The chance of a spurious signal from the direction, distance and magnitude GW170817 lining up with telescope observations by random chance is absolutely astronomically tiny.
let's say a "telescope observation" has to be on the order of a supernova explosion to be observable. There are about 30 supernova explosions per second in the observable universe. To date about 10,000 supernovas have been observed using EM spectra, on average that's about one observed major astronomical event every other day.
GRB170817A isn't a supernova though, so the SNe rate is irrelevant. First estimate of kilonova rates I found online is one per day per Gpc, and with your absurd extrapolation to the entire observable universe you get one per hour in one of the 40 000 degrees^2 of the entire sky.
Why do you think I'm assuming the type of the most well studied optical transient we've had this century? This is a source that was observed by something like a hundred different teams.
I'm ruling out a SN origin on the basis of the (at least) dozen of direct observations that show it isn't.
