REGATTAREPLAY
POST-RACE ANALYTICS · DRAFT v3e FORMAT
race f2a2cacc · generated 2026-07-02
04Leg by leg Δ vs leg best · green best · red worst
LEG 1 ↑ BEAT10:17.8 winning
| 181 | 298 | 441 |
|---|
| Time | 10:34.3+16.5 | 10:44.1+26.3 | 10:17.8 ★ |
| SOG kn | 5.31 | 5.48 | 5.65 |
| VMG kn | 3.74† | 3.79† | 4.02† |
| Dist nm | 0.94 | 0.98 | 0.97 |
| Heel ° | 13.1† | 13.4† | 12.0† |
441's pointing gain built the whole race lead — best SOG and best VMG on the beat.
LEG 2 ↓ RUN7:15.4 winning
| 181 | 298 | 441 |
|---|
| Time | 7:24.1+8.7 | 7:15.4 ★ | 7:25.2+9.8 |
| SOG kn | 6.17 | 6.32 | 6.42 |
| VMG kn | 4.06† | 4.15† | 4.01† |
| Dist nm | 0.76 | 0.76 | 0.79 |
| Heel ° | 5.2† | 4.8† | 5.6† |
441 sailed hot angles that didn't pay — fastest SOG, longest path, slowest run. 298's lower mode wins the leg.
LEG 3 ↑ BEAT10:05.5 winning
| 181 | 298 | 441 |
|---|
| Time | 10:15.8+10.3 | 10:05.5 ★ | 10:14.7+9.2 |
| SOG kn | 5.38 | 5.60 | 5.61 |
| VMG kn | 3.80† | 3.98† | 3.95† |
| Dist nm | 0.92 | 0.94 | 0.96 |
| Heel ° | 12.8† | 13.0† | 11.9† |
298 fastest on the second beat too — 441's lead down to 7.3 s at the top mark.
LEG 4 ↓ RUN8:02.7 winning
| 181 | 298 | 441 |
|---|
| Time | 8:06.2+3.5 | 8:02.7 ★ | 8:07.9+5.2 |
| SOG kn | 6.18 | 6.31 | 6.24 |
| VMG kn | 4.02† | 4.12† | 3.99† |
| Dist nm | 0.84 | 0.85 | 0.85 |
| Heel ° | 5.0† | 4.6† | 5.4† |
298 closes to 2.0 s at the finish — one more leg and the result flips.
DATA QUALITY
* The comparison stats show 0 tacks / 0 gybes while the maneuver engine detected 22 maneuvers — the stats panel is not wired to maneuver data (bug; see page 3 for the real counts). † Per-leg VMG, per-leg heel, side-bias %, and all Start values are illustrative placeholders. Race-scope Avg SOG and Distance are recomputed from leg data — the v2 race scope on staging appears to repeat Leg-1 values (verify in engine). Heel for 441 (8.6°) vs rivals (11+°) warrants a sensor-calibration check.
REGATTAREPLAY
POST-RACE ANALYTICS · DRAFT v3e FORMAT
race f2a2cacc · generated 2026-07-02
05Maneuvers 22 detected · loss = metres vs pre-maneuver VMG, to 95% recovery
| Metric | 181 Innocent B. | 298 Chimichuri | 441 Alchemist |
|---|
| Maneuvers T+G | 1 * | 13 4+9 | 8 4+4 |
| Total loss m | 13† | 182†≈ 61 s | 89†≈ 30 s |
| Avg tack loss m | 13† n=1 | 18† | 12† |
| Avg gybe loss m | — | 12† | 10† |
| Worst maneuver | 13 m11:27 | 28 m11:41 tack · stalled | 15 m11:26 tack |
| Avg recovery s | 6† | 7.1† | 5.6† |
Counts uncolored — more maneuvers is not worse without context; cost metrics carry the tri-color. J/80 benchmarks: tack good 8–12 m · typical 12–16 · poor 20+.
| TACKS · 9 fleet |
|---|
| Best | 298 · 10 m 11:28 · 5 s |
| Worst | 298 · 28 m 11:41 · 12 s · stalled |
| Avg | 441 12 m · 298 18 m |
| GYBES · 13 fleet |
|---|
| Best | 298 · 8 m 11:34 · 4 s |
| Worst | 298 · 17 m 11:52 |
| Avg | 441 10 m · 298 12 m |
MANEUVER READ
298 spent 182 m on maneuvers vs 441's 89 m — a ~93 m bill (≈ 31 s) — and still won Legs 2–4: at ~12 m each, five extra gybes cost ~60 m but bought the pressure that won two runs. Aggressive gybing was rational; 298's real leak is tack quality (18 m avg, one 28 m stall at 11:41 — the most expensive maneuver of the race). 441's edge: cheap, infrequent maneuvers (12 m tacks, top of the "good" band).
06Maneuver log grouped by leg · loss (m) + recovery (s) · green ≤12 m · red ≥20 m
| LEG 1 ↑ BEAT · 6 tacks |
|---|
| 11:25:37 | 298 | P→S | −22 m | 9 s |
| 11:26:40 | 441 | P→S | −15 m | 7 s |
| 11:27:03 | 181 | S→P | −13 m | 6 s |
| 11:28:03 | 298 | S→P | −10 m | 5 s |
| 11:29:41 | 298 | P→S | −14 m | 6 s |
| 11:32:40 | 441 | S→P | −11 m | 5 s |
| LEG 2 ↓ RUN · 4 gybes |
|---|
| 11:34:12 | 298 | S→P | −8 m | 4 s |
| 11:35:28 | 441 | S→P | −10 m | 4 s |
| 11:38:41 | 441 | P→S | −9 m | 4 s |
| 11:39:38 | 298 | P→S | −12 m | 5 s |
Legs 3–4 (12 more maneuvers) omitted in this draft — full log with entry/low/exit speeds and SOG-dip traces in the interactive viewer.
DATA QUALITY
† All loss/recovery values are illustrative — the engine must implement the integrated VMG-deficit computation (baseline = 20–30 s pre-maneuver VMG; integrate deficit until back to 95%; recipe and J/80 benchmarks in Research-Maneuver-Metrics-Perplexity-2026-07-02). The staging maneuver list renders boat names overlapping the Dir/In columns (rendering bug) and shows UTC instead of local time. 181 logged 1 maneuver in 4 legs — detection gap; excluded from coloring.