I selected a number of runners from a few schools and only some had even a single had a result listed in between 2024 XC and 2025 XC - from Indoors. Maybe one had two results. Nothing for the others Indoors or Outdoors. So most runners who progressed since 2024 XC are not reflected in the ratings. A lot of results are missing.
I looked at Hartman, Michalak, Putman of NC St - nothing
Hartman has two results.
yes, I see a week later she does now have the 2 5000 m results. Of course between Putman, Michalak, and Napoleon there is still not one. So those results remain elusive. Only certain meets seem to have been added. And it is not just NC State. Perhaps it just takes time.
I'm sorry there are missing track performances. I've tried to run things a few times but theres just a lot of limitations put up by TFRRS right now. To get those PRs, I need to visit every runner's profile. A lot of profiles are not linked to track performances. I'll try to scrape track performances again this weekend.
I've had a lot of people complain that the 5k ratings don't match up with 5k abilities, particularly on the men's side. Heres what I can say about this: Linking 5k PRs to cross country performances is a little complicated. If you consider a list of 5k PRs to be one big race then you are linking AVERAGE performances to track PRs (not average). Right now I correct for this bias. It seems to work well for women (I think lacctic predicted BU 5k times within a few seconds for the past few years). For men, it does occasionally seem slow for the top end of the field. I think the reality might be that a lot of top-end men are just not running all-out for the entire race. I also don't think we would expect all 30 13:30 guys to be running PR-level races 4 weeks into the season.
I'm sorry there are missing track performances. I've tried to run things a few times but theres just a lot of limitations put up by TFRRS right now. To get those PRs, I need to visit every runner's profile. A lot of profiles are not linked to track performances. I'll try to scrape track performances again this weekend.
I've had a lot of people complain that the 5k ratings don't match up with 5k abilities, particularly on the men's side. Heres what I can say about this: Linking 5k PRs to cross country performances is a little complicated. If you consider a list of 5k PRs to be one big race then you are linking AVERAGE performances to track PRs (not average). Right now I correct for this bias. It seems to work well for women (I think lacctic predicted BU 5k times within a few seconds for the past few years). For men, it does occasionally seem slow for the top end of the field. I think the reality might be that a lot of top-end men are just not running all-out for the entire race. I also don't think we would expect all 30 13:30 guys to be running PR-level races 4 weeks into the season.
Lacctic is great. Don’t listen to the haters. Is it perfect? No. is it awesome to have, yes.
I was not so much criticizing as noting the absence of some data, which if included would affect athlete's ratings, and thus the team rankings. It is a great tool, but it does require data. No doubt by November the absence of the track results will be far less significant.
My understanding is that it takes hours for lacctic to sync with tfrrs. After reading thousands of records, the site needs to process data (in the right format), recalibrate results (in addition to new races, previous races would be slightly affected, a little like AI as we know more about each athlete every time they race), before it can export the most recent results. Somehow the race director of Prenat forgot to upload results to tfrrs until this morning. I anticipate seeing updated results with prenat at lacctic tonight or tomorrow morning. The critical question is a 15:04 TiC for Jane Hedengren, over or under?
. The critical question is a 15:04 TiC for Jane Hedengren, over or under?
It should be under, no question.
We said there was an ~31 sec difference, for the low sticks, between Nutty and the Gans Creek course, giving Hartman an ~18:59 PreNats time. Longisa was awarded a TiC of 15:22 at Gans Creek running the previous 19:07 course record. Hedengren ran 25sec faster than Longisa’s time, so her TiC should be at least ~22sec faster than Longisa’s, giving her a TiC of ~15:00.
Lacctic is so inaccurate. It does not factor in head-to-head results.
For example, UVA beat Colorado handily a few weeks ago but Colorado is ranked higher nationally according to lacctic.
Then Colorado only beat Syracuse by 2 points at Nuttycombe, but Colorado is ranked #3 and Syracuse is ranked #7
It also gives the same course much different conversations...ON THE SAME DAY. And from week to week despite identical conditions.
At Gans Creek a few weeks ago, 14:00 was 23:20 for the Black Race and 23:30 for the gold race. Same conditions. Once pre-nats gets analyzed, it's going to be different as well on the same course.
I’ve already been over this, stating to get the most out of LACCTiC in early season requires measured manual intervention, per my initial posts on the following thread:
Lacctic is so inaccurate. It does not factor in head-to-head results.
For example, UVA beat Colorado handily a few weeks ago but Colorado is ranked higher nationally according to lacctic.
Then Colorado only beat Syracuse by 2 points at Nuttycombe, but Colorado is ranked #3 and Syracuse is ranked #7
It also gives the same course much different conversations...ON THE SAME DAY. And from week to week despite identical conditions.
At Gans Creek a few weeks ago, 14:00 was 23:20 for the Black Race and 23:30 for the gold race. Same conditions. Once pre-nats gets analyzed, it's going to be different as well on the same course.
Of course it factors in head-to-head results, it factors in all the results...
If you'd bother to look at the results you would see Colorado underperformed by 67 and Syracuse overperformed by 83. Similarly the week prior, Colorado had a bad day and Virginia had a good day. You rank teams assuming they show up and have an average to good day.
