons. From that, we could deduce the cratering rate on Mars.

There are some problems, though. We would need to know the orbital history in quite some detail, as the rate of contamination will be critically altitude-dependent, or so it appears to my casual eye. Back-integration of the tidal drag problem may not be accurate enough for this (but at least there are no oceans and no plate tectonics on Mars, which are the main complication for Earth's moon)

Also, there is a fundamental problem:-

Both Phobos and Deimos are being impacted AT THE SAME TIME by the same population of impactors, so the process of gardening of these moons will rapidly acheive a steady state where material gained from Mars is balanced by material lost from the moons.

Nonetheless, it would be an interesting calculation...