He was telling me about his normal run circuit, and the fact that without thinking about it, he matches stride length to surfaces. So for instance running on a hard downhill slope he will subconsciously shorten his stride, protecting his joints from concussion.
On an uphill slope he will extend, as concussion is naturally reduced uphill. These variations in stride, he was very vehement in stating, are nothing to do with his feet - in the sense that they are not adaptations because of sore feet - its all about joints and soft tissue.
If you have good proprioception, then as you cover the ground you receive neural feedback about it - whether it is uneven, concussive, deep - and you adapt accordingly.
Shoes reduce proprioception dramatically, and its only when you run barefoot or (very nearly, as in Andy's case!) that you get this feedback. Ironically, in expensive trainers, research (which we quoted in "Feet First") shows that although proprioception is reduced, concussion isn't - so the result is that your joints suffer more shock, not less because you fail to reduce stride length on harder surfaces.
This brings me of course back to horses. Frankie is a good example - he is a TB with a weak digital cushion and therefore a poor ability to shock absorb. He works nicely on roads at the moment, and trots out perfectly level and tracking up, but his trot stride lengthens slightly if you trot him on grass.
If you were to shoe him again tomorrow, he wouldn't be able to shock absorb any better BUT if you trotted him on tarmac he might well have a longer stride (given his past history he might also go stonkingly unlevel again, but lets leave that aside for now!).
Now, a longer stride and a faster speed add up to even more concussion, and yet many people would automatically assume that a longer stride=comfortable horse=a good thing. They often also assume that a shorter stride=sore feet.
However Andy's experience is that neither of these assumptions are true. I suspect that for many horses, a shod foot+weak digital cushion=reduced proprioception=longer stride=more concussion. Perhaps we should be a lot more analytical when we judge stride length and factor in proprioception and concussion before we judge whether its a good thing or not.