NYTimes research suggests that Nike Vaporfly may actually be 4% faster based on Strava data

I am not sure how much faith people can put in Strava data being all that accurate. Does anyone know the sampling frequency on the GPS that a persons watch or phone operates? I am sure it works fine if going in a straight line on urban streets. My guess is it is far less accurate if someone is running on winding trails as the sampling frequency is probably not all that great in order to preserve your phone's / watch's battery.

Epson error code 0x69

https://fixerrorcode0x69inanepson.wordpress.com/

Almostsbs wrote:

This is dumb... I run all my easy and recovery runs in Hokas . They would appear almost two mins slower... Many folks only race or tempo in 4%

You're dumb. Read the article.

denisbill123 wrote:

I am not sure how much faith people can put in Strava data being all that accurate. Does anyone know the sampling frequency on the GPS that a persons watch or phone operates? I am sure it works fine if going in a straight line on urban streets. My guess is it is far less accurate if someone is running on winding trails as the sampling frequency is probably not all that great in order to preserve your phone's / watch's battery.

Good understanding of GPS here, but not so much of statistics.

agip wrote:

Didn't read the study, but my main concern is lack of accurate 'I wore shoe X' in Strava.

I'd suspect that there is a significant failure rate - of people not entering the shoe they ran in that day.

Probably there is some correlation between hyper careful people about entering what shoe they wore in their running log...and those who buy $250 shoes..

not sure exactly what I'm saying, but probably there is some significant error in self reporting what shoe was worn.

did the piece talk about that?

I still don't fucking understand why people insist on giving their opinion about an article they didn't read! Instead of asking questions and making assumptions, don't be an ignorant douchetwat and read the goddamn article....

I thought the study was done very well considering all the variables involved and seemed to indicate a fairly clear trend. Did not expect that, but did make me more open to try out different shoes that I previously had no interest in (newtons, adidas and even nike's).

Because letsrun.

4% is huge (about 5 minutes off of a 2:06 marathon) if this is anywhere near remotley true, than all Nike athletes must suck because they haven't rewritten the record books in these new shoes.

More likely, Nike shill sends this "study" to lazy NYT reporter and tells him to take credit. Lazy NYT reporter gets NYT to publish this "study" (Nike Ad).

Nike sells more shoes.

It amazes me to this day how many of the genuine intelligent track fans on the board fail to realize that Breaking 2 was just an elaborate shoe commercial.

ex-runner wrote:

stfu JS wrote:

I still don't fucking understand why people insist on giving their opinion about an article they didn't read! Instead of asking questions and making assumptions, don't be an ignorant douchetwat and read the goddamn article....

I thought the study was done very well considering all the variables involved and seemed to indicate a fairly clear trend. Did not expect that, but did make me more open to try out different shoes that I previously had no interest in (newtons, adidas and even nike's).

Because letsrun.

guilty as charged

but

I did so because it's easier to ask you people points of detail than to spend time reading the article.

I scanned the article but didn't feel like spending 20 minutes carefully reading it.

Much of the internet's point is to summarize and give nutshell detail. Wikipedia may be man's greatest creation and that's what it does.

you people are my wikipedia. I asked a quick question, someone gave a quick answer, and done.

My first glance on the study is the shoe is 1% faster than the next fastest shoe in the study. Why that shoe is the Nike Streak and not some other shoe I'm not sure about. Most racing flats are going to be faster than standard trainers. I was surprised more racing flats didn't stand out in this study.

If people blindly don't want to believe the shoe doesn't work because they are antiNike or because Nike in the past marketed running shoes to the masses that weren't about performance at all (think something like Nike Shox) that is fine but don't comment on here. Read the article and then comment

Just read the reviews of LetsRun on the shoe:

They are very positive.

I think it is worth looking at why other racing flats didn't do that well in the study.

I think one thing is for the marathon you need more cushion. I used to try and race the marathon in the lightest flattest shoe I could (what would have been the lightest NB shoe in 2004? Was it the 1500?) and looking back I totally see I needed more cushion. My feet were used to training on dirt and then I went and ran on roads with no cushion. No wonder I was terrible at the marathon.

I find it interesting that there is all this study on a shoe you can't buy. Nike doesn't even have a release date on their own website. The shoe is unavailable at every major chain or running warehouse.

You should read the article. It doesn't claim that it's 4% better than every shoe out there. In fact, it's only "1% better" than then than the next best shoe (Nike Streak). So if you're already running in a fast shoe, your gains will be minimal, but I think the chart listing all of the shoes in the study is very interesting and valuable. If you have a "lower performing" shoe, you might get some benefit by upgrading to a "faster" shoe even if it's not the Vaporfly. Regarding the statistics, I think they did a solid job and did their best to account for biases and imperfect data. There is discussion about the weaknesses of the analyses, and the authors don't claim it's perfect. I agree with your points about advertising, and concur that much "journalism" is thinly veiled marketing. Accepting that, I feel like this analysis was legit and robust. If nothing else, it's a good starting point for discussion and further research. Cycling is full of data analysis and incremental optimization of products. I like the fact that someone is taking a similarly-analytical approach to running gear. Skepticism is valid and encouraged, but we can ask questions while still being supportive of this kind of research, rather than making emotionally-based attacks.

If true, Kipchoge is horrible wrote:

4% is huge (about 5 minutes off of a 2:06 marathon) if this is anywhere near remotley true, than all Nike athletes must suck because they haven't rewritten the record books in these new shoes.

More likely, Nike shill sends this "study" to lazy NYT reporter and tells him to take credit. Lazy NYT reporter gets NYT to publish this "study" (Nike Ad).

