I was taught that oxidized silver was a good conductor, especially in this case with the high impedances involved. I suspect the silvery contact is some other material that needs a minimum wetting current, typically 1mA to 10mA to remain reliable. Replacing contacts inside a relay is not what I would call an elegant solution. My impulse would be to see if the circuit could be modified to increase the contact current.
P.S. While oxidized silver may be a good conductor, it does resists soldering, hence the preference for gold to gold connectors.
One good permanent fix would have been to substitute wercury wetted contact reed relays. But they are a bit expensive and several UNELIGHTENED countries don't allow them any more. Or just plain gold contact reed relay elements could work quite well.
But the fix as posted is quite good, and I would give the detective work a good mark indeed.
I heard a similar story working in Automotive. A certain manufacturer had specified silver and German silver for the contacts in the brake switches but the manufacturer had reversed the materials in some of the parts. As it was explained to me, in the proper configuration when the contacts wiped against each other, any silver transferred between the contacts was essentially plated back onto the correct side. In the defective switches, the lost material wasn't replaced and additional material was also removed with each wipe. This caused brake switch field failures in the trucks at 15-20,000 miles. The situation got very ugly very quickly. Not recommended process for enhancing your career path.
Back in prehistoric times, there was a solution. I don't remember who the supplier was, but they sold "kits" for assembling relays. This you could "mix and match" power and signal rated contacts quite easily. In thsi case, you'd use a pair of contacts: one gold-plated NC, the other silver high-current NO, and hook them up as a "SPDT" pair. Of course, for production you could then have the part as a custom P/N built the way you needed it. I suspect another of the greybeards on this site will remember who the manufacturer was. This isn't really that rare of a problem (at least back in the old days when a lot of machine logic was relay-based).
Blog Doing Math in FPGAs Tom Burke 7 comments For a recent project, I explored doing "real" (that is, non-integer) math on a Spartan 3 FPGA. FPGAs, by their nature, do integer math. That is, there's no floating-point ...