Skip to content

Newport to Bermuda Race — Historical Race Database

Rev: v2.0  ·  Updated: 2026-05-29

What each race in our 6-year set actually did — the weather pattern, the Gulf Stream structure, the winning tactic, and (where we can measure it) how the corridor choice paid off. The spine is the corridor-median table: per-year median elapsed time for the strict 40–45 ft cohort, split by which side of the rhumb each boat sailed. All other figures — winner names, elapsed/corrected times, fleet counts, and every wind-speed and current figure in the per-year prose — come from media/official accounts and are marked carried, unverified — they are context, not measured signal.

TL;DR

  • Only one corridor split in the whole set is statistically real: 2022 EAST (n=29). Every other year's EAST corridor is n=1–2 — a single boat, not a trend. Read those rows as anecdote.
  • Corridor medians are within-year only. Each year had a different wind regime, so a 67 h median (2024) and a 124 h median (2014) are not a speed comparison — they say which side paid that year.
  • The recurring lesson is the Stream, not the side. Across years the decisive move was reading the meander/ring correctly and exiting at the right latitude — chasing a model-only current signal lost time (2016).

Master summary

Winner/elapsed/fleet figures below are carried, unverified — sourced from bermudarace.com results and media accounts (see Sources), not from any script in this repo. Weather codes link the W-* taxonomy; tactic codes link the Routing Tactic Catalog; year links go to the Winner Route Summary.

Year Cohort (n) Pattern Corrected winner (SDL) Gulf Stream Winner tactic
2024 64 W-PF / W-BH / W-RC Carina (Rives Potts) Up to 5 kt; southerly meander (not a CCR) T-PF + T-ME + T-CL
2022 67 W-PF / W-SQ Illusion (Cal 40, S. & S. Honey) Chaotic — forming cold-core ring (oxbow) T-PF + T-EC + T-ME
2018 54 W-LA / W-PL / W-BH Grundoon (Columbia 50, Jim Grundy) Normal; current exploitation decisive T-CF + T-EC
2016 67 W-PF / W-PL / W-CC Warrior Won (Xp44, C. Sheehan) CCR stationary W of rhumb — trap T-NW (avoided ring)
2014* 70 W-LA / W-PF / W-RC Actaea (Hinckley B40, M. Cone) Up to 5 kt; "exit at the right moment" T-PF + T-CF
2012 71 W-BH / W-PF Carina (S&S 48, Rives Potts) Normal — eddy gap near rhumb T-RH + T-CF

*2014 is the held-out seed/check year. Cohort = strict 40–45 ft LOA group on the YB tracker. The Newport Bermuda Race is biennial (even years); 2020 was cancelled (COVID-19).

Corridor medians (verified)

Median elapsed time in hours for the strict 40–45 ft cohort, by side of rhumb, from scripts/analysis/stream_corridor_analysis.py over YB tracker data 2012–2024. (n) is boats in that corridor — read it before reading the median.

Year EAST (n) RHUMB (n) WEST (n)
2024 86.44 (1) 67.72 (52) 67.20 (3)
2022 59.45 (29) 60.63 (30) 75.10 (3)
2018 95.44 (2) 93.52 (39) 97.57 (7)
2016 77.25 (1) 78.74 (21) 83.68 (23)
2014* — (0) 124.04 (28) 123.99 (37)
2012 — (0) 87.84 (47) 84.88 (6)

What the numbers actually support:

  • 2022 EAST 59.45 h is the lowest median in the entire set ("fastest in set") and the only EAST corridor with a real sample (n=29). West cost ~15.7 h vs east (75.10 − 59.45) — but only because SST + AVISO altimetry confirmed the eddy structure first; never trust an RTOFS-only signal here.
  • 2016: chasing the west corridor cost ~5 h vs the rhumb (83.68 − 78.74 = +4.94 h). That west bias was the CCR trap — boats dove for a model-predicted favorable ring that did not deliver.
  • 2018: west cost ~4 h vs rhumb (+4.05 h) in the light-air parking-lot year — corridor barely mattered, finding the Stream core did.
  • 2012: west was marginally faster (−2.96 h vs rhumb) but on n=6 — not actionable.
  • 2024 & 2014: RHUMB dominated the cohort (n=52, n=28). EAST/WEST samples are too thin (or zero) to read.

