Team Racing

What the site does with team racing

• Forecasts. Every team race regatta page has a Ratings & Forecast tab: each side's pre-event strength and boat values, the forecast for every match, who beat it, and the event's measured orderliness.
• Ratings. The power rankings have a Team Race selector: what each skipper brings to one of the three boats, and a school table that adds the three best skippers and the school's own team race term.
• Sailor pages show a team race rating, its history before each regatta, and a regatta-by-regatta surprisal: the record against the expected record, and how surprising it was.
• Plays. An explorer for how team races are won: the mix of finish sets, by play, as a side gets stronger, as one boat is raised, as a star is built, by breeze and level, with a high school toggle.

Three notes explain the work: how team race forecasts work, which team races go to form and Play 1, Play 2, Play 4: how team races are won.

How the forecast works, in one paragraph

A side's strength is the sum over its three boats of 0.62 times the skipper's fleet rating, plus a small team race offset for the skipper and the crew (shrunk toward zero; their fitted spread is only 0.018 and 0.016), plus the school's team race term (spread 0.20) and a home-water edge of 0.13 when the school hosts. The chance that one side beats the other is the logistic of the difference. Fleet ratings alone call 74.5% of held-out matches; the full model calls 75.8%. The model was developed and tested on college data; the high school fit uses the same structure, with crews often missing from the line-ups.

Recent team race regattas

Forecast made before the event; called is the share of matches where the favourite won; orderliness is how much more (green) or less (red) the favourites won than the forecast said, with its standard error. Nothing known in advance predicts an event's orderliness (field, host and wind all fail), so it is a description, not a prediction. One thing does predict a little in high school: championships are more orderly (R² +0.04).

The school term

What a school adds to the sum of its three boats, in rating units, for schools with at least 300 matches (top 10 of 20). These programmes win more team races than their fleet ratings alone would predict, season after season.

School	Team race term	Matches	Win share
St. George's School	+0.61 ± 0.23	337	76%
Point Loma High School	+0.48 ± 0.26	309	79%
Tabor Academy	+0.48 ± 0.22	397	66%
Christchurch School	+0.29 ± 0.15	1,024	73%
Norfolk Academy	+0.26 ± 0.17	651	60%
Broadneck High School	+0.22 ± 0.28	422	44%
New Trier HS	+0.12 ± 0.21	364	52%
Norfolk Collegiate School	+0.12 ± 0.21	736	58%
Bainbridge High School	+0.09 ± 0.19	656	59%
Saint Ignatius College Prep, Chicago, IL	+0.09 ± 0.28	301	52%

How team races are won

Across 12,325 high school matches, the winner held first place in 88% of light-air wins and 92% of heavy-air wins; high school winners sweep 1-2-3 more as the breeze builds, the opposite of college, where winners hold first less and win through Play 2 and Play 4 more. A star with two weak boats wins the same way as three even boats. Explore the mix of finish sets or read the note.

Orderliness over 350 ISSA team race regattas: true event-to-event spread 0.52, split-half agreement r = +0.36, field R² +0.05, wind r = +0.10.