That's one way to do a quick and dirty ("reasonable") estimate (240*0.987 seconds). But I think it overestimates the benefit by about a second, and I'd say 3:58 is the new 4:00. My look at the World Athletics data a few days ago shows that the superspike benefits lessen at higher paces, which his estimate can't account for. It shows what has actually happened to track times.
I looked at 1000th yearly performer times at 800/1500/5000/10000 and compared seven pre-superspike years (2012-18) to 2022 (skipped covid-affected 2019 and 2020; 2021 and 2022 both superspiked years but picked the barely faster 2022 for comparison due more brands with superspikes. Would have had the same results combining 2021 and 2022 as superspike years). It's not a rigorous statistical look, but it was easy to notice that at the 1000th performer level, the data is relatively smooth year to year due to it resting on a ranked list of 999 beneath it. For example, the standard deviation of seven 1500 times from 2012-2018 was only 0.22 second, and the standard deviation of the 800 times from those years was only 0.06 second. (For comparison, if you look higher up the rankings at the100th performer level in the 1500, standard deviation for 2012-2018 is a lot higher at 0.56 second.) The 10000m times were an exception because people started using Vaporflys in 2016/2017 partway through the pre-superpike years.
I made the assumption that that from a physics standpoint, the shoe doesn't know a runner what distance a runner is running, it's just reacting to the forces applied to it. So, I graphed the (log) benefit at different paces and came up with a (highly correlated r2=.9867) regression line, and a table representing that line.
3:58 is the new 4:00. The 0.5 second benefit at 60 second pace was rounded from 0.52 second from the regression line, but I don't want to imply greater precision that I have. And don't forget it's not the same benefit for everyone.
For fun, I'll look up what the World Athletics data shows for the 4:00 mile and give an estimate utilizing other parts of that database than what I've already used.:
80th yearly mile performer looks close to 4:00: 2012 4:01.98 2013 4:00.72 2014 4:00.01 2015 4:00.83 2016 4:01.11 2017 4:00.60 2018 3:59.24 2022 3:57.76 pre-superspike 2012-2018 average: 4:00.64. But the standard deviation is crap at 0.86 second and not usable for the way I'm doing comparisons - so skip this.
3:43.69 1500m is 4:00 mile pace, and pace is what matters to the shoe, so trying the 401st 1500m performer: 2010 3:43.70 2011 3:43.57 2012 3:42.94 2013 3:43.47 2014 3:43.45 2015 3:43.36 2016 3:43.39 2017 3:43.58 2018 3:43.69 2022 3:41.28 pre-superspikes: mean 3:43.46, standard deviation 0.23 second. That's 2.18 second difference at 1500. At the mile, 2.18/1500*1609.344=2.4 second. This aligns close to my table above (0.1 second off). But considering what I'm comparing (mean of 7 numbers to 1 number), I still prefer to round to 2 seconds.
Your detailed analysis is painstakingly done and complex, and appreciated.
But I am missing something. Isn't this correct?A 4:00.00 mile would equate to 3:42.18 for 1,500? Not that 3:43.x you mentioned?
For my purposes, I'm not doing a conversion for effort or worth, just pace equal to 4:00 mile/59.65 per 400, which is 3:43.69. I could just as well do it with 800 times, but World Athletics doesn't log data from 1:59.30 male 800 runners (also 4:00 mile pace) because that time is too slow for their database (except on the women's side... but those are high ranked women with too few numbers to use for this type of comparison). Like I said earlier, from a physics perspective, a shoe doesn't know how far you are running, so that's why I'm grabbing onto pace as representative of the force applied to the shoes. It's all data that can be thrown on a graph of 400 pace vs. benefit (like I have an image of attached above), and since that graph makes a nice line that is highly correlated (r2=0.9867), that's support that 400 pace affects the amount of benefit.
Your estimates are well documented and reasonable. There are confounding factors, however, which include the placebo effect (the runners all know that they are wearing the "superspikes"), the prevalence of pacing lights, which help to even out and hence optimize pace, the increased prevalence of paced races (the BU indoor meets, the Portland Track Festival, the LA meet, the Sound Running meets), and the increased # of older runners in the NCAA still competing instead of retiring and starting a job because of the extra COVID years. In my analyses of indoor times in the winter, there was a major effect of older runners still competing in college and that can affect the world rankings for depth.
The more interesting story really is the claim to a 6% improvement in running efficiency from training in flats. You have to be careful about switching over to flats because of the achilles issues, which I got from workouts in flats about a decade ago, but that's a much larger effect (3x) than superspikes were credited.
The superspikes provide AT LEAST 3 seconds per mile for every distance. The fact that EVERY sub 4, sub 14, etc. runner is wearing them is all the proof you need. People aren't spending 150 dollars on shoes that don't help at all (like some morons claim).
