Singlehanded and keelboat ratings
How the discipline ratings and finishes on a sailor's page work.
What the ratings are
Singlehanded and keelboat (including match race) each have their own rating. Each is on its own scale, so the numbers cannot be compared with the main rating or with each other.
Model
strength of a boat = k × main rating of the helm + crew term + sailor offset + school offset
- The helm is the best-rated sailor aboard. The crew term is the mean main rating of the others times a fitted weight (keelboat).
- The sailor offset is how much better or worse the sailor does in this discipline than their main rating suggests; the school offset does the same for the program (some schools sail keelboats far more).
- Fitted by Plackett–Luce on every race of the discipline, with older seasons down-weighted (half-life about two seasons for singlehanded).
- A temperature (0.8 singlehanded, 1.0 keelboat) shrinks predictions that were overconfident on unseen seasons.
Does it help?
Share of boat pairs ordered correctly on seasons held out of fitting:
| Main rating only | Discipline model | Held-out pairs | |
|---|---|---|---|
| Singlehanded (f24, f25) | 63.9% | 68.1% | 83,229 |
| Keelboat (f24 to s26) | 61.4% | 65.5% | 14,666 |
Rank and results
- Rank. Among sailors who raced the discipline in the last three seasons, with at least 10 races.
- Finish. The boat's rank among all boats in the regatta by total points. A singlehanded sailor is their own boat; a keelboat finish belongs to the whole crew. Points include penalty scores.
Limits
Techscore publishes match-race results as a single final ranking, so there is no round-robin data to rate from. Old singlehanded rows without a class year, and keelboat boats with no skipper entered, are not used for ratings.