Do not compare medians across rows. Different years, different wind regimes. The columns answer "which side paid that year," not "which year was faster."

Per-year detail

2024 — Reaching race, strong current (W-RC)

Front north of the start made a fluky, slow first day; after it cleared, 14–24 kt SW reaching for most of the fleet. A southerly Gulf Stream meander (not a CCR) delivered up to 5 kt NE current. The tactical tension was staying in the current pocket — which meant going east of rhumb — then exiting before being set east and forced upwind to Bermuda. Carina's navigator credited an early Rhode Island coast departure (south of the frontal light patch) as the first key call, Stream exploitation as the second. Cohort median was a tight RHUMB story (67.72 h, n=52).

Carried, unverified (bermudarace.com / media): Pyewacket (Volvo 70) line honors ~59 h 18 m; Carina corrected ~64 h 12 m for her 5th NB win; Hound 2nd at 64 h 25 m incl. a 30-min OCS penalty (actual ~63 h 55 m — start execution decided the overall). Fleet 162 started, 96 in SDL. Safety note: J/122 Alliance was lost after striking a submerged object ~0300 June 23; all crew rescued by J/122 Ceilidh — a collision, not a design defect.

2022 — Post-frontal squalls, forming ring (W-PF / W-SQ)

Front crossed the first night (June 17–18): rough seas, heavy rain, then post-frontal NW ~20 kt that suited the leaders. A developing oxbow meander / forming cold-core ring put adverse current on the wrong side; squalls hit 30+ kt in the Stream. Pre-race analysis (Bohlen, May 2022) flagged the oxbow. The winning move (Illusion, Stan Honey navigating) was to hold a favorable eddy pocket ~7 h (T-ME) and exit at the right position. This is the year the EAST corridor genuinely paid (59.45 h, n=29) — the one statistically meaningful split in the set.

Carried, unverified: Argo (MOD70) line/record ~33 h; Warrior Won (Pac52) monohull line ~56 h 44 m; Illusion (Cal 40) SDL corrected ~51 h 02 m over a 108-boat division (J/122 Zig Zag 2nd, ~2 h back).

2018 — Parking lot (W-LA / W-PL)

A blocking high parked across the rhumb south of the Stream — extreme light air and calms in the middle third. Fast boats stretched a lead, then parked; the race effectively re-started for lighter boats when new breeze filled. Gulf Stream current (up to 4–5 kt in the core) was the differentiator — "the trick was finding the entry point and knowing when to get off the ride." Light air collapses corridor differences (west cost only ~4 h vs rhumb), so finding the core mattered more than the side.

Carried, unverified: Rambler 88 line ~50 h 32 m; Grundoon (Columbia 50) SDL corrected ~112 h 12 m (the huge number reflects the calms); J/121 Apollo won its class and was the best Lupo-size comparable that year. SDL ~85 entries.

2016 — Cold-core ring trap (W-CC)

Pre-race forecasters predicted a dangerous low near Cape Hatteras; seven Gibbs Hill boats withdrew before the start — actual conditions were milder. NE winds mostly under 25 kt. A stationary cold-core ring sat just west of the rhumb (tracked from March 2016, moved ~60 nm west by race week but stayed near the rhumb). Boats that dove in to exploit a predicted favorable current were set back; the ring did not deliver, and the skeptics won. The corridor numbers show it: the west bias cost ~5 h vs the rhumb (83.68 vs 78.74). Comanche (line-honors record) found the gap in the high and went around its edge; Warrior Won won SDL by avoiding the ring (T-NW / CCR-avoid). This is the canonical trust three sources before chasing a ring lesson.

