A 4 minute mile equals 3:56.8 with superspikes

zzzz wrote:

The Outside article says:

I asked Geoff Burns, a physiologist with the U.S. Olympic and Paralympic
Committee and co-author of Joubert’s study, to run some numbers for me. The
calculations are trickier at track-running speeds than they are for marathoners
(for reasons explained here), but a reasonable estimate is that improving your
running economy by 1 percent will increase your racing speed by about
two-thirds of a percent. So a typical 4:00.0 miler in the old spikes might
expect to run 3:56.8 in the new ones—“corrected for inflation,” as Burns puts
it.

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.

Thread:www.letsrun.com/forum/flat_read.php?th...

old spikes/superspikes/benefit (all in seconds)
55 54.7 0.3
56 55.6 0.4
57 56.6 0.4
58 57.6 0.4
59 58.5 0.5
60 59.5 0.5
61 60.4 0.6
62 61.4 0.6
63 62.3 0.7
64 63.2 0.8
65 64.2 0.8
66 65.1 0.9
67 66.0 1.0
68 66.9 1.1
69 67.8 1.2
70 68.7 1.3 (underestimate? Vaporflys mixed into 10000 data 2016+)
71 69.5 1.5 (Vaporflys in data)
72 70.4 1.6 (Vaporflys in data)

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.

zzzz wrote:

The Outside article says:

I asked Geoff Burns, a physiologist with the U.S. Olympic and Paralympic
Committee and co-author of Joubert’s study, to run some numbers for me. The
calculations are trickier at track-running speeds than they are for marathoners
(for reasons explained here), but a reasonable estimate is that improving your
running economy by 1 percent will increase your racing speed by about
two-thirds of a percent. So a typical 4:00.0 miler in the old spikes might
expect to run 3:56.8 in the new ones—“corrected for inflation,” as Burns puts
it.

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.

Thread:www.letsrun.com/forum/flat_read.php?th...

old spikes/superspikes/benefit (all in seconds)
55 54.7 0.3
56 55.6 0.4
57 56.6 0.4
58 57.6 0.4
59 58.5 0.5
60 59.5 0.5
61 60.4 0.6
62 61.4 0.6
63 62.3 0.7
64 63.2 0.8
65 64.2 0.8
66 65.1 0.9
67 66.0 1.0
68 66.9 1.1
69 67.8 1.2
70 68.7 1.3 (underestimate? Vaporflys mixed into 10000 data 2016+)
71 69.5 1.5 (Vaporflys in data)
72 70.4 1.6 (Vaporflys in data)

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.

zzzz wrote:

The Outside article says:

I asked Geoff Burns, a physiologist with the U.S. Olympic and Paralympic
Committee and co-author of Joubert’s study, to run some numbers for me. The
calculations are trickier at track-running speeds than they are for marathoners
(for reasons explained here), but a reasonable estimate is that improving your
running economy by 1 percent will increase your racing speed by about
two-thirds of a percent. So a typical 4:00.0 miler in the old spikes might
expect to run 3:56.8 in the new ones—“corrected for inflation,” as Burns puts
it.

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.

Thread:www.letsrun.com/forum/flat_read.php?th...

old spikes/superspikes/benefit (all in seconds)
55 54.7 0.3
56 55.6 0.4
57 56.6 0.4
58 57.6 0.4
59 58.5 0.5
60 59.5 0.5
61 60.4 0.6
62 61.4 0.6
63 62.3 0.7
64 63.2 0.8
65 64.2 0.8
66 65.1 0.9
67 66.0 1.0
68 66.9 1.1
69 67.8 1.2
70 68.7 1.3 (underestimate? Vaporflys mixed into 10000 data 2016+)
71 69.5 1.5 (Vaporflys in data)
72 70.4 1.6 (Vaporflys in data)

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.

zzzz wrote:

The Outside article says:

I asked Geoff Burns, a physiologist with the U.S. Olympic and Paralympic
Committee and co-author of Joubert’s study, to run some numbers for me. The
calculations are trickier at track-running speeds than they are for marathoners
(for reasons explained here), but a reasonable estimate is that improving your
running economy by 1 percent will increase your racing speed by about
two-thirds of a percent. So a typical 4:00.0 miler in the old spikes might
expect to run 3:56.8 in the new ones—“corrected for inflation,” as Burns puts
it.

