The NCAA data is in - The super spike effect is real

Learn python, LOL. Probably overkill for this scenario.

captainwildcat wrote:

This is FAR from a detailed analysis

1) You only compared one year to 2021 (was 2019 a slower than normal year? Faster than normal? - I don't know)

2) Are all the fastest times run with the "super spikes" (yes, no, some were some weren't)

3) Comparing ONLY the fastest times doesn't take into account that there may be an outlier. A better analysis would be comparing the top 10 or top 15 fastest times. Here again you woud need to look how many of those top runners in 2021 were wearing the super spikes.

You MAY be correct about the super spikes, but your analysis doesn't provide the evidence

His analysis ready is very bad.

He only compared the 50th time? Wtf?

A much better method would be to take the average of the top X runners (his choice of 50 seems like a good choice). Then compare more years. I'd go with at least 10 years to get any meaningful results to run through and see if there is anything of statistical significance.

I wouldn't be shocked if there was, but his current chart just looks like noise. But even there was anything significant, you still have to either accept or try to account for the fact that covid year meets have been drastically different than normal years, so any improvement could be solely from that. The 2022 and 2023 years will likely yield better data without noise from covid.

Do I personally think the spikes make a difference? Yes. But even if Rojo's hack stats there are 100% valid and 100% from the spikes, they must be in the range of .3% to .5% of a benefit.

Correlation =/= causation.

The correct way to do this is a difference-in-difference model comparing 2019 vs 2021, as well as 2018 vs 2019, 2017 vs 2018, etc. Is the trend over the last two years a departure from previous trends?

You couldn't really adjust for this statistically, but the prolonged period of training and the ability to take online classes while training at altitude are surely contributors, as is the 32 athlete field size, as someone mentioned before.

Really sloppy from a statistical standpoint, Robert. To the point where you can't honestly draw many conclusions from it. Athletes are running faster this year, for a variety of reasons. Please subscribe because we need the money.

rojo wrote:

I've crunched the numbers and compared the 2021 and 2019 NCAA times and the distance times are faster in 2021, while the sprint times are not, showing the super spike impact is real.

https://www.letsrun.com/news/2021/05/the-ncaa-data-is-in-the-super-spike-effect-is-real-starting-in-the-1500/

(Sub required).

This is pathetically bad. I thought it was a troll post.

If anything, this disproves the super shoes theory because the longest distance measured has a net decrease in performance when combining both genders.

You can’t say it doesn’t work for sprints, works for middle distance, and then makes you slower for 10k.

rojo wrote:

It would take weeks if not months to do properly.

There's no way a free, ad-supported site could fund that kind of study.

statsprofessor wrote:

My gosh, Rojo just keeps finding new lows.

See if you can hold two independent thoughts (that may seem conflicting) in your head at the same time:

1. Nike's new spikes are the best spikes that have ever been made for distance runners

and

2. They're a natural progression of technology, and do not give an advantage so large as to wiping out the record books.

By your constant criticisms, I can only assume that the only records you would consider are those on cinder tracks with leather shoes.

Fwiw, I think WAVELIGHT is a MUCH bigger advantage for races like the 5k/10k.

Finally, your "statistical proof" proves only one thing, that you don't understand statistics.

They’re a natural form of progression the same way epo is a natural form of progression.

The far fewer number of meets means longer blocks of uninterrupted training. I'd expect times to be much faster.

bugboy98 wrote:

This is preposterous. A high school stats student could do it with a t-test on his/her graphing calculator in an 8-hour work day. Nobody is expecting you to spend weeks/months to get a perfect sample. It just seems weird to claim that you found data to prove that the shoes work and not do any work to back it up. If you had sent a google form to 10 coaches from non nike schools to see if any of their athletes used the shoes and then looked at how those runners' times changed from 2019 to 2021 compared to runners from nike schools, you would actually have something to report. Instead you used data that shows nothing because it doesn't have any control group and not enough data points, and then you defend it like its your life's work.

Yes if he happens to get every coach in the nCAA to fill out a form telling me what shoes their athletes are waering, then it takes no time at all. So if he gets all of the coaches to do the research that would take him weeks, then you are right.

