I think my point wasn't clear. Not just one sailor, but let's have N how many sailors, the same group of sailors, competing against each other. A round robin, if you will.

If a trend appears favoring any particular boat, it would certainly show, and you've reduced your variables.

F'rinstance:
Sailor X sails Formula Z in 5 races, wins four times;
Sailor X sails Formula Y in 5 races, wins two times;
Sailor X sails Formula U in 5 races, wins once;

Formula Z is looking pretty good to me. Now, let's see if Sailor Y and Sailor U show a trend with the same boats, etc.

The situation I suggest is actually basic experimental procedure. My suggestion is of course describing an ideal situation. In fact, it would be made even more accurate by having these people race the same competition every time as well.

See what I mean? My point is, you gotta reduce variability before you make comparisons. Or else get a REAL large, UNBIASED, sample size. As if. That's what the marketing departments are paid to avoid.

sea ya
tami