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.

Thread:www.letsrun.com/forum/flat_read.php?th...

old spikes/superspikes/benefit (all in seconds)
55 54.7 0.3
56 55.6 0.4
57 56.6 0.4
58 57.6 0.4
59 58.5 0.5
60 59.5 0.5
61 60.4 0.6
62 61.4 0.6
63 62.3 0.7
64 63.2 0.8
65 64.2 0.8
66 65.1 0.9
67 66.0 1.0
68 66.9 1.1
69 67.8 1.2
70 68.7 1.3 (underestimate? Vaporflys mixed into 10000 data 2016+)
71 69.5 1.5 (Vaporflys in data)
72 70.4 1.6 (Vaporflys in data)

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.

Cvjhvh wrote:

zzzz wrote:

The Outside article says:

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.

Thread:www.letsrun.com/forum/flat_read.php?th...

old spikes/superspikes/benefit (all in seconds)
55 54.7 0.3
56 55.6 0.4
57 56.6 0.4
58 57.6 0.4
59 58.5 0.5
60 59.5 0.5
61 60.4 0.6
62 61.4 0.6
63 62.3 0.7
64 63.2 0.8
65 64.2 0.8
66 65.1 0.9
67 66.0 1.0
68 66.9 1.1
69 67.8 1.2
70 68.7 1.3 (underestimate? Vaporflys mixed into 10000 data 2016+)
71 69.5 1.5 (Vaporflys in data)
72 70.4 1.6 (Vaporflys in data)

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.

Another way to look at it. in 2019, 35 NCAA runners ran sub 4. In 2023, 35 NCAA runners ran 3:56.9 or faster.

15tg wrote:

zzzz wrote:

The Outside article says:

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.

Thread:www.letsrun.com/forum/flat_read.php?th...

old spikes/superspikes/benefit (all in seconds)
55 54.7 0.3
56 55.6 0.4
57 56.6 0.4
58 57.6 0.4
59 58.5 0.5
60 59.5 0.5
61 60.4 0.6
62 61.4 0.6
63 62.3 0.7
64 63.2 0.8
65 64.2 0.8
66 65.1 0.9
67 66.0 1.0
68 66.9 1.1
69 67.8 1.2
70 68.7 1.3 (underestimate? Vaporflys mixed into 10000 data 2016+)
71 69.5 1.5 (Vaporflys in data)
72 70.4 1.6 (Vaporflys in data)

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.

it if you compare the 10th & 15th best performer each year there is no change in the last 20 years.

king999 wrote:

You know what I never see in these analysis? Every sport has better athletes year over year, not allowed in distance running. Never allowed for..everyone is a drug cheat or has better shoes.

There are over 100M more Americans alone now than 1975.

Coaching is way better, way better years over years.

instant info thru Internet, guys/gals know what the marks are..how many run them, it gives folks a glimpse so much more quickly as to what is possible. You used to have to wait a month back in the day.

Pacing. most of it solid in almost every single race. I think this IS the biggest factor. Not that they never existed, but the sheer volume of them.If you are an Elite... elite HS kid, you can find at least 3 shots in multiple distances for set up races, without even disrupting your season.

If they made that much difference why isn't Ingebrightsen running..3:26 regularly by now. He rarely loses and has beaten everybody in multiple events at some point or other.

I read this piece yesterday, and found much of it soundly valid, but I never see the other factors even mentioned.

I think the shoes help..3.2 seconds at a mile.?.no way no how. I will never be convinced of that..doesn't mean I am not wrong LOL.