diff --git a/docs/ISSUES.md b/docs/ISSUES.md
index 2907c777..4df0c10c 100644
--- a/docs/ISSUES.md
+++ b/docs/ISSUES.md
@@ -118,6 +118,23 @@ to widen the destination-lane time budget during the hold (a budget
 change, out of scope per this round's constraints) — no artifact-shaped
 limiter remains.
 
+**Done bar + next hypothesis (2026-08-17 goal closeout, user-set):** the
+user's benchmark is retail's ~6 s login — at retail's SMALLER default
+window (17×17 = 289 blocks vs our 25×25 = 625) and WITHOUT our
+complete-at-reveal guarantee (retail pops in after reveal). #418 stays
+OPEN until login reaches retail-feel (~6–8 s) with the readiness
+guarantee intact. Next hypothesis, in order: (1) widen the
+destination-lane time budget ONLY while the reveal hold is active (the
+world is hidden behind the tunnel; the ~5 s drip spreads ~500 ms of
+real publication CPU, so a hold-scoped 2 ms → ~8 ms destination-lane
+ceiling should collapse it to ~1–2 s while the tunnel still renders
+smoothly), reverting to the authored budget at reveal; (2) overlap the
+destination-priority mesh uploads (~0.4–1.2 s) and composites (~0.6 s)
+with the drip if they serialize today. Retail's authored tunnel-exit
+choreography (TunnelContinue 2–5 s + two 1 s fades) is retail behavior
+and is NOT a tuning target. Predicted result: gate-ready ~2–3 s +
+authored exit ≈ 5–8 s total — retail-feel at 2.2x retail's window.
+
 ## #417 — World ambience keeps playing (and re-firing) on the character-select screen after the in-world logoff
 
 **Status:** ✅ FIXED 2026-08-17 (logout-audio round; fix + tests in the same