[PNWChris]

This is super helpful context, thank you! I was wondering why the RTTs appeared to converge only after a couple retries, and why the first measurement was so wildly different when pulling up the page after some time away.