Old Dominion University '17
How much better or worse boats do with this sailor crewing than their skipper's rating predicts, shown on the skipper scale so crews can be lined up across teams. Percentiles compare against rostered crews that season (league) or teammates (team).
How crew ratings work →Crew effect over their whole career, not scaled to a team: 85.45 percentile among ICSA crews with 30+ races. Uncertainty ±0.080 rating points from 343 races with rated skippers.
How crew ratings work →No season rating yet: needs 8+ crewed races in a season.
A team race rating is what the sailor brings to one of the three boats: 0.72 x their fleet rating plus a team-race offset learned from every match they sailed, shrunk toward zero so a few matches cannot move it far. Sides are compared by the sum of their three boats plus the school's team race strength.
| Rating | Rank | vs fleet rating | Record |
|---|---|---|---|
| 2.09 (fleet 2.88) | unranked (fewer than 30 matches as skipper, or not active) | −0.00 ± 0.18 about what the fleet rating predicts |
as skipper 0–2 (0%) as crew 111–135 (45%) |
| Season | Regattas | Skipper W–L | Crew W–L |
|---|---|---|---|
| Spring 2017 | 5 | — | 41–32 |
| Spring 2016 | 5 | — | 42–39 |
| Spring 2015 | 5 | — | 22–48 |
| Spring 2014 | 5 | 0–2 | 6–16 |
For each regatta, the boat value that would have made this sailor's record least surprising: the shift applied to their boat alone, in the matches they sailed, with team-mates and opponents fixed at the pre-event forecast. A boat's skipper and crew share the boat, so they get the same number when they sailed every match together. Surprise is in bits at the pre-event value minus at the performance value; P(this good or better) treats the matches as independent at the forecast probabilities. Regattas forecast from fewer than 1,000 training matches are left out.
Over 18 graded regattas (245 matches) the record is 111–134 against 96.4 wins the pre-event forecasts expected (+14.6). 5 regattas clearly above the rating (P ≤ 20%), 7 clearly below (P ≥ 80%); 33 bits of surprise in all.
| Date | Regatta | Role | Value | Perf. | Record | Exp. wins |
Δ | Surprise (bits) |
P(this good or better) |
|---|---|---|---|---|---|---|---|---|---|
| 2017-05-27 | LaserPerformance Team Race National Championship | crew | 3.74 | 4.13 | 8–7 | 7.2 | +0.39 | 0.2 | 40% |
| 2017-04-08 | Prosser Trophy | crew | 3.60 | 4.95 | 10–5 | 5.8 | +1.34 | 4.0 | 2% |
| 2017-03-25 | 2017 Aaron Szambecki Team Race | crew | 3.49 | 5.41 | 18–3 | 11.8 | +1.93 | 7.8 | <1% |
| 2017-03-18 | Graham Hall | crew | 3.42 | 4.21 | 5–8 | 3.5 | +0.78 | 0.9 | 21% |
| 2017-03-11 | St. Mary's Team Race Interconference | crew | 3.47 | 0.47 | 0–9 | 2.1 | ≤−3.00 | 3.3 | >99% |
| 2016-05-28 | LaserPerformance Team Race Nationals | crew | 4.08 | 4.18 | 9–8 | 8.8 | +0.11 | 0.0 | 56% |
| 2016-04-09 | MAISA Team Race Championship/Capt. Prosser Trophy | crew | 3.96 | 5.09 | 10–4 | 7.3 | +1.13 | 2.1 | 8% |
| 2016-03-26 | Aaron Szambecki Team Race | crew | 3.95 | 3.38 | 7–11 | 8.4 | −0.57 | 0.6 | 89% |
| 2016-03-05 | Bob Bavier Team Race | crew | 3.94 | 4.34 | 10–13 | 8.5 | +0.40 | 0.4 | 29% |
| 2016-02-27 | 2016 Jeremy McIntyre Team Race | crew | 4.03 | 1.64 | 6–3 | 8.0 | −2.39 | 3.6 | >99% |
| 2015-04-11 | Captain "Joe" Prosser Trophy | crew | 4.03 | 4.09 | 6–8 | 5.8 | +0.07 | 0.0 | 57% |
| 2015-03-28 | 2015 Aaron Szambecki Team Race | crew | 4.07 | 4.14 | 7–16 | 6.7 | +0.07 | 0.0 | 53% |
| 2015-03-21 | Graham Hall Team Race | crew | 3.98 | 5.00 | 6–9 | 3.8 | +1.02 | 1.7 | 10% |
| 2015-03-14 | St. Mary's Team Race | crew | 4.00 | 3.05 | 1–8 | 1.9 | −0.95 | 0.6 | 92% |
| 2015-03-07 | Bob Bavier Team Race | crew | 4.07 | 2.74 | 2–7 | 3.3 | −1.33 | 1.2 | 97% |
| 2014-04-05 | Prosser Trophy | crew | 3.42 | 4.98 | 6–9 | 2.6 | +1.57 | 4.1 | 2% |
| 2014-03-22 | 2014 Aaron Szambecki Team Race | crew | 3.49 | 0.49 | 0–3 | 0.2 | ≤−3.00 | 0.3 | >99% |
| 2014-03-15 | Graham Hall Team Race | crew | 3.63 | 0.63 | 0–3 | 0.8 | ≤−3.00 | 1.7 | >99% |
Short write-ups of what the ratings and results show, with the numbers and how they were computed.
