2023 CIMAGLIA RANKINGS UPDATE
Hey everyone. It took me a while to get things going this year, but the XC predictor is ready for its next season and I'll start posting predictions and rankings very soon. I'm very excited to see what this year has in store and I have a few changes to the program that I'd like to share.
As always, feel free to email me with any questions or prediction requests, and if you know of any new course changes or runner injuries I'd really appreciate a heads up.
1. Prediction Algorithm Now Takes Race Recency into Account
When the program makes predictions it no longer uses every race you've ran for the season as a basis. Rather, it hones in on more recent races which should make any late season improvements or slumps more evident. To be exact, the program counts the 1st and 2nd most recent races to be twice as important as the 3rd and 4th most recent, which are in turn twice as important as the 5th and 6th most recent.
2. Hill Efficiency No Longer Taken into Consideration
With the above change in mind, I've stopped using individual runners' hill efficiency in prediction calculations. The feature was too finnicky last year and didn't contribute to more accuracy in predictions.
3. Consistency Now Featured in Runner Rankings
I briefly mentioned consistency ratings in my state meet predictions last year as it was a feature that wasn't ready until late season, but now they will be present on the runner ranking page. Simply put, a runner's consistency rating is how many seconds they typically run above or below their average performance. In other words, the closer you are to zero, the more consistent you are.
4. Improvements to Duplicate Runner Finder
I've heavily optimized the part of the program that detects duplicate runners in the system so you can expect to see less errors in predictions and rankings from now on.
5. State Meet Course Now Used as Starting Point for All Other Course Difficulty Calculations
When the program calculates course difficulties, it's given one course to start with a default difficulty of 1.0 and then branches off to calculate all the other courses. The starting course was formerly the Chick-Fil-A Invitational course, but it is now the Cabell Midland course as it has seen much more action and is thus more suited to be the basis for everything else. This comes with the side-effect that course difficulties from last year are now obsolete, so be sure not to compare them to this year's difficulty ratings.
6. Various Efficiency Improvements
The program runs significantly faster now due to some behind the scenes improvements. This won't change anything for you all but it is great for me and will hopefully keep me from getting behind on predictions ;)