DaveR wrote:
Many other fancier curves could be fitted to the data, producing different ranking lists. There is no particular reason why a simple straight-line fit should be the correct thing to use, and significant evidence that it isn't.
Indeed, a complex enough curve could be made to fit exactly through every data point with the result that every competitor would score their average score at every event!
In order to fit fancier curves you need a lot more data. Perhaps if courses typically had hundreds, rather than tens of competitors it could be done. Given we want the ranking system to produce a result for a field of 10 you are pushing the limits to fit a straight line.