It is not the normal cost of the protection : No other AV induce such poor overall performances
You're mixing, or, at least, saying I'm mixing, the "cost" of protection that is reducing (a bit, we wish) the performance with a false positive detection. The false positive is always an error, a bug, that should be addressed. It's not a cost.
If we think about technology, the "detection" technology (blacklist) (in opposite of whitelisting or deny all) will always pay the cost of false positives though, as any error - done with good will - will led to a false detection.
Well, technology philosophy
