diff --git a/docs/re/BACKLOG.md b/docs/re/BACKLOG.md
index 35b3821..c96ccec 100644
--- a/docs/re/BACKLOG.md
+++ b/docs/re/BACKLOG.md
@@ -507,6 +507,18 @@ search cannot find a *schedule*.
   ❔ **Phases 2 and 3 are entirely untested** — every run has stayed in phase 1,
   and phase advance was never located; a phase transition is now the most likely
   place an arrival could exist.
+* 🟡 **(2026-08-24) ANSWER: enemies come into play PER PHASE, deployed at phase
+  start** — [`mission-phase-deployment.md`](mission-phase-deployment.md). Static:
+  phase 1 = 37 squadrons / **42 members**, phase 2 = 36, phase 3 = 49. Dynamic:
+  **`deployed` = 41** from the first instant, never changing. **41 vs 42 — the
+  deployed set IS the phase-1 roster.** That makes everything coherent: all 116
+  records exist at load but only the current phase holds craft; no arrival was
+  ever seen because phase 1 never completed; route times are within-phase fly-in
+  timings, not release times. ⚠️ The **off-by-one is unexplained** and recorded,
+  not rounded away. ❔ **Single remaining question: what advances a phase.** It is
+  directly observable (deployed would jump), did not happen in 435 s with 42
+  kills, and `SUBObjectiveSettings` names 9 objectives but carries no trigger.
+  Unsearched: `StageMessageSet_S02.tbl` (never resolved) and the PE in guest RAM.
   🔴 `SYLPH_HZ=3` refuted as a lever: it gave the LOWEST frame rate (8/s) with the
   most kills, so pilot polling is not the throttle.
 * ~~🚧 BLOCKER: t=210/240 unreachable in one turn~~ — **superseded, see above**;
diff --git a/docs/re/mission-phase-deployment.md b/docs/re/mission-phase-deployment.md
new file mode 100644
index 0000000..38a67c4
--- /dev/null
+++ b/docs/re/mission-phase-deployment.md
@@ -0,0 +1,76 @@
+# Enemies come into play per PHASE, deployed at phase start
+
+Status: 🟡 strongly supported and internally consistent, one unexplained
+off-by-one; ❔ what advances a phase is now the single open question.
+
+This is the answer the wave line of work has been converging on, and it is not
+the one the route timetable first suggested.
+
+## The evidence
+
+**Static**, from `Route_S02.tbl` cross-referenced with `UnitGroup_S02.tbl`:
+
+| phase | squadrons with routes | members |
+|---|---|---|
+| 1 | 37 | **42** |
+| 2 | 34 | 36 |
+| 3 | 44 | 49 |
+| no route | 3 | 3 |
+
+(Squadrons appear in more than one phase — 37+34+44 = 115 against 111 squadrons —
+so the member totals overlap and do not sum to 116.)
+
+**Dynamic**, measured repeatedly across many runs:
+
+```
+deployed = 41 of 116 records, from the first observable instant, never changing
+```
+
+**41 deployed against 42 phase-1 members.** The deployed set is the phase-1
+roster, to within one record.
+
+## What this makes coherent
+
+Every previously puzzling observation falls into place:
+
+* **All 116 records exist at mission load** — confirmed on two stages — but only
+  the current phase's squadrons hold craft.
+* **`deployed` is 41 before the ready room ends and never moves**, because phase 1
+  deploys as a unit at load rather than on a schedule.
+* **No arrival was ever observed** in up to 435 s of verified-live flight with 42
+  kills, because phase 1 never completed in any run — and phases 2 and 3 are
+  where the other 85 members are.
+* **The route times are within-phase fly-in timings**, not release times, which
+  fits the frames reading (Stage 01 phase 2's t = 1500/1800/2100 → 50/60/70 s at
+  30 Hz).
+
+So: **enemies come into play per phase, deployed together at phase start.** The
+route entry gives each squadron its entrance path and the moment it flies in.
+
+## The off-by-one is not explained
+
+41 observed against 42 expected. Candidates, none tested: the player's own record
+may be counted differently; one squadron may hold craft in a way the liveness
+read misses; or one phase-1 member may genuinely not deploy. It is recorded
+rather than rounded away — a one-record gap is exactly the size of the errors
+that have misled this corpus before.
+
+## ❔ The single open question: what advances a phase
+
+This is now the whole remaining question, and it is the same one raised many
+iterations ago in
+[structures/stage-mission-tables.md](structures/stage-mission-tables.md) — the
+stage record declares `Phase_1..3` but nothing found says what ends one.
+
+What is now known that was not then:
+
+* a phase advance is **directly observable** — `deployed` would jump from 41
+  toward the phase-2 roster, which is unmissable with the current probe;
+* it did **not** happen in any run, including 435 s of verified-live flight with
+  42 craft destroyed, so it is not a short timer and not a small kill count;
+* the likely gate is an **objective**, and `SUBObjectiveSettings_S02.tbl` names
+  nine of them but carries only rewards and HUD strings — no trigger.
+
+Places not yet looked: `StageMessageSet_S02.tbl` (never resolved in
+`GP_MAIN_GAME_E.pak`), and the executable, whose decrypted image is only
+available in guest RAM.