There are different ways to try and judge a teams relative strength, but since in California NXN bids are based on the big merge of all the divisions into one race, I tried to make a system based on that. I use the speed ratings on tullyrunners a lot, but it doesn't quite solve this case, because the rating points are mostly linear with time, and in the Cal merge, it is radically different. The difference in points between improving 10 seconds from 14:50 to 14:40 is about 1 point. Improving 10 seconds from 16:30 (roughly 216th place) to 16:20 (171th place) is about 45 points.
Note this a merge score, so the placement is based on after individuals have been removed.
So first I built a table based on tullyrunners speed ratings to their approximate CA merge points.
Then I used used another table on tullyrunners, which mapped speed ratings to 1600m and 3200m times.
I made some small adjustments to the mappings based on historical data, and created CA merge points tables for those events.
Since this is a XC score, low score is better. A 4:00 1600m is a score of 1, meaning that runner is projected to have the fastest time in the merge.
People with experience with this can probably point out one potential flaw. A 1600m time in isolation, can be a lousy indicator of XC times. I have seen lots of runners with excellent 800/1600m times that never were relevant at the state level in XC. But some really good XC runners focus on the 1600m, and never run a fast 3200m, so just ignoring the 1600m would miss a lot of runners. So honestly, I don't have a good answer for it, but I am thinking of a few options.
In the end, there is nothing California specific. It could be used to compare any team in a big merge setting with track times and/or speed ratings. This is just using the Cal state meet merge as the model of the scoring.
I tried these tables on Ventura, which returns all of its runners. So for each runner, I took the current PB at 1600m and 3200m, and also last years best XC speed rating that I could find,
and it seems to check out pretty well.
Some examples:
Anthony Fasthorse has run good times on all distances, but his 4:08 is his best relative time.
Computed score for t1600 of 4:08 = 3
Computed score for t3200 of 9:06 = 11.4
Computed score for speedRating of 188 = 8.5
Best score = 3
Henry Hammel has a best relative time of 9:24, and it shows up as expected
Computed score for t1600 of 4:29 = 178
Computed score for t3200 of 9:24 = 77
Computed score for speedRating of 170 = 117
Best score = 77
Nick DeGeorge excels at XC, and his 15:35 at the state meet is his best mark by far, so that is used
Computed score for t1600 of 4:32 = 240.1
Computed score for t3200 of 9:35 = 136
Computed score for speedRating of 179 = 47.7
Best score = 47.7
I tested using 3 teams, using their top 6 and letting the program determine the top 5 best marks.
For some runners with PBs from the previous year, I used those.
The scores (only the top 5 count):
###Team Scoring###
-----------
Ventura 247.7
Anthony Fasthorse 3
Micah Grossman 3
Nick DeGeorge 47.7
Henry Hammel 77
Hollis Costa 117
Blake Harris 196.9
-----------
Great Oak 309.5
Westin Brown 18
Gabriel Rodriguez 20.4
Jeffrey Keeney 27.3
Michael Rodriguez 97.6
Jack Paradise 146.2
Jacob Brown 160.8
-----------
San Clemente 360.2
Brett Ephraim 10
Pierce Clark 69.7
Taj Clark 77
Kai Olsen 83
Dallin Harrington 120.5
Isaac Gould 202.3
For reference, here are the top schools in last years merge:
1 Newbury Park 52 1-2-3-14-32(52)(169)
2 Great Oak 214 19-37-41-58-59(97)(125)
3 San Clemente 264 22-25-48-63-106(115)(137)
4 Granada 303 17-55-68-73-90(120)(123)
5 Oakdale 361 5-16-29-87-214(596)(739)
6 Ventura 442 7-24-43-121-247(429)
The projected scores look high because they is still more improvement coming in track times and speed ratings next XC season.
I'll add some more teams, and after Arcadia some top individuals.
If anyone has any particular teams they want modeled, let me know.