I just did a search on the forum and found it. The highest count recorded was 5700. I calculated around 4mi on a 15" wheel. Normally counts were around 2,000 by morning after the counter was reset at night.
The data was a single magnet that closed a reed switch every one rotation. Considering the wheel can spin a few times after the animal jumps off the number is probably a bit higher than if it were to detect whether or not the animal was actually running.
I have plans to build a more sophisticated system but I don't know when I'll actually do it. That one will measure quarter rotations and will log RPMs to help weed out false positives.