Have you tried reducing the frame rate in the demo to see if the same effect is true for 30 fps as well?

Also, I wonder if a simple square is enough to prove the "sameness" of motion blurs across different framerates. A fast camera pan across a landscape may yield different results.