>If the instability is something that develops over time as a genuine change in behavior of the chip, and not merely an artifact due to the evidence of instability taking time to pile up, then the extreme voltages are by far the most plausible culprit.

That's all fine and dandy, but can you please cite some evidence to support those claims?

This should not be such a farfetched request.