Learnings from retiring
If you’ve sailed Gotland Runt, your brain carries a hard-wired association: Almagrundet means you’re nearly done.
Finish, Sandhamn, beer, sleep.
Twenty years of conditioning.
The Baltic Sea Race course put that exact mark a quarter of the way into 635 nm. With 430 nm of heavy upwind and reaching still to go. So when we bore away at Alma after 36 hours of racing, leading IRC Two, two-handed and FinRating, we could practically hear the music from Seglarhotellet across the water.
Homer covered this three thousand years ago.
The sirens’ trick isn’t that they sing. It’s that they sing your song. Odysseus knew willpower would lose, so he didn’t rely on it: wax in the crew’s ears, himself lashed to the mast, standing orders to ignore the captain when the captain begs. Behavioral economists later named it a Ulysses contract.
We didn’t have one.
We had a car parked at Bullandö Marina.
We’ve since done a fairly brutal debrief; tracking data, weather archive, and an uncomfortable amount of psychology. Our conclusion: the decision was probably right for that race.
The process was wrong in ways that would bite much harder in a race that mattered more.
Here’s what we learned, in case your boat has its own Almagrundet.
1. Decide what the race is for. Before the start. Asked separately how we’d actually treated BSR, we both said ~80 % “we’re here to win”. Asked how we wish we’d defined it: 40 % racing, the rest experiment and training. When the fleet collapsed and the best shorthanded teams retired, the 80 %-frame made retiring rational. A goal you never wrote down can be renegotiated at 02:00 by whoever is wettest.
2. Chart the sirens along with the course. Every course has places where quitting is cheap; a home harbor, a familiar marina near a mark. If the course rounds something your brain files under “finish”, expect the music. The kicker: we had identified this exact mechanism before the race – about other races. Knowing about the sirens doesn’t help. Wax helps.
3. The conditional pact IS the decision. Somewhere mid-morning we agreed: “if the forecast still looks like this at Alma, we’re done.” That sentence felt like postponing the decision. It was the decision. From that moment, every forecast check asked “do we get to stop?” instead of “should we?”. The burden of proof had flipped – twelve hours before the mark, and neither of us noticed.
4. Classify before you argue: RED, AMBER, GREEN. RED = safety, damage, dangerous fatigue: you stop, no philosophy required. AMBER = the sporting equation genuinely changed: analyse it. GREEN = cold, wet, bored, unmotivated, far from home: not a reason. That’s the sport you entered. Most of what our brains served up that night was GREEN wearing AMBER’s foulies.
5. Double-handed consensus comes fast. Too fast. Both of you doubting privately, neither wanting to be the weak one. Then one careful “are you thinking what I’m thinking?” And the answer is relief. Two private doubts become one joint decision in minutes, and it feels like truth because you agree. Nobody made the case for continuing. Make it a rule that someone must, properly.
6. Nobody will hold you accountable. That’s the trap. We posted our RET with a screenshot of the leaderboard. 188 reactions: “wise”, “smart”, “love the honesty”. One single critical comment – aimed at the organizers. Retiring while leading pays out socially like a win, at a fraction of the cost, which quietly teaches your brain to do it again. External accountability, measured: zero. The only accountability that exists is what you write down before the start.
7. Sometimes retiring is right. So decide in advance which races deserve the grind. Annie Duke’s research says most people quit too late; walking away at the top is rare and sometimes optimal. The skill isn’t grit – it’s aim. We now carry a one-page decision contract: the frame in percentages, what success looks like without a podium, the sirens marked on the chart, and a protocol for the moment “should we retire?” comes up. First live test: next race. If you want the template, see below.
Worth remembering how the story ends: Odysseus got to hear the song and sail past. That’s the actual goal. Not a crew that never retires, but a crew that knows why it continues, knows why it retires, and can tell both apart from how it happens to feel at 02:00.
Decision contract — template v1
Built by the crew of J/99 BLUR after retiring from the RORC Baltic Sea Race 2026 while leading. Not a promise never to retire. A promise to know why we continue, why we retire, and to tell both apart from how we happen to feel at 02:00.
Fill in BEFORE the start, per race. Ten minutes, both crew present.
One copy in the race folder, one on paper aboard.
Before the start
1. Frame (adds up to 100 %, both sign): RACING ___ % · EXPERIMENT ___ % · TRAINING ___ %
2. Success without a podium. This race was worth doing even if we don’t win, provided: ______
3. What’s at stake. Qualification, season goals, anything else: ______
4. The sirens on this course. Where along the track is quitting cheap? Home waters, a comfortable harbor near a mark, a place your brain has filed under “finish”. Mark them NOW: ______
5. Retirement classes:
- RED = safety or capability: injury, illness, dangerous fatigue, critical gear failure, weather outside our margins. We stop. No discussion needed.
- AMBER = the sporting or strategic value of the race has genuinely changed. May be examined, but only through the protocol below.
- GREEN = cold, wet, bored, unmotivated, a long way to go, missing the shower. Never a reason on its own. This is the sport we entered.
During the race: when “should we retire?” comes up without a RED
a. No conditional pacts. A sentence like “if the forecast still looks like this at X, we retire” IS a retirement decision. It triggers the protocol now, not at X. (A pact quietly flips the burden of proof: from then on you check whether you’re allowed to stop, not whether you should. That is how Almagrundet was decided, hours before Almagrundet.)
b. No consensus in the first conversation. Separately, in writing, five minutes: my reasons, each classed RED / AMBER / GREEN, plus “if I were deciding alone, my probability of retiring is __ %”.
c. Steelman both directions, out loud. The best possible case for continuing AND for retiring. Not “we should probably push on a bit”. The real version.
d. Check the facts against data. Competitor picture from the tracker (course vectors count as verification), forecast from the models. Then the goal test: “Does retiring affect what we set out to achieve in points 1–3?”
e. Decide no earlier than one full watch cycle after (b). Until then the plan stands; the default is continue. The burden of proof is on retiring: “What relevant change means this contract no longer applies?” If that question can’t be answered concretely, the answer is GREEN, and GREEN doesn’t retire from races.
After the race
f. One line in the debrief: did the contract hold? Was the question raised, and was the protocol followed? What changes in this template for next time?
v1 — untested.
Further reading: Annie Duke, Quit
I actually listened to her first book, Thinking in Bets, just a few weeks ago when we thought about our decision system around weather. Should have read this one instead.
Annie Duke played poker for a living before she started writing about how people decide badly. Her main finding is that most people quit far too late, not too early. Leaving feels like admitting the whole thing was a mistake, so we don’t. Grit gets all the credit; knowing when to walk away is the harder skill and nobody gives you a trophy for it.
