Dave's running a live switch cycle on his second channel. I watched it for a few minutes. Tried to calculate the resistance with the formulas shown but looks like its 1.7K ohms the one way a little better the other. Guessing I am missing something as I think he said they could only read up to 10 ohms or so. So much computing power and they resort to sticky notes.
He also cleaned off the accumulated dust/debris after every bunch of cycles.
If that's fiberglass dust then it's an abrasive, just sayin'.
I don't think it's fiberglass - solder mask more likely. In normal use the meter would be picked up and put down potentially dislodging the dust anyway after every use or so - you cannot replicate normal usage easily. These sorts of test can only ever be a rough guide to wear rate, and you would need to test many to get some sort of average.
I was under the impression that this round of testing was done live, non-stop, 50K half cycles, no cleaning with the shim installed correctly. I only saw the last few minutes so if you saw him cleaning it again, that's a problem. If your pulling that out of your ass and don't really know, we should ask Dave.
I don't need to run more of the same meter to see the difference in wear. Between the pictures and resistance metric, it's good enough for me. Some have certainly cut deeper than mask. Then there is that detent spring, it's not always the contacts that are the problem.
Then again, I have been told many times that I can't draw a conclusion from the transient tests I run because the sample size is one.