Collegiate cross country computer rankings and simulations (LACCTiC)

Very cool. It looks like computers are finally starting to reach their potential.

It looks like you are missing a decent number of runners who are returning and include others who have moved on. Hopefully it will be easy for you to clean it up through a script after another month of races when you can see who is actually running.

I think I have all the runners who have raced a cross country race - but I’m missing people who have done track but not XC (scraping all the track meets would have been a nightmare). But I guess it’s possible the scraper missed some people? Tfrrs is kind of a mess of a website.

People will hopefully be added as they start to race. Right now I’ve included a lot of people who probably graduated but still have eligibility just in case. Im going to set anyone who doesn’t race in the next two weeks as “inactive” which should clean stuff up. You can also manually delete them in your custom simulation.

Good work.

Except it doesn't factor in all of the good altitude teams and runners who compete at altitude instead of sea-level. A guy running 15:00 on the track in Gunnison, CO (7700 ft) would be better than a guy running 14:30 at sea-level.

It is a nice tool, but at first glance you seem missing some - unless you deliberately omitted those listed as seniors last year. Saw this for Stanford, BYU, NC State just to start. For example, Lawson, Heymach, Steelman, Orton and others are missing all of whom ran XC and also 5,000 outdoors I think.

I have incremented any reported year from last year. Anyone marked as a SR on results last year is now marked as graduated. Is this incorrect? Are those runners returning?

For example, Heymach has a profile here:

https://www.lacctic.com/runners/18604

But she does not appear on the roster because I assumed seniors graduated. I think TFRRS reports eligibility, but I could be wrong.

I searched by teams, not by athlete. But last year's XC seniors can come back this year as many of them are. same goes for Indoors I think and I am not so sure about Outdoors. Heymach, Lawson, Orton, and Steelman (and many others) are listed on rosters this year. An odd year, to be sure.

Washed-up Grad Student wrote:

I have incremented any reported year from last year. Anyone marked as a SR on results last year is now marked as graduated. Is this incorrect? Are those runners returning?

The Covid extra year has certainly messed with your algorithm on who has eligibility. I suggest just making sure you are flexible and have the ability to add athletes you weren’t expecting as results show up in TFFRS.

Also, I’d wait until schools have at least two meets before archiving anyone. And like with Covid, be ready to put a few back in when they get healthy and run.

Ok, I've set everyone back to seniors unless their most recent race was earlier than 2020. Hopefully thats more accurate

Nice work again Bijan! I always love your simulations. Hopefully, this will be updated throughout the season because I will be constantly checking it.

What is the difference between Current TIC and Season TIC? Why do older times have a lower Current TIC than a similar more recent time?

Great work! Tough to do the algorithms with the Covid eligibility complicating things. Hopefully after a couple of races will be able to get more clarity on who is still running vs has graduated

littlebigplanet21 wrote:

What is the difference between Current TIC and Season TIC? Why do older times have a lower Current TIC than a similar more recent time?

Each performance is adjusted to an equivalent 5k for the previous season. If I run a cross country race in 2019, it will be adjusted to an equivalent 5k in 2019 . The modern TiC will be adjusted to the 2021 track season

These equivalences are built by comparing what people run, so there is a natural drift. Everything is in terms of percentages, but at around 15:00 people on average improve 2-3 seconds per year. That is, a 14:57 from a year ago is about equal to someone running a 15:00 today. From 2020-2021 there is a sudden jump of around a 13-14 second improvement in 5ks in the 15:00 range. Personally I believe it is because of an advancement in shoe technology.

It may seem like this adjustment unfairly rewards people who haven’t raced for a while - but the reality is that we would have expected someone to have improved after a year. Of course, older performances are lightly weighted in my system because of their weaker correlation, so new races will erase old marks quickly.

Washed-up Grad Student wrote:

I have incremented any reported year from last year. Anyone marked as a SR on results last year is now marked as graduated. Is this incorrect? Are those runners returning?

Many are returning because COVID has allowed an extra year if eligibility.

Yeah, thanks for this info! I've set most people to "active" and "seniors" now to bring them back. The only people who are set as graduated are people who competed as "seniors" in 2019. There might be some exceptions but they'll get put back in if they race.

just fyi Ella Donaghu of Stanford is one of those who is NOT coming back

@rojo @wejo @letsrunwebguy

SPONSOR THIS MAN'S WEBSITE IMMEDIATELY

It is a national treasure and cannot be lost!

I LOVE this concept. I think normalizing things to 5000m times is a great way to get solid predictable data. Making that happen in an accurate way is obviously difficult but kudos on where you're at on this.

I will say that looking at my teams specifically that there are a LOT of people that aren't on my roster and haven't been for a while. There are also people that have have projected 5000m PRs that are slower than their ACTUAL 5k PRs. Not trying to be critical, just some feedback for you.

The good part about the people not being on rosters is that you can hand create simulations, and while that's more work, it's at least an option. I guess one option he could give us is to manually subtract people from simulations, i.e. if we know someone is hurt or graduated.