You're sort of on the right track with this (and I'm not sure why you're getting downvoted). The difference in talent of the top 250 or so CEOs is actually very small, but the huge size of modern companies means that even a small difference in CEO talent makes a huge difference in dollars.
Gabaix and Landier [0] run a counterfactual (based on their model) and suggest that "if firm number 250 could, at no extra salary cost, replace for a year its CEO by the best CEO in the economy, its market capitalization would go up by only 0.016%. ... this talent difference implies that the pay of CEO number 1 exceeds that of CEO number 250 by 530%. Substantial firm size leads to the economics of superstars, translating small differences in abil- ity into very large differences in pay."
250 doesn't sound like a large enough statistical sample to eke out even the most trivial performance differences in a single person when you consider the size of the companies being discussed.
Gabaix and Landier [0] run a counterfactual (based on their model) and suggest that "if firm number 250 could, at no extra salary cost, replace for a year its CEO by the best CEO in the economy, its market capitalization would go up by only 0.016%. ... this talent difference implies that the pay of CEO number 1 exceeds that of CEO number 250 by 530%. Substantial firm size leads to the economics of superstars, translating small differences in abil- ity into very large differences in pay."
[0] http://qje.oxfordjournals.org/content/123/1/49.abstract