I'm studing Saff's papers about behavior of zeros corresponding to the best approximation of general continuous functions published in his web, but I think more specific and better results for polygonal functions might be found.

http://www.math.vanderbilt.edu/~esaff