Highest Tully rating Boys and Girls from Foot Locker. Most competitive fields all year confirmed by Bill.
Speaking of east coast biases. The top 4 boys’ performances were not done by guys from the northeast, but it’s par for the course from Bill. But hey, at least the ratings give XC parents who know nothing about the sport something to talk about.
Running in standing water on a tough course earned a whole 1.5 bump in rating from the year before. That seems totally reasonable.
So Bill’s model is biased against the whole NXN meet or just the east coast kids in that meet? I’m confused.
Speaking of east coast biases. The top 4 boys’ performances were not done by guys from the northeast, but it’s par for the course from Bill. But hey, at least the ratings give XC parents who know nothing about the sport something to talk about.
Running in standing water on a tough course earned a whole 1.5 bump in rating from the year before. That seems totally reasonable.
So Bill’s model is biased against the whole NXN meet or just the east coast kids in that meet? I’m confused.
Well, start with the fact that his entire analysis is based on comparing everyone to the SUNY Utica Tech race course from 1999. Add in that he is most familiar with NY races and athletes, and does analysis for all of them while admitting he ignores top races from other areas (due to time constraints, which is fair) and you start to see where problems occur.
In my opinion, his entire model is inherently flawed. Its entire purpose is to compare one runner to another. To do this, he takes highly individualized athletes with highly individualized performances and attempts to normalize them by applying general (or “standardized”, in his terms) adjustments. The problem with this is that it assumes every athlete responds to different courses and conditions exactly the same. If a course is muddy, it assumes every athlete is an equally strong mudder. If a race is at altitude, it assumes every athlete responds to altitude changes the exact same way. If it’s 95 degrees during the race, it assumes every athlete will be affected the same, and so forth. That simply isn’t reality. Different conditions affect different athletes differently. In my view, what Bill’s analysis shows is not how one actual runner has performed relative to another, but how one hypothetical athlete who is perfectly standard in all things (according to Bill’s own determination) would have performed had they run the same time. To me, that isn’t helpful.
So Bill’s model is biased against the whole NXN meet or just the east coast kids in that meet? I’m confused.
Well, start with the fact that his entire analysis is based on comparing everyone to the SUNY Utica Tech race course from 1999. Add in that he is most familiar with NY races and athletes, and does analysis for all of them while admitting he ignores top races from other areas (due to time constraints, which is fair) and you start to see where problems occur.
In my opinion, his entire model is inherently flawed. Its entire purpose is to compare one runner to another. To do this, he takes highly individualized athletes with highly individualized performances and attempts to normalize them by applying general (or “standardized”, in his terms) adjustments. The problem with this is that it assumes every athlete responds to different courses and conditions exactly the same. If a course is muddy, it assumes every athlete is an equally strong mudder. If a race is at altitude, it assumes every athlete responds to altitude changes the exact same way. If it’s 95 degrees during the race, it assumes every athlete will be affected the same, and so forth. That simply isn’t reality. Different conditions affect different athletes differently. In my view, what Bill’s analysis shows is not how one actual runner has performed relative to another, but how one hypothetical athlete who is perfectly standard in all things (according to Bill’s own determination) would have performed had they run the same time. To me, that isn’t helpful.
I'm still unclear on how you think that the ratings for NXN are biased towards east coast runners. Nevertheless, all the variables are normalized with such large data sets. That's what statistical models do. It's not perfect - it's still just a model - but speed ratings have been proven to be highly predictive. Certainly more than anything else.
Well, start with the fact that his entire analysis is based on comparing everyone to the SUNY Utica Tech race course from 1999. Add in that he is most familiar with NY races and athletes, and does analysis for all of them while admitting he ignores top races from other areas (due to time constraints, which is fair) and you start to see where problems occur.
In my opinion, his entire model is inherently flawed. Its entire purpose is to compare one runner to another. To do this, he takes highly individualized athletes with highly individualized performances and attempts to normalize them by applying general (or “standardized”, in his terms) adjustments. The problem with this is that it assumes every athlete responds to different courses and conditions exactly the same. If a course is muddy, it assumes every athlete is an equally strong mudder. If a race is at altitude, it assumes every athlete responds to altitude changes the exact same way. If it’s 95 degrees during the race, it assumes every athlete will be affected the same, and so forth. That simply isn’t reality. Different conditions affect different athletes differently. In my view, what Bill’s analysis shows is not how one actual runner has performed relative to another, but how one hypothetical athlete who is perfectly standard in all things (according to Bill’s own determination) would have performed had they run the same time. To me, that isn’t helpful.
