I'm sure this has been posted somewhere but I have not been able to find it. Has anyone put together a reliable 5k time distribution for the American population? For example, the number of people who can (currently, not pr) run a 5k under 20min, 19min,18min...
Just a wild guess but I'm thinking anything under 20 minutes would amount to only 1 or 2% of the population.
Why stop there? We can do it by age and gender. Heck, we could do it by weight of the runner.
If you look at big 5ks (>1000 entrants), 20 minutes is about the 10-11th percentile for men and about the 2nd percentile for women. Below that it's difficult to get good numbers because the turnout of faster runners is extremely biased and selective
Okay, here is a bit of a rambling response as I like math and running. Check out this link:
http://runningtimes.com/Article.aspx?ArticleID=6385&CategoryID=&PageNum=1
. I know this doesn't directly answer your question, but there is some cool information.
"A two-year study of more than four million high school students during the early 1980s found that the average boy took 7:40 to run a mile. The average girl took 9:51."
A 7:40 mile would correlate to 25:52.1 5k, although most people without much running background, or with backgrounds in other sports, or who are less fit would comparatively do better in the shorter events than the longer events. Granted, at least at my school, a lot of people just walked the mile run or didn't put in their best effort. But also, most people are considerably heavier and less active than they used to be, as this study was done 30 years ago.
For most of the races I run, a 21:xx minute 5k ( 7 minute pace) would be about the 95-99 percentile for a girl and probably the 70-90 percentile for a guy, among a self-selected, at least semi-trained population of people. I'm guessing most of the top runners on average run a lot more races than the slower runners. Granted also a lot of runners in my running club may run a race as a training run or run with another runner who is slower, but I'm assuming the majority of the runners in most races are putting in 90-100% effort.
Here is a distribution of times from a large 5k that I ran recently ( Somerville Jingle Bell Run in MA) : Temperatures were in the high 20s.
3640 women
Average - 32:34
top 10% - 26:02
top 5% - 24:33
top 1% - 21:54
top 1/10 of a percent (4th) - 18:45
Winning time: 18:30
1879 men
Average: 28:42
Top 20%: 23:37
Top 10%: 21:37
Top 5%: 20:10
Top 1%: 18:06
Winning time: 16:37
For some reason, clydesdales and athenas were listed as additional genders in the results so I did not account for them.
There were some walkers, but looked like 90% were mostly or completely running (11:xx min pace or faster.
It would be cool create a database that would compile results from certified 5ks within a given state by residents of the state for a given year and create a top 1000 list. Doesn't seem like it would actually be all that hard to add a feature like that to Athlinks or Coolrunning.
Try this link:
http://www.pace-calculator.com/5k-pace-comparison.php
Even though the sampling for this may be a bit off, it's still fun to play around with.
If you really want the time distribution for the population (and not the population of people who enter races) you are out of luck. You will need to make some huge assumptions.
Maybe for the slower times, but for the faster times (17:00 for men for example) its probably safe to say you have run a race in the past year. So if you compiled all races within a year you would have a pretty accurate estimate for the quicker times.
Rushed wrote:
Maybe for the slower times, but for the faster times (17:00 for men for example) its probably safe to say you have run a race in the past year. So if you compiled all races within a year you would have a pretty accurate estimate for the quicker times.
What? So you are saying take the total number of people who have run sub 18 in the last year and divide that by the US population?
Thats the idea. Obviously there are some problems with this method (there might be people capable of running that time, but dont). But it should give you a good idea of the distribution. It would be cool for someone to put together a database of 5k times across the county. A "performance list" so you could see where you stand up against, well everyone.
