ANALYTICS V6 — COURSE · SIDE BIAS · CHRONOLOGICAL

Chronological spine · position-only efficiency (rhumb/path) · runs added · mark segments · real G3 RCYC 6/15 · f2a2cacc · all boats POSITION-BASED (immune to N2)
RCYC · R1 JUN 15 '25 3/3
COURSEv6 · chronological · efficiency
S ↑ L1 WM1▲ ↓ L2 G▼▼ ↑ L3 WM2▲ ↓ L4 F
OverallStartL1 ↑L2 ↓L3 ↑L4 ↓Wind
181Innocent B.POSITION 298ChimichuriPOSITION 441AlchemistPOSITION
OVERALLefficiency across 4 legs
Mean efficiency rhumb/path0.715 0.7060.690
Most efficient legs4 / 40 / 40 / 4
Race finishP3P2P1
The efficiency signal separates straightness from speed. CAN181 is the most efficient boat on every leg (4/4) yet finished last. CAN441 won the race (P1) despite being the least efficient on every leg — higher boatspeed compensated. This axis is invisible to elapsed time alone.
STARTgun + 2-min window
2-min cross-track window: pending scaffold. Position-only cross-track at T+2:00: what side had each boat committed to? Same GPS-track pipeline as side-bias on beats. SOG ±30 s from v5 is the base; this block adds the side read at T+120 s. binding: trackAt(T+120s) → crossTrackSigned → eL/eR chip · same pipeline as LegMetrics.sidePercent
LEG 1 · UPWINDreal · clean roundings
Side of course % timeL 59%L 61%L 57%
Mean offset m from CL65 L59 L84 L
Vs fleetMIDRIGHTMOSTLEFTMOST
Tacks112
VMG kn · pos2.722.702.78
Efficiency rhumb/path0.522 0.5100.510
Leg time+21.0+29.715:21.0
EFFICIENCY · RHUMB / PATH ↑ = STRAIGHTER LINE
181
0.522★
298
0.510
441
0.510
441 won L1 (best elapsed) and was the leftmost boat (84 m mean) — the whole fleet worked left, but 441's higher boatspeed more than offset the extra distance sailed. 181 was the most efficient (0.522★) but not fastest. VMG ranks identically to elapsed (constant rhumb) so efficiency is the genuinely new per-boat axis.
WM1 ▲approach · rounding · position
Entry tack last 200 mRRR
Rounding orderP2P3P1
Position change (200 m before / after): pending. Who moved up or down at WM1? Rank delta at −200 m and +200 m from the mark. binding: markRoundingPositionDelta(WM1, −200m, +200m) → rank change Δ
LEG 2 · DOWNWINDrun · real efficiency
Side of course—†—†—†
Gybes242
VMG kn · pos5.325.38 5.31
Efficiency rhumb/path0.864 0.8560.829
Leg time+5.47:14.0 +6.0
EFFICIENCY · RHUMB / PATH
181
0.864★
298
0.856
441
0.829
Run-side (port / starboard): pending calibration. Cross-track algorithm validated for upwind legs. Downwind sign convention calibration in progress. binding: SidePercent module · downwind sign convention calibration pass pending
181 was most efficient on L2 (0.864★) yet finished +5.4 s behind 298. 298 won with most gybes (4) and highest VMG — routing paid. 441's L2 MAJOR MOMENT gybe (see Maneuvers) explains its 0.829 worst efficiency.
LM1 GATE ▼▼choice · escape
Gate rounded L/RLRL
Escape timing + position change: pending. Gate choice available; centerline-crossing escape seconds still refining at the rounding baseline. GateChoice escape refinement + markRoundingPositionDelta(LM1)
L/R looking downwind · † interpolated pending refinement
LEG 3 · UPWINDreal · headline beat
Side of course % timeL 93%R 56%L 62%
Mean offset m from CL262 L25 R70 L
Vs fleetLEFTMOSTRIGHTMOSTMID
Tacks122
VMG kn · pos3.763.84 3.77
Efficiency rhumb/path0.702 0.6870.675
Leg time+12.010:08.8 +10.7
EFFICIENCY · RHUMB / PATH
181
0.702★
298
0.687
441
0.675
181 committed hard left (93%, 262 m mean) and paid the most (+12.0 s). 298 was the only boat right of centerline and won the leg. Efficiency gap 181–441: 0.702 vs 0.675 — the largest of any leg. This is the coaching notebook story in pure position data.
WM2 ▲approach · rounding · position
Entry tack last 200 mRRR
Rounding orderP3P1 P2
Position change (200 m before / after): pending. markRoundingPositionDelta(WM2)
LEG 4 · DOWNWINDrun · real efficiency
Side of course—†—†—†
Gybes232
VMG kn · pos4.58 4.514.41
Efficiency rhumb/path0.773 0.7720.744
Leg time9:21.6 +8.7+21.4
EFFICIENCY · RHUMB / PATH
181
0.773★
298
0.772
441
0.744
181 swept L4 — best VMG, best efficiency, best time. Efficiency gap 181 vs 298 is near-zero (0.773 vs 0.772); 441 trails at 0.744, mirroring the L4 major-moment state on Maneuvers (+21.4 s deficit). The L1 gap (0:29) was larger than L4's recovery range.
WINDcourse orientation
RC set068° courseSetup.orientation — real
Fleet-sailed wind: computed, not yet served. Engine derives fleet consensus wind (beat-angle bisection + confidence) — E-C1 read-through hasn't shipped. FleetConsensus.windDirection + confidence · E-C1 = S (read-through)
COURSE SETUPrhumb geometry
Leg 1 rhumb m1 319
Leg 2 rhumb m1 202
Leg 3 rhumb m1 202
Leg 4 rhumb m1 324
rhumb = CourseLeg.distance (mark-to-mark haversine) · constant across boats · denominator in efficiency · numerator in leg VMG
Line square / course length / setup stats: pending E-C series. courseSetup.{startLineLength, startLineBias, courseSetupStats} · E-C series
iPhone · 390pt · full scroll · REAL G3 · position-only efficiency

