🤣 Okay. Let’s say hypothetically the paid relay performance is 100%. Is there another reason? Is there some subset of legitimate users I would be missing?
Discussion
Same rationale as relays in general. Diversify where your data is.
Let’s say a paid relay goes under, rugs you, or goes down for maintenance. You have others.
Other users may be using different relays too that you may want to interact with.
I think we’re still really early and there is experimentation in both paid relay selection and administration. YMMV.
Redundancy. If I write to N relays following the emerging NIP-65/Gossip model you should read me there, if one is down or the returned data is inconsistent, try another.
Can you ELI5 the gossip model? Are clients responsible for connecting to all the relays they want their kind 10002 events to be seen by?
