How the model works
No black box. This is the whole method, in plain words.
The pool
Every men’s Championship Tour career with at least rated heats since : careers. A career that began before counts from on.
The stats
Ranked against the field he surfed against
A raw number means little on its own: winning half your heats in a stacked year is not the same as in a thin one. So each stat is ranked season by season against the other CT surfers that year with at least rated heats. That gives a percentile from 0 to 100: the higher it is, the more of that season’s field he was ahead of.
A career value is the average of those season percentiles, weighted by heats: a full season counts more than a short one.
Why #1 isn’t at 100
A career percentile is an average of season standings, so it almost never gets near 100: nobody tops the field every single year. That’s why a career value well short of 100 can still be #1 of : nobody’s average is higher.
Measuring the distance
Put a career’s values side by side and you have its fingerprint. To compare two fingerprints we first put every stat on the same scale (how far each career sits from the pool’s average), so no single stat outweighs the others. The distance is then the straight-line (Euclidean) distance across all stats at once.
Smaller means more alike. The model distance has no unit: it only means something next to other distances. The closest career is the twin.
What we leave out
- The WSL Finals. A one-day event in a different format.
- Wave-pool events. No ocean, a different kind of contest.
- Walkovers. Heats nobody really surfed: no performance to measure.
- Incomplete events. Events we don’t hold in full. Half a picture would skew a season.
- Wildcard appearances. A surfer’s own stats count only from seasons on tour. His wildcard heats still count for the surfers he faced.
Why heats, not waves
Heat results go back to ; wave-by-wave scores in our data start much later. Working heat by heat lets careers from different decades sit in one pool.
Why from
Our heat-by-heat data starts in . Scores from before then aren’t comparable with what came after, so careers are measured from on. Even after the scoring scale changed, which is why every heat total is compared with its own event’s average.
What it isn’t
It doesn’t say who was better. Two careers sit close when they won in a similar way, not necessarily at the same height. And it only sees what heat results can show: no style, no single waves, no world titles.
Found an error? DM @surf.stats on Instagram and we’ll check it.