To all of the haters who have been basically saying the same thing which is like the quote below - please answer the following:

duck sauce wrote:

i'd probably guess something along these lines if you asked me how much faster the #50 mark would be after ~9 months of no racing and with 5 full classes of guys with eligibility rather than 4

Why haven't the sprinters also improved in similar fashion?

Game. Set. Match. Rojo.

PS. I never tried to claim this was some scentifically valid study. I just literally thought, "Let's look at the 50th time. WOw, they are faster this year versus 2019 in the distance races but not the sprints (yes the 10k is messed up)."

rojo wrote:

Bad Wigins wrote:

Don't you get revenue from shoe companies advertising here? Conflict of interest.

God people are stupid. We've received money directly from two shoe companies - neither of whom has a super shoe. I'm literally promoting the technology of Nike.

+2600

Isn't it possible the sprinters will add a lot more top times at their conference meets and in the championship season than the distance runners? Feels weird to not wait until the championships season where sprint times >>> distance times.

Sorry for piling on with my snarky comment. I agree that the shoes seem to work as advertised. Wish they were really for sale to the public so I could experience higher speeds again.

rojo wrote:

To all of the haters who have been basically saying the same thing which is like the quote below - please answer the following:

duck sauce wrote:

i'd probably guess something along these lines if you asked me how much faster the #50 mark would be after ~9 months of no racing and with 5 full classes of guys with eligibility rather than 4

Why haven't the sprinters also improved in similar fashion?

Game. Set. Match. Rojo.

PS. I never tried to claim this was some scentifically valid study. I just literally thought, "Let's look at the 50th time. WOw, they are faster this year versus 2019 in the distance races but not the sprints (yes the 10k is messed up)."

I'm sure you're just baiting all of us.

You're choice of a control group being "sprinting events" is not appropriate. How is that relevant? Sprint times are not comparable to distance times in many ways. They are much more affected by weather. They can come at almost any time of the season. And sprinters have many chances to run their best time each season. You don't have to travel across the country to run a fast 100m. Plus, I would guess that age is much less of an advantage in sprinting. Plus you haven't shown that sprinters haven't improved. Your sample size is too small and you're not controlling it for enough variables.

I'm actually interested in your analysis and I appreciate the effort. I just can't tell if you really want to do a good job or if you're intentionally just trying to get us excited.

We all know the shoes work. You really just need to see how many people are wearing them to prove that. Runners know what makes them faster and they wouldn't wear them otherwise.

juvie wrote:

rojo wrote:

I did note it. I can't help it that you didn't actually read the article.

So far we have a bunch of criticisms from people who haven't even read the article.

Lol. Tough crowd.

Not really. Science requires a rigorous defense of the thesis, and,most important of all, replication of the original data and analysis.

Rojo might be wrong, but prove it first.

sunflower wrote:

I'm actually interested in your analysis and I appreciate the effort. I just can't tell if you really want to do a good job or if you're intentionally just trying to get us excited.

No, here is what happened. Rojo had a good thought that he could use the descending order lists to try to quantify the shoe impact. He also went deep enough to exclude outliers, which was another good thought.

But what you saw was the best he could produce because he has no analytical training and wouldn't know what to do next.

And then he got defensive when everyone pointed out this is a really weak analysis that doesn't prove what he thinks it does.

TL;DR: There are many possible effects.

If Skip Bayless can make a career of poorly researched but assertively voiced opinions, I see no reason why Rojo can't. Unfortunately, I think there are a lot of factors that may or may not have contributed to the relatively fast mid-distance times this year. I've done exactly as much research as Rojo on this (i.e., none) but here are a few possibilities. I'll leave the causal determinations to smarter people.

1. MORE RUNNERS THIS YEAR. Covid means that there are far more 5th and 6th years running this year than any other year. At the bottom I've included all of the sub-3:40 men this year with their high school class and/or birthdate. I'm not making any judgments, but it would be interesting to know how this distribution compares to typical years.

2. AGE. Not only are there more runners, but assuming that runners get faster with age, we might expect more runners near the top of the regional qualifiers lists. Another effect is that even the true underclassmen now have the advantage of being in races with faster runners. I'm not sure how severe that effect is, but if you looked at high schoolers or pros you might avoid the effect entirely.

