Also, why does Lyellian relative dating match very well with radiometric dating?
Lyell used the extinction rates of European molluscan faunas to originally define the Cenozoic epochs (a few were added a bit later in the early 1800s, after people found them).
<30% extinct: Pleistocene–radiometric dating gives 1.800-0 Ma
~40% extinct Pliocene–5.333-1.800
~50% extinct: Miocene–23.03±0.05-5.333
~70% extinct: Oligocene–33.9±0.1-23.03±0.05
~90% extinct: Eocene–56.0-33.9
100% extinct: Paleocene–66.0-56.0
Other parts of the world have different rates, here are those for the US East Coast:
<5% extinct: Chibanian-Recent (middle-late Pleistocene and younger) (<700 ka)
15% extinct: early Calabrian (early-middle Pleistocene) (1.6 Ma)
60% extinct: early Gelasian (late Pliocene, under Lyell’s age-definition) (2.4-1.8 Ma)
67% extinct: late Piacenzian (mid-late Pliocene) (3.0-2.6 Ma)
75% extinct: late Zanclean to Piacenzian (early to mid Pliocene) (~4-3.2 Ma)
100% extinct: Tortonian (late Miocene) (~7-~5 Ma)