Ask and you shall receive. I think this is about right. Mind you, its for a man in his physical prime, apparently.http://www.letsrun.com/forum/flat_read.php?thread=34511275K Percentile Rarity (1 in ?? Men25:00 50 2 Average male in prime21:37 84 620:10 90.9 1119:30 97.22 3619:00 97.72 4418:40 97.959 49 Will often win age-group for 50 yr16:50 99.865 74116:25 99.9407 1689 Will win most local races15:45 99.9666 3,00015:08 99.9966 30,00014:25 99.9998 500,000 Top college runners13:50 99.999966 3,000,000 Top 50 in USSub 13 Top 15 in world 2009
Rushed wrote:
I\'m sure this has been posted somewhere but I have not been able to find it. Has anyone put together a reliable 5k time distribution for the American population? For example, the number of people who can (currently, not pr) run a 5k under 20min, 19min,18min...
I think the distribution above is likely confounded by the fact that much of the US population is unfit and that many talented runners are never discovered. Here is an estimated distribution for a very fit population of men aged 18 to 33 (extrapolated from a combination of Japan's best 10k times and from a study of the Ache people in South America).
Mean 5k time for men aged 18 to 33---19:00
Two thirds come in between------------17:30 and 20:30
95% come in between------------------16:30 and 22:30
% WR speed----Top X%----1 in X men-----200m-----400m-----800m-----1 mi*----2 mi*----5 km-----10 km
63%-------------74%------------------------30.42-----68.45-----2:39------5:50-----12:29----20:00----41:39
65%-------------63%------------------------29.66-----66.74-----2:35------5:41-----12:10----19:30----40:37
66%-------------50%------------------------28.90-----65.03-----2:31------5:32-----11:52----19:00----39:34----average young man
68%-------------37%------------------------28.14-----63.32-----2:27------5:24-----11:33----18:30----38:32
70%-------------25%-------4---------------27.38-----61.60-----2:23------5:15-----11:14----18:00----37:29----top 5 in average Ache tribe
72%-------------14%-------7---------------26.62-----59.89-----2:19------5:06-----10:55----17:30----36:27
74%-------------7%--------14--------------25.86-----58.18-----2:15------4:57-----10:37----17:00----35:24----fastest in average Ache tribe
76%-------------3%--------33--------------25.10-----56.47-----2:11------4:49-----10:18----16:30----34:22
79%-------------1%--------100-------------24.34-----54.76-----2:07------4:40-----9:59-----16:00----33:19
81%-------------------------453-------------23.58-----53.05----2:03------4:31------9:40----15:30-----32:17
84%-------------------------2600-----------22.82-----51.34-----1:59------4:22------9:22----15:00----31:14
87%-------------------------22,000---------22.05-----49.63-----1:55------4:14------9:03----14:30----30:12
90%-------------------------300,000--------21.29-----47.91-----1:51------4:05------8:44----14:00----29:09
93%-------------------------7 million---------20.53----46.20-----1:47------3:56------8:25-----13:30----28:07----olympic finals
97%-------------------------342 million------19.77----44.49-----1:43------3:47------8:07-----13:00----27:04----gold medal
100%-----------------------------------------19.19----43.18-----1:41------3:41------7:53-----12:37----26:18----world record
*Converted from 1500m and 3000m times
This spread is probably what you would expect if everybody realized his full running potential. Looking at the 'average' guy's times, they look just somewhat slower than what you would expect from a DIII college soccer player.
One way to define a 'serious' runner might be such that you have enough talent that your VO2max (or your sprint speed) could be useful to your tribe if you were a hunter gatherer. This should put you roughly in the top 3 to 5 young men in your tribe, so about at 70% of world record pace (18:00 for the 5k).
Persistence hunting with the top 3 men in a Khoi-San tribe:
http://www.youtube.com/watch?v=826HMLoiE_o
Study of the Ache:
http://anthropology.missouri.edu/people/walkerpubs/phys12.pdf
Source of Japanese 10k times:
take with a grain of salt wrote:
Try this link:
http://www.pace-calculator.com/5k-pace-comparison.phpEven though the sampling for this may be a bit off, it's still fun to play around with.
Not sure I would place much reliability on this. Playing with it, I input my age group 60-64 Male. And only got 1 runner running faster than a 7:20 pace. In a recent local race, probably about 3-400 max, we had 4 males in that age group run 7:02 or faster and the fastest guy in town was not even in the race, and I know of a couple of other local guys that can run that pace.
