A statistics project on college sailing results, 2008 to 2026

College sailing ratings

Results come from Techscore, the ICSA and ISSA scoring sites. I fit a rating to every sailor and test how well it predicts races it hasn't seen. 1.5 million boat-race results since 2008.

Quinn Brighton · Applied Math–CS, Brown University · Class of 2027 Email LinkedIn

100K+

Training races

24K+

Sailors modeled

13.7M

Parameters

70%

Held‑out pair accuracy

What this shows

• Statistics. A Plackett–Luce ranking model fit to 1.5 million boat-race results, evaluated only on races it did not see: held-out regattas, season-ahead forecasts, and championships with ratings frozen the week before, scored against Elo and simple baselines with bootstrap confidence intervals.
• Modeling. Each regatta is simulated thousands of times to turn ratings into win and podium probabilities. The simulation includes sailor form and boat effects fitted to real championships, and its probabilities are checked against what actually happened.
• Markets. A pretend prediction market (fake coins, no real money) turns those probabilities into odds with a house margin and settles from published results.
• Software. Python (PyTorch, SQLite, Flask), a pipeline that re-collects results, retrains and redeploys every night, and more than 300 automated tests. Details on How it's built.

Predictions

Win probability from 5,000 simulated regattas on the predicted lineups. One event per weekend, the one with the strongest field.

Oct 10–11

Open Atlantic Coast Championships Round 1A

St. Mary's · 18 schools

2

13%

College of Charleston

1

18%

Stanford University

3

10%

Roger Williams University

Oct 10–11

Sister Esther Whalen Trophy

Salve Regina · 14 schools

2

16%

Tufts University

1

37%

Boston College

3

16%

Yale University

Oct 10–11

Stu Nelson @ Connecticut College 2026

Connecticut College · 10 schools

2

16%

Harvard University

1

38%

Yale University

3

15%

Brown University

Oct 17–18

Captain Hurst

Dartmouth College · 10 schools

2

19%

Dartmouth College

1

19%

Brown University

3

16%

Boston College

Oct 24–25

Oberg Trophy

NU/MIT · 12 schools

2

24%

Roger Williams University

1

35%

Brown University

3

13%

Dartmouth College

Oct 30–31

Nickerson Trophy

Tufts University · 10 schools

2

21%

Yale University

1

38%

Brown University

3

8%

Boston College

Predicted lineups and simulated standings for every regatta.

Ratings over time

Method

Every sailor gets a skill rating for each week they race, θ[sailor, week]. I fit the ratings so the observed finishing orders are as likely as possible under a Plackett–Luce model, which handles a full finishing order rather than just win or loss. Penalties on week-to-week change keep a rating from jumping on one regatta. About 13.7 million parameters, trained in PyTorch.

The Boat Speed Analyzer uses the same data. It compares a school's fleet-race results across two time windows and tests whether the difference is larger than race-to-race noise.

Accuracy

I held out a random 1 in 20 regattas before fitting. On those races the model picks the faster of two boats 70% of the time. Forecasts a full season ahead, and the comparison with Elo and simpler baselines, are on the methodology page.

Why

• Most college sailing rankings are win totals or opinion. I wanted ratings I could check against results.
• I'm Club Captain of Brown Sailing and compete regularly (my profile). I use this tool to size up a fleet and put my own performances in context.

More