REGATTAREPLAY — DESIGN REVIEW HUB

END-TO-END REVIEW · LATEST LOCKED VERSION OF EVERY UI SURFACE · UPDATED 2026-07-02
BRAND LOCKED: SIGNAL AZURE #58B4FF · SAIRA CONDENSED / ARCHIVO / MARTIAN MONO · DARK MAPBOX

Viewer — current standard (v2)● LIVE STANDARD

Mobile viewer📱 PHONE FRAME
☰ command menu · Map/Data split spectrum (Map · 75/25 · 50/50 · Data) · contextual transport + Focus bar · decluttered edge-to-edge map.
/v2/phone-preview Desktop viewer🖥 OPEN ON DESKTOP
Fleet rail + Live/Stats/Maneuvers/Video panel, dark recolored basemap, FOCUS presets. (Same file responsive — on a phone it renders the mobile shell.)
/v2/fleet

Analytics standard (v6) — chronological + efficiency● IN REVIEW 2026-07-04

Race Analytics📱 MARK SEGMENTS
Leg N · Direction ↑↓ naming. 2-min checkpoint. WM1/WM2/Gate mark segments (entry tack, rounding order, position-change scaffold). Efficiency row per leg.
/v6/race.html Maneuver Analytics📱 SOG SPARKLINE
16-point stored SOG trace (−8 s → +30 s) with entry/low-point/recovery markers. MAJOR MOMENT trace in red. M2 maneuver consistency (fatigue analysis).
/v6/maneuvers.html Course / Side Bias📱 CHRONOLOGICAL
Chronological spine S→L1→WM1→L2→Gate→L3→WM2→L4→Wind→Setup. OVERALL efficiency block. Runs L2/L4 added. Position-only VMG + efficiency per leg (immune to N2).
/v6/course.html v6 index page↗ ALL THREE
v6 landing page. 181=0.715★ most efficient · 298=0.706 · 441=0.690 (won race). Efficiency vs elapsed is the coaching frame.
/versions/v6/

Analytics standard (v5) — engine-grounded● LOCKED 2026-07-03

Race Analytics📱 REAL G3 · QUALITY-GATED
v3c spine, grounded. Bug A/B killed by binding; Start block adds line-position + bias gems + SOG ±30 s; honest N1b "leg boundary uncertain" state. Quality-flagged metrics can't win/lose the row.
/v5/race.html Maneuver Analytics📱 REAL G3 · KB RECOVERY
Headline = KB-strict recovery seconds (SOG-robust → trusted on bad-heading boats). Real 441: 12–19s, 2 major-moment timeouts of 8. GPS-course badge mutes count/TWA. Detail keeps ALL fields.
/v5/maneuvers.html Course / Side Bias📱 REAL G3 GEOMETRY
Real cross-track: 181 Beat-2 = 93% left / 262m; all three entered WM from the right. Position-based → immune to N2. Honest pending states for derived-wind + gate-escape.
/v5/course.html v5 index page↗ ALL THREE
Landing page for the engine-grounded analytics standard. Supersedes v3/v4a where they conflict.
/versions/v5/

Analytics standard (v3) — superseded by v5○ LOCKED 2026-07-02

Race Statistics📱 SPEC + ANNOTATIONS
Comparison-scroll blocks: Race Overview (segment winners + LEAD progression) → Full Race (Elapsed first, x̄ column) → Start → Legs 1–4. Tri-color best/worst, ties neither.
/v3/analytics.html Maneuvers📱 SPEC + ANNOTATIONS
Overview → Tacks → Gybes → leg-grouped log. Headline metric: metres lost + recovery seconds (research-backed); J/80 benchmark color bands.
/v3/maneuvers.html Post-race report🖥 PRINT LAYOUT
The same blocks in a 3-page light print theme — this is also how desktop arranges the mobile blocks multi-column.
/v3/report.html Report — sample PDF⬇ PDF
Generated output for Race 1 (real staging data, race f2a2cacc): podium, overview + lead chart, full race, start, legs, maneuvers.
/v3/report-sample.pdf

Implementation — live app● CONVERGING TO SPEC

Staging viewer — Race 1📱 + 🖥
The running product (regatta-replay-v4, auto-deployed from main). Compare against the specs above; v2-port + QA Round 4 in flight.
staging.regattareplay.com/view/f2a2cacc… Staging — race list📱 STRESS TEST
Pick a large-fleet race for a density check: the design must hold at N=20+, not just 3 boats.
staging.regattareplay.com/races

Archive

v1 baselineFROZEN 2026-07-01
Locked-azure baseline the fidelity audit ran against. Kept for comparison.
/v1/fleet · /v1/phone-preview v3 index page
The analytics-standard landing page with spec notes.
/versions/v3/
How to review: open the 📱 links on a phone (or 390px devtools) and 🖥 links on a laptop. The annotation columns on the v3 pages are part of the spec — they state the rules (labels, tri-coloring, benchmarks, no hidden scroll) the implementation must follow. Numbers marked † are illustrative pending engine work; everything else is real Race 1 data.
TarioGroup/regatta-viewer-concept · gated preview · VERSIONS.md is the changelog · analytics research: vault → Research-Maneuver-Metrics-Perplexity-2026-07-02