There's literally NOBODY at the top level who isn't using them. You'd think there'd be people that are old school and don't want the extra assistance. There's always runners that want to make things more challenging than they are (running barefoot, ultra marathons, blue jean mile, beer mile).
The elite athletes won't even pretend like the shoes don't make a massive difference.
This happens every time a new spike is released. The previous model gets phased out and everyone races in the new one. Every company now has a “super spike” as their flagship racing shoe, so of course that’s what people are using.
The 3 seconds per mile figure is nonsense. That makes 3:27.20 the new 3:30 and 12:50.7 the new 13:00. So in your view, we’ve had zero sub-3:30s and only one sub-12:50 in the last 4 seasons…
The superspikes provide AT LEAST 3 seconds per mile for every distance. The fact that EVERY sub 4, sub 14, etc. runner is wearing them is all the proof you need. People aren't spending 150 dollars on shoes that don't help at all (like some morons claim).
There's literally NOBODY at the top level who isn't using them. You'd think there'd be people that are old school and don't want the extra assistance. There's always runners that want to make things more challenging than they are (running barefoot, ultra marathons, blue jean mile, beer mile).
The elite athletes won't even pretend like the shoes don't make a massive difference.
This happens every time a new spike is released. The previous model gets phased out and everyone races in the new one. Every company now has a “super spike” as their flagship racing shoe, so of course that’s what people are using.
The 3 seconds per mile figure is nonsense. That makes 3:27.20 the new 3:30 and 12:50.7 the new 13:00. So in your view, we’ve had zero sub-3:30s and only one sub-12:50 in the last 4 seasons…
Just to add to this, compare the last 4 years since super spikes (2019-22) to the previous 4 year period (2015-18).
19-22: 0 sub 3:30, 1 sub 12:50, 12 sub 13:00
15-18: 12 sub 3:30, 3 sub 12:50, 25 sub 13:00
So men’s distance has fallen off a cliff since the super spikes were released.
The superspikes provide AT LEAST 3 seconds per mile for every distance. The fact that EVERY sub 4, sub 14, etc. runner is wearing them is all the proof you need. People aren't spending 150 dollars on shoes that don't help at all (like some morons claim).
There's literally NOBODY at the top level who isn't using them. You'd think there'd be people that are old school and don't want the extra assistance. There's always runners that want to make things more challenging than they are (running barefoot, ultra marathons, blue jean mile, beer mile).
This.The elite athletes won't even pretend like the shoes don't make a massive difference.
This happens every time a new spike is released. The previous model gets phased out and everyone races in the new one. Every company now has a “super spike” as their flagship racing shoe, so of course that’s what people are using.
The 3 seconds per mile figure is nonsense. That makes 3:27.20 the new 3:30 and 12:50.7 the new 13:00. So in your view, we’ve had zero sub-3:30s and only one sub-12:50 in the last 4 seasons…
Exactly Finding old spikes is a pain in the ass. I've worn the "super spikes" they're definitely nice, but nothing special aside from the extra cushion they provide. People are going to buy what's the newest and most widely available, which at the moment happens to be Dragonfly's and Air Zooms.
This happens every time a new spike is released. The previous model gets phased out and everyone races in the new one. Every company now has a “super spike” as their flagship racing shoe, so of course that’s what people are using.
The 3 seconds per mile figure is nonsense. That makes 3:27.20 the new 3:30 and 12:50.7 the new 13:00. So in your view, we’ve had zero sub-3:30s and only one sub-12:50 in the last 4 seasons…
Just to add to this, compare the last 4 years since super spikes (2019-22) to the previous 4 year period (2015-18).
19-22: 0 sub 3:30, 1 sub 12:50, 12 sub 13:00
15-18: 12 sub 3:30, 3 sub 12:50, 25 sub 13:00
So men’s distance has fallen off a cliff since the super spikes were released.
do you know how drugged up running was in 2015-2018? Kiprop, Maflhouki, Centro, etc. That was prime Salazar drug era.
There is variation per person, but the superspikes are proving to match or better the benefits that doping has.
This happens every time a new spike is released. The previous model gets phased out and everyone races in the new one. Every company now has a “super spike” as their flagship racing shoe, so of course that’s what people are using.
The 3 seconds per mile figure is nonsense. That makes 3:27.20 the new 3:30 and 12:50.7 the new 13:00. So in your view, we’ve had zero sub-3:30s and only one sub-12:50 in the last 4 seasons…
Just to add to this, compare the last 4 years since super spikes (2019-22) to the previous 4 year period (2015-18).
19-22: 0 sub 3:30, 1 sub 12:50, 12 sub 13:00
15-18: 12 sub 3:30, 3 sub 12:50, 25 sub 13:00
So men’s distance has fallen off a cliff since the super spikes were released.
Um, you are aware that way, way fewer high-level races occurred from 19-22 due to covid, right?
