So can you now say with some confidence that competition performance doesn't correlate with job performance? That's still kind of interesting, although less so than the original conclusion.
The explanation of the effect did seem a bit too convenient.