Actually, he did take into account the different angular velocities of the wheels. I did my own calculation of both the rotational (1/2 * I * omega^2) and translational (1/2 * m * v^2) energies of all three wheel sizes and my numbers are in agreement with Keith's.