I've downloaded the last couple weeks of March 2018 runs from Rob Pope's Strava account to analyze. The GPX files and analysis results are on my Google Drive:
https://drive.google.com/open?id=1qfbEbcJZ4XGnI7ZuocHfO898t9gjjALq
A first pass through the data shows that there are some issues with the cadence/stride lengths but which are explainable. We can see in the cadence histograms that there are quite a large number of less than 30 rpm cadences (zeros) as indicated by the far left bar of data. When combined with his speed data, this results in stride length histograms that have a very large number of them at 15 feet to infinity, corresponding to when he is moving at speed but his cadences are at or near zero.
https://drive.google.com/open?id=129yLLgiYu6KWVNJQS1MNGwzpMrzzpOl9
https://drive.google.com/open?id=1fO0p_RkEs7FQex_EtPFZ_ycObMhaGGJN
Looking at his unsmooth cadence line plots, we can see that they often oscillate from 0 and then right to ~60 rpm:
https://drive.google.com/open?id=18oqIow9LG8JUh56LKIAnXcfm8eLmtwfk
We saw these 0-60 rpm cadence oscillations and infinite stride lengths in the Samir Singh, faith runner, investigation where the evidence strongly showed Singh used outside assistance in the form of a bicycle or similar. The difference here is that Rob's speed histograms are reasonable and he is openly using a push stroller, unlike Samir Singh. Rob's speed data shows he typically stayed below 8 mph with very few data points above that speed (similar to Pete Kostelnick's data). This is in contrast to Samir Singh's data where he had very large portions of his speed data greater than 13 mph.
https://drive.google.com/open?id=1EZaLcw2y4ZO8TAlewLZp-94rCCOSWED9
In Rob Pope's case, what appears to be happening is that sometimes he has the hand with his Polar GPS watch on the handlebar of his push stroller he uses to carry all his gear. When this is the case, the cadence does not register like normal because the oscillations of the watch are greatly reduced like as if one had his hands on bicycle handlebars.
So a first pass through Rob Pope's data passes my sniff test and looks legit. I'll have to do some code changes to my analysis tool to download all the rest of his Strava data for the runs since there are way more than I like to do by hand and to correct the timestamps inside the GPX files.