Skip to content

Navigator Walkthrough - How to Use This Site

Rev: v3.0  ·  Updated: 2026-05-29

Navigator of Lupo Di Mare — what to read, when, and what to do with it, from T-10 to the finish.

TL;DR

  • The spine is one command: .venv/bin/python scripts/daily_brief.py runs fetch → classify → match → brief and writes a dated brief to data/daily-brief/. Run it daily from T-10. It automates the GFS/ECMWF/ICON pull, the W-* pattern call, and the analog ranking — see Pre-Race Data Workflow.
  • What stays manual is the judgement layer: Gulf Stream imagery (RTOFS/SST/AVISO), Expedition routing runs, and the per-gate decisions below.
  • Protect the three gates (Stream-entry sail, current confirmation, model agreement) and pull the right onboard card for each phase.
  • Rule line: all AI/analog analysis is pre-race only — no shore routing advice once the gun fires (RRS 41 modification).

Quick reference — what to open when

Situation Open this
Just assigned as navigator Crew Orientation, then Navigator Playbook
Daily forecast pull (T-10 on) Run daily_brief.py; method in Forecast Analog Matcher
Classifying the pattern by hand Weather Analog Framework
Pattern → tactic shortlist Model → Tactic Synthesis
Detailed tactic profiles Routing Tactic Catalog
Stream analysis (T-7→T-5) Gulf Stream Framework + Crossing Strategies
Cold-core ring confirmation CCR Guide
Warm-core eddy concern WCE Guide
Pre-race crew brief Crew Orientation + all 10 onboard cards
Approaching the Stream Card 03, then Card 04
Squall developing Card 09
Something unexpected Navigator Playbook → Failure modes

Three gates to protect

Gate Where Rule
1 — Stream-entry sail Card 03, T-2h before north wall Entry sail locked 2 h before the SST gradient — fix it in the 50-nm approach window. A peel inside the Stream is skipper-cleared, not a watch default.
2 — Current confirmation Card 04, at the north wall Do not commit T-CCR or T-ME on RTOFS-only data. SST + altimetry must confirm. If not: default T-RH, brief Skipper.
3 — Model agreement Before any major routing call GFS vs ECMWF arrival Δ >6 h → HIGH UNCERTAINTY. No high-commitment call (meander entry, ring exploit) until they converge.

The one-command spine

.venv/bin/python scripts/daily_brief.py        # fetch → classify → match → brief

Writes a dated brief to data/daily-brief/. The steps below that used to say "pull GFS/ECMWF/ICON, classify the pattern, rank the analog" are now this command — the brief's trend block tracks model agreement day-over-day (it replaced the old hand-kept comparison log). Read the Pre-Race Data Workflow for the full day-by-day procedure; this page is the navigator's reading/decision overlay on top of it.


Pre-Race: T-10 (Jun 9) → T-0 (Jun 19)

Gate Run / read Decision output
T-14 (Jun 5) — foundation Crew Orientation, Navigator Playbook, Polars (plan at S3 / 10% degradation), Sail Inventory Solution space defined before Expedition opens
T-10 (Jun 9) — baseline daily_brief.py; cross-read Weather Analog Framework Pattern classified; 2–3 analog years; long-range signal only — do not commit
T-7 (Jun 12) — hypotheses Re-run brief; Model → Tactic Synthesis (Pattern Matrix → Model filter → Stream filter → Expedition). Base meta-tactic is T-CL. 3–5 routing hypotheses; preferred strategy at S3
T-5 (Jun 14) — Stream lock Gulf Stream Framework + Crossing Strategies; RTOFS+SST+AVISO North wall, CCR/WCE, entry latitude + crossing-sail commit
T-3 (Jun 16) — critical gate Re-run brief; compute GFS vs ECMWF arrival Δ Δ ≤6h → proceed; Δ >6h → HIGH UNCERTAINTY, brief Skipper, dual contingencies. Crossing window locked
T-48 (Jun 17) — crew brief Everyone present Pattern, first 24 h, Stream plan, 30/20/10 nm squall protocol, sail plan; hand out all 10 cards
T-24 (Jun 18) — final Last cycle; finalize Expedition waypoints Waypoints to St. David's Head locked; nav-computer checklist (nav/PRE_RACE_CHECKLIST.md); watch/sleep plan set
T-0 (Jun 19) — start Card 01 Crew briefed on first 12 h; routing in Expedition; navigator rested
Stream-lock detail (T-5)
  • North wall: SST + RTOFS together. Agree → plan entry directly. Disagree → note uncertain, plan a conservative default.
  • CCR: SST + RTOFS + altimetry — all three must agree before routing through it. RTOFS-only is not actionable (CCR Guide).
  • WCE in path (31–34°N): default avoid; an adverse limb can cost 20+ nm (WCE Guide).
  • Meander exploit (T-ME): SST + RTOFS together, with a pre-calculated exit latitude. Never enter without knowing where to exit.
Crew-brief detail (T-48)

Cover: analog + fleet-typical behavior + Lupo's strategy; first-night conditions; Stream position and plan; squall protocol (each crew member states the action at every threshold without looking); sail plan; and confirm every crew member has read all 10 onboard cards before departure.


Onboard: Start to Finish

Card Phase Elapsed (typ.) Key navigator decision
01 Start to Block Island 0–12h First-night conditions, tidal current, first decision point
02 Offshore approach 12–30h Second night, offshore wind trend, sleep schedule
03 Pre-Gulf Stream 20–35h Stream-entry sail commit, crew brief, role assignments
04 Gulf Stream entry ~35–40h Current confirmation, routing commit, squall protocol active
05 Inside Gulf Stream 35–50h Radar cadence, sail freeze, current observation
06 Gulf Stream exit 45–55h Exit confirmation, post-Stream wind transition
07 Mid-ocean transition 55–70h Tactical reset, model update, crew rest
08 Bermuda approach 70–85h Final 150 nm, approach angle, WCE avoidance
09 Squall & frontal response Any 30/20/10 nm thresholds, all-hands protocol
10 Tactical scenarios Any Sail selection, reef timing, bash-vs-deflect

Cards 03 and 04 are the highest-stakes. If anything from the pre-race plan changed, work through their contingency sections before crossing the north wall.

  1. GRIB/RTOFS update on the briefed schedule — not "when interesting."
  2. Current cross-check: GPS SOG − boat log = experienced current; compare to RTOFS; flag persistent divergence to Skipper.
  3. Wind cross-check: instrument TWS/TWD vs forecast; flag persistent divergence.
  4. T-CL gate check: any deviation needs a named data trigger, a counted maneuver cost, and Skipper concurrence.
  5. Watch Captain at handoff: route, next decision point, current trend, forecast change.
  6. Before sleeping: brief Skipper on the next forecast inflection so the watch can act without waking you.

Post-Race: Debrief

After finish, complete the Debrief Template — actual vs pre-race route, which analog held, Stream-feature accuracy (SST/altimetry), and what to do differently.


Sources

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