I'm still unclear on how you think that the ratings for NXN are biased towards east coast runners. Nevertheless, all the variables are normalized with such large data sets. That's what statistical models do. It's not perfect - it's still just a model - but speed ratings have been proven to be highly predictive. Certainly more than anything else.
I never said NXN ratings are biased for east coast runners. I’m saying east coast runners over perform in his ratings because the entire analysis is based on Bill’s focus on and familiarity with east coast runners and courses. His entire analysis is based on comparing performances throughout the country to the course he knows best. In doing so, he completely undervalues mountain west athletes. There isn’t a single performance from Utah’s state meet in the top 200, and only Simmons and JoJo broke 190 there. Part of that can be explained by Utah runners performing better at meets outside the state (which always seems to be the case with Utah athletes under Bill’s ratings). But for arguably the best state with a very deep field to not have one performance from its state championships end up in the top 200 is suspect at best.
People worship speed ratings like they were given from on high. To say they aren’t perfect is quite the understatement.
I never said NXN ratings are biased for east coast runners. I’m saying east coast runners over perform in his ratings because the entire analysis is based on Bill’s focus on and familiarity with east coast runners and courses. His entire analysis is based on comparing performances throughout the country to the course he knows best. In doing so, he completely undervalues mountain west athletes. There isn’t a single performance from Utah’s state meet in the top 200, and only Simmons and JoJo broke 190 there. Part of that can be explained by Utah runners performing better at meets outside the state (which always seems to be the case with Utah athletes under Bill’s ratings). But for arguably the best state with a very deep field to not have one performance from its state championships end up in the top 200 is suspect at best.
People worship speed ratings like they were given from on high. To say they aren’t perfect is quite the understatement.
Can you give examples of the best performances in 2023 that are underrated. What should the top results be?
I never said NXN ratings are biased for east coast runners. I’m saying east coast runners over perform in his ratings because the entire analysis is based on Bill’s focus on and familiarity with east coast runners and courses. His entire analysis is based on comparing performances throughout the country to the course he knows best. In doing so, he completely undervalues mountain west athletes. There isn’t a single performance from Utah’s state meet in the top 200, and only Simmons and JoJo broke 190 there. Part of that can be explained by Utah runners performing better at meets outside the state (which always seems to be the case with Utah athletes under Bill’s ratings). But for arguably the best state with a very deep field to not have one performance from its state championships end up in the top 200 is suspect at best.
People worship speed ratings like they were given from on high. To say they aren’t perfect is quite the understatement.
Can you give examples of the best performances in 2023 that are underrated. What should the top results be?
The Utah State Championships. In the final season rankings, Utah had six runners with ratings above 190. According to Bill though, only two of the six were able to achieve a 190 or above while racing in-state.
This post was edited 5 minutes after it was posted.
Well, start with the fact that his entire analysis is based on comparing everyone to the SUNY Utica Tech race course from 1999. Add in that he is most familiar with NY races and athletes, and does analysis for all of them while admitting he ignores top races from other areas (due to time constraints, which is fair) and you start to see where problems occur.
In my opinion, his entire model is inherently flawed. Its entire purpose is to compare one runner to another. To do this, he takes highly individualized athletes with highly individualized performances and attempts to normalize them by applying general (or “standardized”, in his terms) adjustments. The problem with this is that it assumes every athlete responds to different courses and conditions exactly the same. If a course is muddy, it assumes every athlete is an equally strong mudder. If a race is at altitude, it assumes every athlete responds to altitude changes the exact same way. If it’s 95 degrees during the race, it assumes every athlete will be affected the same, and so forth. That simply isn’t reality. Different conditions affect different athletes differently. In my view, what Bill’s analysis shows is not how one actual runner has performed relative to another, but how one hypothetical athlete who is perfectly standard in all things (according to Bill’s own determination) would have performed had they run the same time. To me, that isn’t helpful.
I'm still unclear on how you think that the ratings for NXN are biased towards east coast runners. Nevertheless, all the variables are normalized with such large data sets. That's what statistical models do. It's not perfect - it's still just a model - but speed ratings have been proven to be highly predictive. Certainly more than anything else.
With respect to your “highly predictive” comment, take at look at Bill’s predictions for NXN teams and individuals from his own site below. His ratings weren’t even close to predicting the outcome with any sort of accuracy.
With respect to your “highly predictive” comment, take at look at Bill’s predictions for NXN teams and individuals from his own site below. His ratings weren’t even close to predicting the outcome with any sort of accuracy.
