DI and DIII Cross Country Nationals Simulations

This is pretty sick. For D1 preditions I woul dgo through and weigh some region meets less because some teams chilled or sat guys out. Take Wisconsin for example, they sat their #2 and if he had run they would've won the meet easily.

Yeah, I currently weigh the regional meet the most, I am working on re-running weighing it about half as much as conference meets. I'll also take Wisconsin's top 7 from conference so I have the correct runners.

How cool is this? Looking forward to come back to this post-race. Creative. Nice work.

Seufer winning would be crazy. He's my dark horse pick to win.

Great work. But why would Falasco be at 51 after winning the AR?

Have you tried running your analysis on data from previous years and comparing your predictions to the results?

Can this software run a simulation where I ask out all the girls at my school and find out which day they might be most receptive?

Also I would go through and reweigh some other meets. Most seem correct but looking at the national catholic meet on your list it isn't a slow course. It is the same as Joe Piane invite.

settle down there noah we all know you have a good kick.

settle down there noah wrote:

settle down there noah we all know you have a good kick.

honestly, I'm just a guy in his 60s who ran D3 decades ago and have a particular interest in the AR. Seemed curious to me. I have no connection to Noah or his school. (And I'm only responding because I didn't mean to implicate the kid.)

plz kind sir wrote:

Can this software run a simulation where I ask out all the girls at my school and find out which day they might be most receptive?

They all said no

Chapman is missing off your d3 west region rankings. I assume it's because you used team scores from last year's regional and they didn't have a full team compete. But they did have full teams this year! Awesome rankings!

These ratings are amazing, but can somebody please explain to me how Peter Seufer has a 42% chance to win the meet and is by far the favorite here, while Edwin Kurgat (the odds on favorite in everything else I am reading) is rated 7th?

DIII and DI have different sources of instability.

In DIII there is not a lot of inter-region competition. So knowing how good each region is relative to others often comes down to one or two teams running at one meet. If those teams have an off day, or a great day, the rest of the region gets a penalty/boost.

In DI the relative evaluations of meets is much more stable because everyone races everywhere. The problem in DI is a lot of teams jog/don't take races seriously. Because I go based on normalized time and average performances, this can lead to some outliers. I'll look into Peter's scores, see if one of them is a huge outlier from the others, and report back.

Nice model.

These DIII predictions are comical, although largely for reasons that are just intrinsic to the coaching style of many of these programs or reasons that are difficult to weight with these kind of models when these teams rarely face one another during the regular season. Emory and Berea will both finish near last. Hopkins and MIT are both over-peaked. NCC will not be third. Williams might win, but the scores will be tighter. It is crazy to St. Olaf and haverford near to the bottom of this list, but that might be right. when I was in college, those were two of the best teams in the country. It shows how absurd it is that NCC has maintained it's excellence for so long. Although they do not fight many of the same academic plights as some of these other schools. That is not to say Al Carius isn't one of the best coaches in america.

Cheers

Its a down year for Haverford. They will reload soon enough so seeing them that low is appropriate. They nearly got 4th at their region meet and the team that did get 4th didn't even get a bid to natty. Haverford ranked 22 in the natty pole this week while Etown lost by 9 pts and didn't even rank and had a better team average. Hop got demolished by CMU so makes sense they are overpeaked.

The simulation and average ranking posted some time on the 20th and there at the time of my posting here is missing BYU's top runner Whittni Orton and also missing Stanford's top runner Fiona O’Keeffe. Both were rested for regionals. Orton was also rested for Pre-Nats.

The prior version posted earlier in the day on the 20th was missing Arkansas's 5th runner Lauren Gregory.

I did not look to see which men may have been rested for regionals and thus may be missing. I know that these women were rested because it was mentioned in articles here at Letsrun or mentioned by Flotrack.

So without knowing everyone who was rested at regionals, especially the teams' top 5, the ranking will be off. If will be interesting to see how close the simulation was after the race and after the simulation has the correct set of runners for each team.

I've made the following changes:

11/20 changes:

Lindsey Rudden -> India Johnson (Michigan State)

Adoette Vaughan -> Lauren Gregory (Arkansas)

11/21 changes:

Paxton Smith -> Alec Hornecker (Colorado)

Added noise to bootstrapping simulations for better probabilities

Removed any meet with the word “alumni” in the title

More 11/21 changes:

Fahy, Kristin -> O’Keeffe, Fiona (Stanford)

Sargent, Sadie -> Orton, Whittni (BYU)

Brown, Reed -> Hocker, Cole (Oregon)

Also now you can click on peoples names to go to their tfrrs, which is nice

Here's another simulation:

http://seanfischer.me/posts/men_2019_predictions.html

This model differs from Bijan's, but between the two of them I think there is a pretty good picture of the men's meet!

This model likely undervalues the Atlantic and Mideast regions to some extent (due to a good amount of inter-regional racing followed by weaker conference and regional competition), and overvalues the West and South (lack of inter-regional racing). This model produces a slightly different look at the title race, with Williams and NCC appearing essentially tied. As with Bijan, there is also a chance that some runners need to be switched out for some teams' top 7s.

The model only includes the team race, but on an individual level Aidan Ryan has the highest rated race of the year (Conn College on 10/19) and is undefeated.