[foldr]

>The exceptions were when the whole class did poorly and no one would have gotten better than a C - then they adjusted to ensure some people got at least a B - but usually not based on a distribution.

See, that illustrates the problem. If cheating pushes grades up, then the difficulty won't be adjusted, even if no-one would actually have gotten high grades if it wasn't for cheating. The professor might think the test was fine because at least some people got Bs, but what if all those people cheated?