Athletes on different surfaces: Asphalt, gravel, grass. It matters.

Hopefully you have noticed that went you are on a crushed limestone trail that your foot tends to spin and slip, kind of like being on ice, real rough ice.  Some people spin more than others but most people will notice some degree of slip or spin. For the LUCKY runner who is almost purely sagittal, they will strike the ground (more likely midfoot when the surface is a little slippery or unstable) and roll off the toe with minimal spin at the ball of the foot during toe off. They might get a little posterior slip at push off which is why we change our gait to be a little more apropulsive on slippery or unstable surfaces and why we see some hip extensor-flexor relationship problems in those that chose a limestone path to run on.
But for many, many who are not as lovely sagittal progressers as the lucky few, there can be much spin slip, typically external rotation foot slippage as we engage our glutes (external rotators) to push off. This can be a result of increased foot progression angle and many other things, but a biggie is that runners are often engaging the more economical narrow based step width, something we like to call the cross over gait (search our blog for dozens of references). When we have a narrow step width, our foot knee and hip are not stacked, the foot is inside the hip width spacing at foot contact. This means our foot attacks the ground with more supination, and thus more pronation afterwards and these accentuations engage more of the frontal and rotation axes in the foot, ankle and the lower limb. This can lead to more aggressive progression to high gear toe off (medial tripod) and thus an induced spin.  And so, when you engage frontal and rotational axes when you are moving forward on a slippery surface, spin results at the contact interface, more spin than we want, and that means changes in the ground reactive events leading to apropulsive gait mechanics and other nasty things we do not want such as toe gripping and clenching, a futile febble attempt to gain better purchase on the ground (in fact a useless strategy, because your foot interfaces with the sock liner of the shoe first, the sole of the shoe doesn't get to see it).
If you are a slipper, head for the asphalt road, so that you do not engage the apropulsive gait cycle biomechanics that can lead to a myriad of attempted stabilization co-contractions, quite frequently in the hip and pelvis stabilizers as one tries to nullify some of the excessive rotation in the limb-foot.  At the very least, mix up your running surfaces, give your body the exposure to variable surfaces, and as this article suggests, different motor patterns and strategies. It just might make you more durable.

-Shawn and Ivo, the gait guys

From the Dolenec study:
"Six male and two female runners participated. The participants ran at a freely chosen velocity in trials on asphalt while in trials on gravel, and grass surfaces they were attempting to reach similar velocities as in the trials on asphalt. Muscle activation of the peroneus brevis, tibialis anterior, soleus, and gastrocnemius medialis of the right leg was recorded. Running on asphalt increased average EMG amplitude of the m. tibialis anterior in the pre-activation phase and the m. gastrocnemius medialis in the entire contact phase compared to running on grass from 0.222 ± 0.113 V to 0.276 ± 0.136 V and from 0.214 ± 0.084 V to 0.238 ± 0.088 V, respectively. The average EMG of m. peroneus brevis in pre-activation phase increased from 0.156 ± 0.026 V to 0.184 ± 0.455 V in running on grass in comparison to running on gravel. Running on different surfaces is connected with different activation patterns of lower leg muscles. Running on asphalt requires stiff ankle joints, running on gravel requires greater stability in ankle joints, while running on grass is the least demanding on lower leg muscles."

Coll Antropol. 2015 Jul;39 Suppl 1:167-72.

Activation Pattern of Lower Leg Muscles in Running on Asphalt, Gravel and Grass.

Dolenec A, Štirn I, Strojnik V.
https://www.ncbi.nlm.nih.gov/pubmed/26434026