I think in running we need to give times an Era Adjustment. Just like in Baseball with OPS+ or ERA+ to compare how dominate players were relative to their Era. Like Jim Ryun is still the best American Middle Distance runner of All Time because relative to his time he was so much better than Most Americans and Significantly better than most High Schoolers . While for example Hobbs Kessler may have run a faster time than Ryun, there were multiple High school kids who broke 4 during the same Era. We need to bring this type thinking to track. We already have this in the 100m/200m with wind adjusted times. I think it's time to take a baseball approach to track analytics.
You know what I never see in these analysis? Every sport has better athletes year over year, not allowed in distance running.
Not true. Take baseball, for example. In the 90s and 2000 , you had to hit 55 or more home runs to be the national league home run champion. Now if you hit 40, you win. Players are not as good.
Your estimates are well documented and reasonable. There are confounding factors, however, which include the placebo effect (the runners all know that they are wearing the "superspikes"), the prevalence of pacing lights, which help to even out and hence optimize pace, the increased prevalence of paced races (the BU indoor meets, the Portland Track Festival, the LA meet, the Sound Running meets), and the increased # of older runners in the NCAA still competing instead of retiring and starting a job because of the extra COVID years. In my analyses of indoor times in the winter, there was a major effect of older runners still competing in college and that can affect the world rankings for depth.
The more interesting story really is the claim to a 6% improvement in running efficiency from training in flats. You have to be careful about switching over to flats because of the achilles issues, which I got from workouts in flats about a decade ago, but that's a much larger effect (3x) than superspikes were credited.
Thanks for this post. Those factors I boldened are huge reasons for the increase in sub-4 miles in the NCAA. I’m sure new spikes are faster than older spikes, but people are terrible at entertaining more than a single superficial variable.
Only for amateur high schoolers who run 4:00/4:20 with the build and coordination of a bag of spoiled milk, maybe not for the absolute specimens of humanity like Kipyegon, Athing, Ing, et al
I know he's just reporting something that was presented but these conversations are nauseating. Just shut the f--k up on this topic already.
All these assumptions - and that's what they are, are so dependent on other factors it makes the conversation almost pointless. These numbers are all related to running economy savings, but there are multiple other variables involved I don't know how any "scientist" can settle on a figure.
The first one is the track surface itself and if Tokyo (and closer to home the track at BU) hasn't shown this already, not all tracks are created equal. The resiliency of the surface has a far bigger impact on elastic return (key component of this "%" figure) and unlike running on asphalt where there is basically little to no variation when and where you run, tracks have large variance.
Another factor is the athlete that actually has to run in these spikes. A lot of people inside the industry suspect that the biggest benefit of these well cushioned spikes (which is what they really are) is the ability to train in them with the same level of performance as "dated" elite racing spikes such as the Vic/Matumbo without destroying the lower libs through impact fatigue. Where this gets difficult to apply a quantitative benefit to, is knowing for example if the athlete actually trains in the spikes (vs flats) and what amount of damage (as all training is a form of controlled damage) they actually impart on themselves (ie. do some athletes work out more on grass vs a track etc).
And to cap it off unfortunately just looking at yearly performances en masse, while interesting, just isn't overly helpful in this particular case.
You can say "hey look since 2019 the performances of the xxxxth athlete have dropped this much therefore" but again - soooo many factors at play. An era of higher aspiration? (J.I showing that a young European can be the best in the world). An era of greater depth which lifts everyone from the top right through to high school ranks. Athletes training faster through to other product that utilize similar tech. Covid-19 which sparked mass desperation just to even compete which has just snowballed post that era. It's impossible to quantify.
Quite simply, new product has made a difference - I'm not saying it hasn't. It's had an impact on the road and it's having an impact on the track. But attaching a blanket figure to products (which btw differ across brands so who is at the top of x% value and who isn't) is simply impossible and also not fair to a lot of athletes where the narrative becomes "oh you are only doing this because...."
It's almost like it's the convenient narrative of jilted retired 4.02 miler that wants to feel better about high schoolers out there running 5 seconds faster than he ever could. The tech has improved the sport like it should. Like EVA did. Like mesh uppers did. Like synthetic tracks did. Can we just leave it at that?
The superspikes provide AT LEAST 3 seconds per mile for every distance. The fact that EVERY sub 4, sub 14, etc. runner is wearing them is all the proof you need.
This guy should be disqualified from this discussion. All the fast dudes wore Victories and Matumbos 10-15 years ago, was that in and of itself proof that they provided 3+ seconds per mile?
Final word on this thread: it figures that Alex Hutchinson was a 3:42.43 1500 guy (3:43.3 per WA profile, but states 3:42.43 in an interview). 3:42.43 is a 4:00.22 equivalent.
This post was edited 1 minute after it was posted.