I'm not sure what your definition of accurate is. He basically gets the top 8 and bottom 8 right, so if you plot his prediction vs outcome, the correlation will be in the 95-97% range, which is pretty good.
With quick scan, his two big misses were Riverton and Herriman, plus cba doing their annual choke. Obviously Herriman outperformed and Riverton underperformed.
A prediction is by definition going to be different than the outcome, that's why they run the race.
I'd have to dig around and look at more Utah races. He sets the results based on where the mid pack runners are vs historical races, and the front runners calculate off of that. One normally assumes that these top national guys just coast in the local races, so big invites, states and the nxn/fl races are where they go all out. You'll note Meylan liked Simmons for the win at NXN, despite his not having fastest rating. The ratings don't rate a runner' s capabilities, rather just their performance in one particular race.
With respect to your “highly predictive” comment, take at look at Bill’s predictions for NXN teams and individuals from his own site below. His ratings weren’t even close to predicting the outcome with any sort of accuracy.
I'm not sure what your definition of accurate is. He basically gets the top 8 and bottom 8 right, so if you plot his prediction vs outcome, the correlation will be in the 95-97% range, which is pretty good.
With quick scan, his two big misses were Riverton and Herriman, plus cba doing their annual choke. Obviously Herriman outperformed and Riverton underperformed.
A prediction is by definition going to be different than the outcome, that's why they run the race.
I'd have to dig around and look at more Utah races. He sets the results based on where the mid pack runners are vs historical races, and the front runners calculate off of that. One normally assumes that these top national guys just coast in the local races, so big invites, states and the nxn/fl races are where they go all out. You'll note Meylan liked Simmons for the win at NXN, despite his not having fastest rating. The ratings don't rate a runner' s capabilities, rather just their performance in one particular race.
Of course you have to run the races. But prediction is what the other poster is hanging his hat on.
Bill was way off on Herriman missing by almost 100 points and three places. His AF, South Lake, and Belen picks were close place-wise but were also obvious. No one needed a computer to tell them they were top 4 teams. He completely whiffed on Niwot, CBA, and Downers. He was actually much closer on Riverton than you give him credit for. His individual ratings were a mess from a predicative standpoint. Three of his top five guys coming in ended up finishing 11th, 20th, and 42nd. He had JoJo and Westfall outside the top-20 in his ratings yet they ended up 1st and 4th. I’m not sure what your definition of accurate is either. I can tell you this though, if the BroJos has done an NXN prediction contest, Bill would have been much closer to the middle of the pack than winning any sort of prize. And LetsRun middle of the pack is not something to be proud of let alone worthy of the adoration speed ratings are given.
Other than Simmons and JoJo, no one else in Utah had the luxury of holding back at state. They were all running for scholarships. This is why I say that Bill is undervaluing that race. This is further highlighted when you consider that he had JoJo rated 22nd coming in even though he had only lost all season to Simmons, the reigning Gatorade national runner of the year. Westfall, whose speed rating was 35th going in to NXN, has a similar story.
Foot Locker Champion Drew Griffith, NXN Champion JoJo Jourdon 1-2 In Final Season Cross Country RankingsBy Keenan Gray and Doug Binder of DyeStatIt isn't easy choosing the top-ranked runner in the nation when there are multip...
Griffith is ranked high based on his win over Pajak. Pajak is ranked high based on his win over Griffith. JoJo is ranked in between because he beat seven of the top 10 including the defending Gatorade runner of the year and the previously number one speed rated runner. I know which one is more impressive to me. If only we could have seen them all in the same race.
Griffith is ranked high based on his win over Pajak. Pajak is ranked high based on his win over Griffith. JoJo is ranked in between because he beat seven of the top 10 including the defending Gatorade runner of the year and the previously number one speed rated runner. I know which one is more impressive to me. If only we could have seen them all in the same race.
Maybe JoJo should’ve run in the national championship then.
Griffith is ranked high based on his win over Pajak. Pajak is ranked high based on his win over Griffith. JoJo is ranked in between because he beat seven of the top 10 including the defending Gatorade runner of the year and the previously number one speed rated runner. I know which one is more impressive to me. If only we could have seen them all in the same race.
Maybe JoJo should’ve run in the national championship then.
Seven of the top 10 ran NXN, according to Dyestat. It’s pretty clear to those with eyes which race was the national championship. It’s been a few years now since FL produced the undisputed national champ. Time to keep up with the times.