Why would you think different races on the same day would get the same ratings? Look at prelim heats for a 1500 at a conference meet. One heat might go hard from the gun while another goes tactical. The B heats in XC usually get a slower course rating because you don't have a lot of guys going out hard dragging the field with them.
LACCTIC is an excellent and interesting model. No, it is not 100% accurate in ratings or simulations, but it is extremely good. None of the many established stats models for other sports like MLB, the NFL, or the NBA are 100% accurate either. If people like you - the guy who says “Lactic is so inaccurate” - keep complaining and ignorantly criticizing it then the guy who runs it, for no financial benefit to himself, is probably going to stop doing it.
Lacctic is so inaccurate. It does not factor in head-to-head results.
For example, UVA beat Colorado handily a few weeks ago but Colorado is ranked higher nationally according to lacctic.
Then Colorado only beat Syracuse by 2 points at Nuttycombe, but Colorado is ranked #3 and Syracuse is ranked #7
It also gives the same course much different conversations...ON THE SAME DAY. And from week to week despite identical conditions.
At Gans Creek a few weeks ago, 14:00 was 23:20 for the Black Race and 23:30 for the gold race. Same conditions. Once pre-nats gets analyzed, it's going to be different as well on the same course.
Then build something better yourself genius. Not like this guy is getting paid for this. So entitled.
Lacctic is so inaccurate. It does not factor in head-to-head results.
For example, UVA beat Colorado handily a few weeks ago but Colorado is ranked higher nationally according to lacctic.
Then Colorado only beat Syracuse by 2 points at Nuttycombe, but Colorado is ranked #3 and Syracuse is ranked #7
It also gives the same course much different conversations...ON THE SAME DAY. And from week to week despite identical conditions.
At Gans Creek a few weeks ago, 14:00 was 23:20 for the Black Race and 23:30 for the gold race. Same conditions. Once pre-nats gets analyzed, it's going to be different as well on the same course.
Anytime you apply math analysis to sports, humans tend to get in the way of the accuracy.
Lacctic is a great tool that uses track 5k and recent xc results to predict future races.
It can't factor in injuries, improvement, who actually runs the race on race day, and how runners perform in the upcoming meet - most importantly the ncaa meet.
I'd still much rather have lacctic than not, as it gives a 'data driven' view of how teams line up.
Lacctic can't tell US which mens team will actually run great at nats: UVA, Co, or Cuse nor can it tell US what to expect from Bunnage, Olemomoi, Noe, or Hutchins.
Lacctic is so inaccurate. It does not factor in head-to-head results.
For example, UVA beat Colorado handily a few weeks ago but Colorado is ranked higher nationally according to lacctic.
Then Colorado only beat Syracuse by 2 points at Nuttycombe, but Colorado is ranked #3 and Syracuse is ranked #7
It also gives the same course much different conversations...ON THE SAME DAY. And from week to week despite identical conditions.
At Gans Creek a few weeks ago, 14:00 was 23:20 for the Black Race and 23:30 for the gold race. Same conditions. Once pre-nats gets analyzed, it's going to be different as well on the same course.
Of course it factors in head-to-head results, it factors in all the results...
If you'd bother to look at the results you would see Colorado underperformed by 67 and Syracuse overperformed by 83. Similarly the week prior, Colorado had a bad day and Virginia had a good day. You rank teams assuming they show up and have an average to good day.
You proved my point. If Colorado "underperformed" by Lacctic's standards at the two most important meets of the year, then why are they ranked 3rd in the nation by lacctic?????
Lacctic could continue working with its current functionality and it would still be the greatest gift to the collegiate running community since TFRRS. Can't believe the amount of people whining about such a helpful tool that was created by one of our own and not a company seeking to extract profits.
New competitor from USTFCCCA. The XCRI: Cross Country Rating Index. I would guess the model is very similar to Lacctic, though I haven't read much on the methodology. It would be great to hear a breakdown of the differences from the top minds themselves. (This also one has the benefit of being powered directly by athletic.net and eliminates web-scraping, which is probably a huge convenience.)
This is a great addition. I suspect Lacctic will be a better predictive tool, but this may help incentivize competition if it can push teams and athletes to compete more during the XC season.
New competitor from USTFCCCA. The XCRI: Cross Country Rating Index. I would guess the model is very similar to Lacctic, though I haven't read much on the methodology. It would be great to hear a breakdown of the differences from the top minds themselves. (This also one has the benefit of being powered directly by athletic.net and eliminates web-scraping, which is probably a huge convenience.)
New competitor from USTFCCCA. The XCRI: Cross Country Rating Index. I would guess the model is very similar to Lacctic, though I haven't read much on the methodology. It would be great to hear a breakdown of the differences from the top minds themselves. (This also one has the benefit of being powered directly by athletic.net and eliminates web-scraping, which is probably a huge convenience.)
This is FAR more accurate because it factors in each runner individually instead of a team average. As we know, this is more important for XC.
You can have the fastest top 4 in the NCAA that make your team average really fast, but if your 5th runner is in the 200s at NCAAs you aren't winning or coming anywhere close.