diff --git a/Sleep/2026-08-09.md b/Sleep/2026-08-09.md
new file mode 100644
index 0000000..fb142a9
--- /dev/null
+++ b/Sleep/2026-08-09.md
@@ -0,0 +1,49 @@
+---
+date: 2026-08-09
+type: sleep
+tags: [sleep, alcohol, short-sleep, hrv-low]
+---
+
+# Sleep — 2026-08-09
+
+**4h 33min** · 05:12→09:45 · REM: 24% · Deep: 29% · HRV: **29ms** 🔴 · Resting HR: 72 bpm · Score: 74/100
+
+> The 5am drinking night (Sat night → Sun morning). Bed at 05:12, up at 09:45 — 4h33min, less than half the reference range. **HRV at 29ms is the lowest ever recorded — by a wide margin** (previous worst: 71ms on the Jul 8 heavy drinking night). Resting HR at 72 bpm is 23 bpm above the 48-59 personal range. This is the classic alcohol-hangover cardiovascular signature: sympathetic overdrive, body still processing the night's intake through the measured window.
+>
+> Paradoxically, the sleep *structure* for the hours logged was normal: REM 24% (1h 5min), deep 29% (1h 20min), light 47%, 0 wake-ups, breathing quality 100, SpO2 98%. The body packed a structurally fine 4.5h — there just wasn't nearly enough of it, and the cardiovascular system was clearly hammered (HRV 29ms, HR 72).
+>
+> Score 74 / better than 33% — the watch called it below-average, and the HRV/HR pair is the loudest warning signal in Gutca's entire sleep history. This is what a 5am drinking night costs.
+
+## Details
+
+| Metric | Value | Reference |
+|--------|-------|-----------|
+| Night sleep | 4h 33min | 6-10h (Below average) |
+| Bed → Rise | 05:12 → 09:45 | — |
+| Deep sleep | 29% (1h 20min) | 20-60% |
+| Light sleep | 47% (2h 8min) | <55% |
+| REM sleep | 24% (1h 5min) | 10-30% |
+| Deep continuity | 80 points | 70-100 |
+| Times woke | 0 | 0-1 |
+| Breathing quality | 100 points | 70-100 |
+| HRV | **29ms** 🔴 | 77-135ms (Far below range) |
+| Resting HR | **72 bpm** 🔴 | 48-59 bpm (Above range) |
+| SpO2 | 98% | 96-99% |
+| Respiratory rate | 16 brpm | 12-16 brpm |
+| Score | 74 / 100 | Better than 33% |
+
+### HRV progression
+
+| Date | HRV | Context |
+|------|-----|---------|
+| Jul 5 | 135ms | Sober, clean night |
+| Jul 20 | 107ms | Heavy weed |
+| Jul 23 | 111ms | Post-fröccs |
+| Jul 27 | 135ms | First sober night post-festival |
+| Jul 29 | 140ms | Light dose, all-time high |
+| **Aug 9** | **29ms** 🔴 | 5am drinking night — all-time low |
+| Aug 10 | 139ms | Recovery night after |
+
+## Connected
+
+- [[Sleep/2026-08-10]]
diff --git a/Sleep/2026-08-10.md b/Sleep/2026-08-10.md
index 25bcb7a..cbf8975 100644
--- a/Sleep/2026-08-10.md
+++ b/Sleep/2026-08-10.md
@@ -8,11 +8,11 @@ tags: [sleep, recovery, naps]
 
 **8h 32min** · 00:24→09:19 · REM: 7% · Deep: 38% · HRV: **139ms** 🟢 · Resting HR: 49 bpm · Score: 87/100
 
-> Recovery night after a brutal weekend. Saturday night was out until 5am — but on **only 0% beers** (Gutca voiced concern about drinking nights being too common, then ran the experiment: no alcohol, same schedule). Sunday was a write-off on 4-5h sleep. This night: 8h32min, deep sleep at 38% (3h 15min) — the body paid back physical debt hard. HRV at **139ms is the second-highest ever recorded** (140ms all-time high on Jul 29), sitting *above* the personal range of 78-136ms — parasympathetic rebound after sleep deprivation, with no alcohol in the system to suppress it.
+> Recovery night after a brutal weekend. The 5am drinking night (see [[Sleep/2026-08-09]]) crushed HRV to **29ms — the lowest ever recorded** — with resting HR at 72 bpm. This night the body rebounded hard: 8h32min, deep sleep at 38% (3h 15min) — physical debt paid back. HRV at **139ms is the second-highest ever recorded** (140ms all-time high on Jul 29), sitting *above* the personal range of 78-136ms — full parasympathetic rebound once the system was clean.
 >
 > REM is the odd one out: only 7% (33 min). Classic post-deprivation pattern — the body prioritizes deep sleep first, and REM debt gets paid back *during the day*. Gutca reported napping a lot, struggling to wake, and dreaming heavily at the end of naps (couldn't recall the dreams) — that's daytime REM rebound, not a bug. Sleep inertia on top of a 09:19 wake made the morning feel heavier than the numbers suggest.
 >
-> **The cleanest takeaway:** the wrecking ball wasn't the alcohol — it was the 5am schedule. Same late night, zero booze, same wrecked Sunday. Worth honoring that out-loud observation: the *frequency* of drinking/late nights is the real pattern to address, not just the substance.
+> **The cleanest takeaway — the contrast is everything:** drinking night (Aug 9) = HRV 29ms · HR 72 bpm · 4h33. Sober night (Aug 10) = HRV 139ms · HR 49 bpm · 8h32. That gap is the measured cost of a 5am drinking night — the worst recovery signal in Gutca's entire logged history. The good news: one clean night restored HRV to near-record levels, so recovery capacity is fully intact. The frequency of drinking/late nights is the pattern to cut, not the substance alone.
 
 ## Details
 
@@ -42,8 +42,10 @@ tags: [sleep, recovery, naps]
 | Jul 23 | 111ms | Post-fröccs |
 | Jul 27 | 135ms | First sober night post-festival |
 | Jul 29 | 140ms | Light dose, all-time high |
-| **Aug 10** | **139ms** 🟢 | Recovery night after 5am 0%-beer night, above range |
+| **Aug 9** | **29ms** 🔴 | 5am drinking night — all-time low |
+| **Aug 10** | **139ms** 🟢 | Recovery night after, above range |
 
 ## Connected
 
+- [[Sleep/2026-08-09]]
 - [[Runs/2026-07-29]]