Ouch! Hmm ... depending on just how it was failing, that might even have worked. If the forward resistance was increasing as it cooked, then the new one would limit the voltage in the forward direction, so it might cook more slowly. But if the reverse resistance was decreasing, it would do no good at all. I never actually took the time to analyze the failure mode of the few I saw die -- just took off power and replaced them as quickly as possible. :-)
I do have one (probably still good) which came out of the base of a small Sanford surface grinder to power the mag chuck. It also had an electrolytic capacitor which was *supposed* to be 8 uF, but was closer to 0.002 uF when measured. I replaced it with a much bigger HV electrolytic, and a silicon bridge rectifier (and a redesign of the switching circuitry) to maximize the holding capacity of the mag chuck -- before I replaced it with a permanent magnet chuck.
Enjoy, DoN.