One more point, but I think important. When we talk about reprodicibility of a statisical result, one of the most important has to be reproducibility in the results of the particular athlete. As these ratings seem to be based on the Beyer speed ratings in horse racing, to use an example, we can trust the speed ratings for Secretariat because they happened over and over and correspond to race records that still stand in all three triple crown races.
With Tuohy we see again repeated comparable performances and the course records to back it up that gives us a measure of confidence in the ratings she achieved.
With the Trotter and Fairchild ratings the issue has always been that they are such outliers per that particular athlete's individua performances that something seems amiss.
So let's take this race's ratings. Before today the highest SR either athlete achieved was 162, and that in a FL regional against what turned out to be a very beatable opponent. Dudek only had a 159.
But now all of the sudden we get a 169 and 168 where they finished within one second of each other where three girls ran a sub 17 in the same race. This again seems to be the definition of a result that is not reproducible.
Anyway.