That's a standard fallacy I often run into: over estimating second order effects. Yes, unless the ball is perfectly smooth or in a vacuum you will get slight curvature. But it would be so slight that it would look like B.