Carried, unverified: Comanche line/record ~34 h 53 m (broke Rambler's 2012 mark by ~4 h 46 m); Warrior Won (Xp44) SDL corrected ~70 h 40 m over 70 boats (Cal 40 Flyer 2nd, ~53 min back). Safety note: J/122 Alliance suffered water ingress and was abandoned (a separate incident from 2024).

2014 — Light air, frontal zone, strong current (W-LA / W-RC) — seed year

"One of the slowest races in recent history" — light air and "airless sinkholes," a frontal zone that shut down the east side of the course. Conservative west for pressure was the right call, then exploit up to ~5 kt of favorable Stream and exit before being set east (same exit lesson as 2024). In the cohort, WEST and RHUMB medians are essentially tied (123.99 vs 124.04 h) on large samples — corridor was a wash; pressure and Stream-exit timing decided it. Held out of the analog matcher as the seed/check year.

Carried, unverified: Shockwave (R/P Mini Maxi) line ~63 h 04 m; Actaea (Hinckley B40 yawl) SDL by a ~45-min margin. No J/122 cohort results confirmed in available sources. ~164 starters.

2012 — Fast reaching race (W-BH / W-PF)

Post-frontal cool, dry NW/N air gave the whole fleet a long port-tack reach; big boats averaged 16+ kt. The Stream played a secondary role — Carina threaded a narrow gap between an eddy and the rhumb. Mid-fleet saw the northerly moderate near Bermuda ("Happy Valley") with a developing squall line on approach. The race was about sailing angles, not the corridor: west was marginally faster (−3 h vs rhumb) but on only n=6.

Carried, unverified: Rambler (90 ft) line/record ~39 h 39 m; Carina (S&S 48) SDL corrected ~76 h for her 2nd straight Lighthouse Trophy — famously only ~2 jibes and 2 tacks in 635 miles. 48th running; ~165 started.

Reading this as a Lupo crew

Lupo Di Mare (Italia 12.98) has no entry in any 2012–2024 fleet. Benchmark against the strict 40–45 ft cohort (the medians above), refined to ±5% TCC once the ORR cert is confirmed. Sister hull Italia 11.98 raced as CHRISTOPHER DRAGON in 2022 and 2024 — treat as a single data point, not a prediction (different crew, sails, rating). See Historical Fleet Tracks for the per-year cohort maps.

Do not compare Lupo to Comanche, Rambler, Argo, Pyewacket, or any ORMA/MOD/pro maxi — their tactical conclusions do not transfer. The Cal 40s, Xp44, J/12x, and Columbia/Hinckley winners above are the relevant analog band.

Caveats

  • EAST is anecdote except 2022. n=1–2 in every other year. Do not generalize from a single boat's corridor.
  • Within-year medians only — never cross-year speed comparisons.
  • Winner/elapsed/fleet figures and all wind/current values are carried, unverified — confirm against the official bermudarace.com ORR results PDFs and archived weather data before citing as fact.
  • Corridor and cohort numbers are the verified layer — they come from the YB tracker pipeline and match the strict 40–45 ft cohort sizes (2024=64, 2022=67, 2018=54, 2016=67, 2014=70, 2012=71).

Sources

  • Corridor medians & cohort sizes — scripts/analysis/stream_corridor_analysis.py over YB tracker binaries 2012–2024 (strict 40–45 ft cohort); cached data/yb_cache/stream_corridor_stats.json.
  • Winner tactics & per-period encoding — winner_route_summary.md; tactic codes routing_tactic_catalog.md.
  • Weather pattern taxonomy — weather_analog_framework.md.
  • Gulf Stream structure (rings / meanders / RTOFS) — gulf_stream_framework.md (#ccr, #wce, #rtofs, #crossing-strategies).
  • Cohort track maps — historical_fleet_tracks.md.
  • Field-definition schema (archived Wave 1) — archive/05_historical_races/historical_schema.md.
  • Race results & accounts (carried, unverified) — bermudarace.com official results/press releases; Scuttlebutt, Sailing World, Cruising World, North Sails, J World race reports; Bohlen Gulf Stream pre-race notes (2016, 2022).

PRE-RACE RESEARCH — not race-period routing advice. Updated 2026-05-29.