As someone who likes math and running your math is a joke. You said a 7:40 mile would correlate to 25:52.1 5K, this is awful how wrong you are. A 5K is 3.1 miles, so let's do the math. if you are running a 7:40 mile, that means you would do 3 miles in 23 minutes. So by your math, the .1 of the race is going to take you 2:52 seconds to complete? Absolute stupidity. At a pace of 7:40 mile, it is going to take you on average, 46 seconds to do .1 of a mile so that would put your 5K pace at 23:46 for a 5K, not as you stated 25:52.1.
I ran a 5K at a pace of 7:56 a mile and completed it in 24:34. So it took me 1:18 less than your estimated time and I had a 16 second longer pace.
Good thing you like math :)
Disappointed wrote:
As someone who likes math and running your math is a joke. You said a 7:40 mile would correlate to 25:52.1 5K, this is awful how wrong you are. A 5K is 3.1 miles, so let's do the math. if you are running a 7:40 mile, that means you would do 3 miles in 23 minutes. So by your math, the .1 of the race is going to take you 2:52 seconds to complete? Absolute stupidity. At a pace of 7:40 mile, it is going to take you on average, 46 seconds to do .1 of a mile so that would put your 5K pace at 23:46 for a 5K, not as you stated 25:52.1.
I ran a 5K at a pace of 7:56 a mile and completed it in 24:34. So it took me 1:18 less than your estimated time and I had a 16 second longer pace.
Good thing you like math :)
*Facepalm*
He said that if you could run a 7:40 1 mile run, then you are in good enough shape to run a 25:52 5k. I assume your 5k PR is at a slower pace than your 1mi PR.
V1per41 wrote:
Disappointed wrote:As someone who likes math and running your math is a joke. You said a 7:40 mile would correlate to 25:52.1 5K, this is awful how wrong you are. A 5K is 3.1 miles, so let's do the math. if you are running a 7:40 mile, that means you would do 3 miles in 23 minutes. So by your math, the .1 of the race is going to take you 2:52 seconds to complete? Absolute stupidity. At a pace of 7:40 mile, it is going to take you on average, 46 seconds to do .1 of a mile so that would put your 5K pace at 23:46 for a 5K, not as you stated 25:52.1.
I ran a 5K at a pace of 7:56 a mile and completed it in 24:34. So it took me 1:18 less than your estimated time and I had a 16 second longer pace.
Good thing you like math :)
*Facepalm*
He said that if you could run a 7:40 1 mile run, then you are in good enough shape to run a 25:52 5k. I assume your 5k PR is at a slower pace than your 1mi PR.
Yup disappointed in Disappointed.
HuffnPuffn wrote:
Unless you are over 60 you shouldn't even mention the fact you "ran" a 5K in over 24 minutes on a running forum. We are all laughing at you.
Elitism and arrogance are ruining this forum. You are a bully and for that I couldn't care less whether you run a 14 minute 5k.
HuffnPuffn wrote:
Unless you are over 60 you shouldn't even mention the fact you "ran" a 5K in over 24 minutes on a running forum. We are all laughing at you.
Elitism and arrogance are ruining this forum. You are a bully and for that I couldn't care less whether you run a 14 minute 5k.
Many people have responded with the parametrization independent quantiles you wanted. I was looking for evidence that race finish times are normally distributed, as part of the justification for defining "average" speed as the distance weighted harmonic mean. Hope the following helps:
Finish times are (approximately) normally distributed, see
http://www.rrtc.net/finishline/theory.html
That should help build the cumulative distribution function and the quantiles.
(This is meant to be part of a mini-lecture on how to meaningfully define "average" and variance of a random variable: first figure out what function of the random variable is (if at all) normally distributed, calculate _its_ mean and variance and interpret those. The distribution of speeds will be highly skewed and not at all normal.)