Nike sells more shoes.

It amazes me to this day how many of the genuine intelligent track fans on the board fail to realize that Breaking 2 was just an elaborate shoe commercial.

Someone mentioned weight, which makes sense up to a point. I would think if that were to dominant factor, there would be a grouping of lightweight shoes on the higher end, but they spread out well.

Even though the research makes a valiant attempt to control for selection bias, I still feel there is something in there they aren't catching. That's not to say the shoes near the top of their scale are bad, but they are pretty consistent across metrics. Something is going on, as this clearly isn't random noise.

Now that they've done this little study, I'd love to see a breakdown of each shoe (weight, flexibility, cushion, mileage, etc) and overlay that with what they have and see if something sticks out.

Aaaaand... It just hit me: mileage. The VFs are pretty much raceday only shoes, and fairly new at that. So the new question is what happens when you take into account the age of the shoe? I doubt there are many examples of have people running a race in VFs with 100 miles on them. (I don't remember seeing anything in the article about that, but let me know if I missed it.)

All that being said... With Strava collecting such vast amounts of data, I wonder what else we can learn?

agip wrote:

Didn't read the study, but my main concern is lack of accurate 'I wore shoe X' in Strava.

I'd suspect that there is a significant failure rate - of people not entering the shoe they ran in that day.

Probably there is some correlation between hyper careful people about entering what shoe they wore in their running log...and those who buy $250 shoes..

not sure exactly what I'm saying, but probably there is some significant error in self reporting what shoe was worn.

did the piece talk about that?

I would think the big error in reporting comes from Strava demanding a default shoe. From the moment you enter a shoe, a default needs to be selected and that shoe is assigned to every run automatically. You have to manually change it to a different shoe. This could explain we there are so many heavier trainers listed in the study. I specifically added a "default shoe" so I don't unintentionally attribute a run to the wrong shoe and easily identify the ones I need to edit.

Good points. Yes, there will always be selection bias unless the treatment (which shoe) is assigned randomly, i.e. the study is an RCT. It's unavoidable. There will always be RCT purists, but RCTs have short comings as well. Particularly, they often have small sample sizes which are typically not representative of the entire population. For example, a sports science study researching the impact of LT training might be based on 30+ college aged women with limited training history. The results of that study cannot necessarily be applied to 40 year old male who's been running competitively for 20+ years.

While RCTs are certainly the gold standard for identifying causal effects, in real world practice, large retrospective database analyses compliment RCTs and may even have advantages of RCTs. Here this study included a broad spectrum of individuals fast runners that represent a significant share of runners actually completing marathons.

Additionally, in many instances it is not feasible to conduct a RCT for either ethical reasons or in this case for practical reasons... how can you randomize the shoes without the athletes knowing what shoe they are wearing? It won't be a blind RCT so there will be some placebo effect, although with less selection bias.

Thus while the retrospective study may not be the perfect tool, it may often be the only available tool, and accordingly produces the best available evidence. In other words we have to work with what we got and should approach such evidence with curious skepticism. The spirit and design of this study I would argue represents just about the best possible way to address the question of how effective are the VF 4%s given the existing data and inherent challenges. This study combined with the laboratory study conducted at Colorado make a compelling case for the effectiveness of the VF 4%: The shoe is somewhere between 1% to 4% better than the Nike Streak. (1%-4% is a really large range however).

I agree. I am very curious to see what else we can learn from Strava. Does anyone have experience using their API to pull data? How difficult would it be to pull all the activities from all participants (on Strava) in the 2018 London Marathon for example?

I am not an author of this study but design and evaluate data from RCTs, retrospective studies and databases for a living.

I did not proofread this post. Back to work.

your problem was the 1500 can't bounce off the floor and practically touch the ceiling like the vaporfly

https://video.twimg.com/ext_tw_video/1007958930595565568/pu/vid/360x640/6Ohbn53rmY-O0p0Y.mp4

mad props for doing a marathon in the 1500s, once did it in asics hyperspeeds with much regret after mile 20

They should have compared the 4% with more racing flats such as Saucony Type A and Mizuno Ediken.

wejo wrote:

Urban B wrote:

I still don't see anyone talking about the OTHER shoe in this study that also stands out.

Anyone run in the Streak? And why this one and not one of the other more standard racing shoes? I feel like the study is still missing something.

My first glance on the study is the shoe is 1% faster than the next fastest shoe in the study. Why that shoe is the Nike Streak and not some other shoe I'm not sure about. Most racing flats are going to be faster than standard trainers. I was surprised more racing flats didn't stand out in this study.

[/quote]

This^

Well, Streak is more than 10% and 20% lighter than both NB 1400 and 1500 respectively. 1500 is a lightweight trainer, whilst Streak is a facing flat, not a fair comparison between the two.

Interesting...switching to Streak has a better chance of getting a pb than to the 4%.

I am not sure how much faith people can put in Strava data being all that accurate. Does anyone know the sampling frequency on the GPS that a persons watch or phone operates? I am sure it works fine if going in a straight line on urban streets. My guess is it is far less accurate if someone is running on winding trails as the sampling frequency is probably not all that great in order to preserve your phone's / watch's battery.

https://printertechsupportnumbers.com/blog/how-to-fix-hp-printer-offline-errors-in-windows/

morpheusss wrote:

What do people think?

https://www.nytimes.com/interactive/2018/07/18/upshot/nike-vaporfly-shoe-strava.html

I think that if you are an elite marathoner, the Vaporfly might give you a tiny edge.

I think everyone else would be wasting their money.

The best way to get faster is to train to run faster.