3. UNINTERRUPTED TRAINING. There has already been quite a bit of discussion about this. Yes, the sprinters had uninterrupted training too, but I'm not so sure a year of lacking consistent access to tracks for training or competition would be advantageous for them.

4. CROSS COUNTRY IN THE WINTER. Not sure if this is beneficial or not, but it seems hard to control for.

5. REGIONAL QUALIFYING RULES. Usually most athletes in the top 50 nationally can train through the regular season without worrying a whole lot about getting regional qualifying marks. It seems reasonable that the threat of only 36 athletes in each region qualifying motivated some coaches to get their athletes running fast early in the season. I would think 50th place is about the worst place to look for this reason. You really need to look at more of the distribution of times.

6. OLYMPIC TRIALS. Athletes in the top 50 might also be chasing Olympic Trials standards.

7. SPIKES IN THE 800. Are we saying that the new spikes aren't helpful here?

8. 10K RACE OPPORTUNITIES. Before the season, I predicted that the 10k regional marks wouldn't be quite as competitive this year simply because there was no Stanford Invite (at least not in the traditional sense) and Payton Jordan. I'm actually surprised the men's 10k has been so competitive this year. Either way, it's really hard to say that the spikes conclusively made everyone faster when the 10k ``effect'' is so significantly negative.

And probably many other effects that I haven't considered... Either way, I can't complain; these times have been fun to watch, and we should be excited about this generation of college runners.

3:35.96 - Nuguse, Yared ( hs: 2017)

3:35.97 - Teare, Cooper ( hs: 2016)

3:36.00 - Kipsang, Eliud ( dob: Sept, 1996)

3:36.47 - Hocker, Cole ( hs: 2019)

3:36.53 - Suliman, Waleed ( hs: 2017, dob: Sept, 1998)

3:36.97@(1) Clinger, Casey ( hs: 2017)

3:37.17 - Garcia Romo, Mario (dob: June, 1999)

3:37.21@(1) Bons, Lucas ( hs: 2018)

3:37.55 - Franco, Talem ( hs: 2013)

3:37.66 - Adams, Ryan ( hs: 2015, dob: July, 1996)

3:37.73 - Ryan, Paul ( hs: 2015, dob: May, 1997)

3:37.76 - Comber, Casey ( hs: 2015, dob: Nov, 1996)

3:37.96 - Mantz, Conner ( hs: 2015)

3:38.07 - Knevelbaard, Kasey ( hs: 2015, dob: Sept, 1996)

3:38.09 - Herrera, Eduardo ( hs: 2016, dob: Sept, 1997)

3:38.17 - Hunter, Charlie (dob: July, 1996)

3:38.29 - Brandt, Robert ( hs: 2015)

3:38.52 - Kioko, Athanas (dob: June, 1995)

3:38.58 - Tanner, Sam (dob: Aug, 2000)

3:38.60 - Dolan, Sean ( hs: 2018)

3:38.72(1) Salisbury, Jack ( hs: 2016)

3:38.79 - Fogg, Adam (dob: Jan, 1999)

3:38.87 - Hamlin, Jesse ( hs: 2019)

3:39.02 - Grijalva, Luis ( hs: 2016)

3:39.03 - Zarate, Diego ( hs: 2015, dob: Sept, 1997)

3:39.17 - Anstey, Jack ( hs: 2016, dob: Apr, 1997)

3:39.27 - Ratcliffe, Thomas ( hs: 2016)

3:39.43@(1) Troutner, Aidan ( hs: 2018)

3:39.44 - Stallings, Zach ( hs: 2017, dob: Aug, 1998)

3:39.49 - Stanovsek, Mick ( hs: 2015)

3:39.73 - Tubby, Brandon ( hs: 2016)

3:39.79@(1) Osterstock, Nate ( hs: 2017)

3:39.85 - Davis, Jonathan ( hs: 2016, dob: Oct, 1997)

3:39.87 - Facioni, Zach ( dob: Apr, 1999)

3:39.91 - Bedard, Simon ( dob: Jan, 1997)

3:39.93 - Lagat, Festus (dob: Oct, 1996)

I claim there is a pink unicorn orbiting Neptune. I might be wrong, but prove it first.

Also rojo do you know what a p-value is?