WHAT'S NEW IN V6 — COURSE PAGE

  • Chronological order, not wind-first. v5 led with "What the wind really was." v6 follows the race: Start → L1 → WM1 → L2 → Gate → L3 → WM2 → L4 → Wind → Setup. The scrollable spine at the top makes the structure scannable at a glance.
  • OVERALL efficiency block (new). Mean rhumb/path per boat across all 4 legs. 181 = 0.715★ (most efficient, every leg). 441 = 0.690 (least efficient, every leg). Yet 441 won. The efficiency vs elapsed tension is the coaching frame. FINDINGS-Per-Leg-VMG §5
  • Position-only VMG + efficiency per leg (new). leg_VMG = rhumb/elapsed (kn); efficiency = rhumb/path (0–1). Zero heading, zero wind — immune to N2 on all three boats. Efficiency ranking differs from elapsed on 3/4 legs. Note: VMG ranking ≡ elapsed (constant rhumb) so efficiency is the genuinely new axis. Proposed E-G under bundle:R-analytics-math.
  • Runs added (L2, L4). v5/course.html showed beats only. v6 adds both downwind legs with VMG, efficiency, gybe count, and leg time. Run-side (port/starboard) is an honest pending block.
  • Mark segments (WM1, WM2, LM1 Gate). Between-leg blocks show entry tack (last 200 m) and rounding order. Position change (±200 m window) scaffolded as a pending block.
  • Tack/gybe count per leg. GPS-course boats (181/298) render with ·gps treatment on counts; 441's counts are precise.

DATA STATUS

  • All side-bias, VMG, and efficiency numbers are real engine + corpus data (G3 f2a2cacc). Gate choices marked † pending rounding refinement. Run-side (—†) pending downwind calibration. Mean efficiency = (L1+L2+L3+L4)/4 per boat.