fastest production 14foot dw racer

laceys lane said:

"i'm suggesting testers stay off their 'own boards'. there's more to be gained by not paddling them imo".

This is an interesting and valuable suggestion, and it is easy to see why it might be made: Perhaps people will "sandbag" on all other boards than their own, or their familiarity with the board would mean that they are disproportionately good. It might be a particular problem where a majority of the testers are familiar with a particular board or brand.

But can I please suggest instead that the test be run with every paddler paddling every board, but that a record is made of how familiar each rider is with each board?

We can then examine statistically for an effect of this familiarity. It would allow us to see if familiarity is indeed an important factor in speed. Moreover, there might be a statistical interaction between familiarity and certain boards - in other words, that you need to learn how to get the best out of some boards, but with others, you can be fast straight out. It seems to me that this would be a valuable thing to know.

We can then statistically remove from the data the variance attributable to familiarity, and can then be as sure as we can be that the results we get are unaffected by a person's experience with a particular board (or indeed, brand).

It is much better to collect a dataset where all riders ride all boards, with no gaps in the data. For small trials like these, gaps in data are pretty fatal and you can easily end up with a dataset from which nothing can be deduced, and the whole effort is wasted.

It would be better to measure "familiarity" (say, rank order of the boards for each rider, according to the number of hours they will have spent on it) than "ownership" because, as has already been pointed out, many people have more than one board, or might have e.g. recently sold a board, or often used a friend's etc.

More broadly, I'd say that there are likely to be pretty sensible solutions to most of your concerns that can be taken care of when designing the experimental trial. So by all means express them, and we can collectively discuss the best way forward. Maybe if you have all had a say in the design of the trial, you may view the results with less scepticism.

And if there are those that doubt the results that PT Woody (or anyone else for that matter) discovers, or wants to compare the performances of boards not included in a particular trial, I would be very happy to help them design and analyse their own trial so that they can see for themselves.