Maybe.

The problem is we don't know how many hoaxes are there and have _not_ been discovered. If there are a lot, then the model is failing, if there are only a few, then it is working.

Alas, without that data, we can't make a call on the wikipedia article either way.