Partial page loads and CrUX: what field data quietly drops

The table comparing the Lighthouse clock with the CrUX field clock is the part I will keep. It makes the mismatch concrete: one controlled full navigation against an aggregate of eligible real experiences, with background tabs, early abandons, and bfcache restores quietly dropped from the comparison. The line about stopping the claim that field disagrees with Lighthouse also landed well. Lab measured a controlled run and field measured a population, so the honest move is to label which clock each number came from before selling a fix.