diff --git a/docs/FLOOD_FORECASTING.md b/docs/FLOOD_FORECASTING.md
index 3774614..9ae4ebc 100644
--- a/docs/FLOOD_FORECASTING.md
+++ b/docs/FLOOD_FORECASTING.md
@@ -291,6 +291,21 @@ training half. So the model was retrained on data **ending 2024-08-31** and aske
 to forecast September–November 2024 cold, with no knowledge of the event that
 followed. This is the closest thing to a real operational test available.
 
+![Observed P.1 level vs the model's 24 h-ahead predicted peak through the
+October 2024 flood, with the warning probability below](img/backtest-2024-p1.png)
+
+The render above shows the whole event hour by hour. Top: the observed level
+(blue) against the 24 h-ahead predicted peak the model issued at each hour
+(amber, dashed) — the amber line leads the blue one into both flood waves,
+which is the entire point of the system. Bottom: the model's probability of
+flooding within 24 h; it fires its **first alert at 01:00 on 24 September, a
+full day before the river crossed the 3.70 m flooding line**, stays pinned near
+1.0 through both waves, and stands down between and after them. Also visible,
+honestly: the predicted peak tops out ~0.4 m short of the actual 5.30 m record
+(the extreme-compression limitation discussed below), and the prediction is
+noisier on the recession limbs. The same model track drives the dashboard's
+"Replay Oct 2024 flood" feature, so this chart can be watched live on the map.
+
 - **25 September cold start.** P.1's first warning crossing of the episode was
   alerted **24–26 hours ahead**. This is the genuinely impressive case: the river
   was in normal state, and the alert came from upstream routing alone.
diff --git a/docs/img/backtest-2024-p1.png b/docs/img/backtest-2024-p1.png
new file mode 100644
index 0000000..610544d
Binary files /dev/null and b/docs/img/backtest-2024-p1.png differ
diff --git a/src/static/dashboard.html b/src/static/dashboard.html
index de5500b..d2401a2 100644
--- a/src/static/dashboard.html
+++ b/src/static/dashboard.html
@@ -870,7 +870,12 @@
                     alertEl.textContent = pWarn == null ? '' : 'model: no flood expected';
                     alertEl.className = 'replay-alert note';
                 }
-                i += 2;                                 // 2 h per frame (~40 s total)
+                // Variable speed: race through quiet days, slow to hour-by-hour while
+                // the model is detecting and the flood unfolds, so the alert sequence
+                // (elevated risk -> model alert -> flooding) is readable.
+                const critical = (pWarn != null && pWarn >= 0.5) || (state.p1Now != null && state.p1Now >= 3.5);
+                const elevated = (pWarn != null && pWarn >= 0.2) || (state.p1Now != null && state.p1Now >= 2.8);
+                i += critical ? 1 : elevated ? 2 : 4;
                 if (i >= n + 40) endReplay();           // hold the last frame briefly
             }, 120);
         } catch (error) {