My C-bit prefilter is a winner, my friend.
Convinced? ⇒ bc1qvmw6hxf7fhatxstf7vg53cd3n2a4jfa8at9wa6

You still do the same amount of hashing, so there's zero saving. We don't need to derive an address. The H160 is always enough. And it's always also necessary.
That comment misses the real win: we’re not paying the full cost of address derivation + Base58 on every candidate, only on the tiny fraction that passes the c-bit filter.