Originally Posted by cyccommute

The roller isn’t wearing. The pin wears and allows the roller to follow something of an elliptical path as it goes over the cog and chainwheel. New chains follow a circular path so they engage the pitch of the cog and chainwheel properly so they don’t wear the edges of the teeth like a chain where the rollers can shift.

The question I have for you is how do your propose to measure roller wear if that is what you think is causing the problem? I can think of no way to measure roller wear separately from pin wear. It is possible, and even quite common, to measure secondary effects to measure a primary phenomena. In chemistry, for example, we often react two chemicals together and then further react the products of that reaction to see what the concentration of the unknown of the first two chemicals are. Pin wear and roller wear are related. Measuring pin wear also measures the effect of the rollers on the cogs.

Frankly, if you are worried about cog wear, you should probably be using one of these. I have one but find them to be mostly useless.