I'm sure you know this already BoSkin, but to the benefit of others.
Generally the tracking is easier to do at lower speeds, because the servo and correction mechanism have more time to react. Also, disc speed related problems are smaller at lower read speeds.
So, a test at a lower speed probably gives you a better idea of how well the disc can be read in best possible situation on this drive and other drives like this. It could even be (no proof yet either way) that a test at a lower speed is a more telling of the actual problems the media has, as opposed to the drive.
However, testing at max speed can also be useful, because it simulates the average usage situation better: people usually use their drives at max speed for reading, so errors occuring at high speed in KProbe testing are more likely to appear in normal high speed use as well.
My current working hypothesis based on what I know is that testing at a lower speed tests more the quality of the burn+media (as opposed to stressing the capabilities of the reader) than when testing at a high read speed.
However, one wants to try and test is the key factor in deciding which test speed to use.