Lacctic ratings

They look very good. I think those will be the spacing at nationals.

They take the course measurement at it's word for what is posted. The work study student who painted the course is not as meticulous as people on this board or who work at Lacctic think. Also projecting the 5k shape these runners are in using courses that vary greatly, relative place to others, and what may not be max efforts are wild projections. Valby goes hard from the gun and makes sure she doesn't get beat by anyone who is not as fast. Her projections are always going to be relatively fast in this system. It's like no one who believes these ratings have participated in a race before act like coaches specifically hold back top runners in early season races "Katelyn, it's just ACC's, stay with the leaders for at least 4/5k before making a move. You should be able to run away from everyone at that point." Just like some people/teams are not racing all out at regionals when all they need to do is qualify, some teams are good enough to do that in their region with just a 90 percent effort or holding out some of their top 7.

I didn't suggest Valby or Tuohy were faster or slower than each other. I just suggested they raced differently. I'll let the ladies figure out who's faster in a couple of weeks in Charlottesville.

In college where many of the very best do not run flat out often it might underestimate their fitness. But for the vast majority it is good. What it misses I think is a big change in fitness where a runner has a race with lower fitness being included in the average (current) rating. See Bush of NC St who had a slow result at ND early this year which weighs on her current rating. But overall a very good tool.

I sought other data to try and see how realistic Valby’s rating is. I looked at past Nuttycombe times by various athletes, and their corresponding BU indoor time that same year:

ATHLETE……….Nutty……..BU
Weini Kelati……19:49…….15:15
Taylor Werner…19:52……15:11
Katie Izzo………..19:59…..15:13
Ednah Kurgat…..19:39…..15:15
Weini Kelati………19:36…..15:15
Allie Ostrander….19:40…..15:16
Ednah Kurgat……20:22……15:38

ATHLETE…………Nutty……..BU
Weini Kelati……..19:49…….15:15
Taylor Werner…..19:52……15:11
Katie Izzo…………19:59…….15:13
Ednah Kurgat…..19:39…….15:15
Weini Kelati………19:36…….15:15
Allie Ostrander….19:40……15:16
Ednah Kurgat……20:22…….15:38
Katelyn Tuohy……19:44…….15:16
Katelyn Tuohy……19:30……..???
Parker Valby………19:17………???

The site doesn’t aim to estimate 5k fitness. I’m sure Bijan will hop on here at some point and clarify, but my understanding is he’s just trying to compare performances and a 5k mark makes sense to most people. It’s not supposed to be a prediction of what they would run on the track (although it’s probably not too far off)

The BU times comparison is an interesting exercise...

Apalachee (ncaa finals so hard effort and only 2 weeks from BU)

Orton 19:25 .... 15:08 (she was targeting 15:00)

Chmiel 19:34 ... 15:27

Hays 19:48 ... 15:33

Larkin 19:47... 15:41

Denner 19:37 ... 15:27

Cohen 19:58 ... 15:33

Delay 19:37... 15:36

Kimelo 19:43 ... 15:43

I applaud analyzing data. Different weather for different years of course. The correlation of this data is not the best at 0.6. Basically a bunch of 5000 times around 15:15 for a range of Nutty times with a 23 second range (ignoring slowest Kurgat time). But I'd conclude their times seem to indicate faster than 15:15 for both Valby and Tuohy - maybe close to 15:00 especially Valby. Too bad no Monson 5000 time with either her 19:33 or 19:39 times - but she did have the 8:45 which points to a 15:10ish 5000 probably? She beat Kelati by 10 seconds at Nutty I think. So bottom line Valby and Tuohy are very fit (duh!).

Per the Daniels Calculator, Monson’s 8:46 at Milrose is equivalent to a 15:14.

The women’s rankings are pretty accurate but the men’s individual rankings are way off.

Lol. Ok, a few things here:

1) Nobody works at LACCTiC. It’s me running code on my laptop when I have time. It costs about 50-60 dollars per month to pay for the server hosting etc. I bring in 5-10 dollars a month from donations. “We” are not a company.

2) LACCTiC doesn’t care about course distances. All it does is look at relative times, which lets you translate between performances.

3) As someone already pointed out (thank you), the goal of LACCTiC is not to predict 5k performances in the spring. The goal is to give one rating that can easily be compared across courses. Instead of choosing a gold standard course or a meaningless number (like with speed ratings), I put everything in terms of 5ks. If you are better at cross country, you might get a rating much better than you ever run on the track.

4) I try to counter the 90% effort problem a few different ways. The first is that I have a way of scoring the “importance” of a race. The second is that I try to base ratings off the best season performances (sometimes including last season or track if the athlete doesn’t seem to have run all out yet). Of course, this is all done computationally. I don’t do this by hand.

-Bijan

It doesn't care what length the course is... it could be a 4k or it could be a 10k. Irrelevant to the algorithm. It just looks comparatively between the people's prior ratings in the race.

You mention coaches holding runners back. Pretty much everyone is trying. You can't make a comprehensive rating system tailored to just the top people. It does great for everyone except for those few people. Also, realize that it weights each performance differently. If someone has a very fast performance, but the rest of their season is meh, the fast performance will be 100% of their weight.

Great site. Most coaches, runners, and fans can't process race results that way. They focus on distance and time. A guy can run 24 minutes and finish 100th and then run 25 minutes and finish 75th against the same field and they think the 24 minute performance is much better.