ANALYTICS V3e — MANEUVERS · RESEARCH-BACKED METRICS

Headline metric per research: metres lost (integrated VMG deficit vs 20–30 s pre-maneuver baseline, to 95% recovery) + recovery seconds · dip-% demoted to detail · REVIEW DRAFT — not pushed
‹RACE 1 JUN 15 3/3
MANEUVERSscrolls by block
OverviewTacksGybesDetail⇅ By leg
181Innocent B. 298Chimichuri 441Alchemist
MANEUVER OVERVIEWrace · 22 detected
Maneuvers T+G1 *13 4+98 4+4
Total loss m13†182†≈ 61 s89†≈ 30 s
Avg tack loss m13† n=118†12†
Avg gybe loss m—12†10†
Worst 13 m11:27 28 m11:41 tack 15 m11:26 tack
Avg recovery s6†7.1†5.6†
loss = metres given up vs pre-maneuver VMG, until back to 95% · J/80 tack: good 8–12 m · typical 12–16 · poor 20+
298 spent 182 m on maneuvers vs 441's 89 m — a ~93 m bill (≈ 31 s) — and still won Legs 2–4: their lane and pressure gains outweighed it. 441's edge is cheap, infrequent maneuvers.
TACKSbeats · 9 fleet · good ≤12 m
Count1 *44
Avg loss m13† n=118†12†
Avg recovery s6†8†5.5†
Best 13 m11:27 10 m11:28 11 m11:32
Worst 13 m11:27 28 m11:41 · stalled 15 m11:26 · slow exit
Retention x̄ low÷in70%63%†68%†
441 averages 12 m per tack (top of "good" band) vs 298's 18 m — ≈ 6 m per tack, ≈ 24 m over the race's beats. 298's 11:41 stall (28 m, 12 s recovery) was the most expensive maneuver of the race.
GYBESruns · 13 fleet · good ≤12 m
Count0 *94
Avg loss m—12†10†
Avg recovery s—5†4†
Best — 8 m11:34 9 m11:38
Worst — 17 m11:52 11 m11:54
Gybes are cheap for both boats (10–12 m) — which is exactly why 298's 9-gybe strategy worked: at ~12 m each, five extra gybes cost ~60 m but bought the pressure that won two runs by 15 s (~45 m). Aggressive gybing was rational.
ALL MANEUVERSgrouped by leg · loss + recovery
LEG 1 ↑ BEAT · 6 TACKS
11:25:37298⤸ P→S−22 m9 s
11:26:40441⤸ P→S−15 m7 s
SOG THROUGH MANEUVER · ±30 s · LOSS = SHADED VMG DEFICIT
3.5 kn baseline VMG
in 5.7 → low 3.5 → out 5.5 kn · retention 61% · recovered in 7 sslow exit†
11:27:03181⤹ S→P−13 m6 s
11:28:03298⤹ S→P−10 m5 s
11:29:41298⤸ P→S−14 m6 s
11:32:40441⤹ S→P−11 m5 s
LEG 2 ↓ RUN · 4 GYBES
11:34:12298⤾ S→P−8 m4 s
11:35:28441⤾ S→P−10 m4 s
11:38:41441⤿ P→S−9 m4 s
11:39:38298⤿ P→S−12 m5 s
LEG 3 ↑ · LEG 4 ↓ — 12 more maneuvers · tap to expand
Data quality: † all loss/recovery values illustrative — engine must implement the integrated VMG-deficit computation (recipe in the research note). Loss color bands: green ≤ 12 m · plain 12–16 · red ≥ 20 (J/80 benchmarks). 181: 1 maneuver in 4 legs — detection gap, excluded from coloring.
☰ MENU
Map◫Data
v0.43
iPhone · 390pt · full scroll

WHAT THE RESEARCH CHANGED (V3d → V3e)

  • Headline metric is now metres lost per maneuver — the integrated VMG deficit vs a 20–30 s pre-maneuver baseline, counted until VMG is back to 95%. This is what Vakaros Atlas 2 ("how many meters were lost… after the maneuver is completed") and Expedition ("Tack/gybe loss metres") surface as first-class. It captures your speed-in/speed-out/loss intuition in one number a racer feels: "that tack cost 2 boat lengths."
  • Secondary: recovery seconds (time back to 95% of baseline). Depth × duration = distance lost; the pair covers both failure modes (deep dip vs slow rebuild).
  • The old efficiency % (low ÷ entry) is demoted to a muted "Retention x̄" row and the expanded detail — research: racers respond to metres, "percent speed retained… demands more cognitive effort."
  • Benchmark color bands from J/80 values: tacks good 8–12 m / typical 12–16 / poor 20+; recovery good 5–7 s. Loss numbers in the log are colored against the band, not against the fleet.
  • Phase labels, not phase numbers: worst maneuvers get a one-word diagnosis ("stalled", "slow exit") derived from the trace — no third metric column.
  • Strategy insight the new metric unlocks: 298's five extra gybes cost ~60 m but won two runs by ~45 m-equivalent — aggressive gybing was rational; their real leak was tack quality (18 m avg, one 28 m stall).
  • Full research (18 sources, computation recipe, benchmarks): Research-Maneuver-Metrics-Perplexity-2026-07-02 in this project folder.