Sub 5 minute Mile or Sub 3 hour Marathon. Which is a better achievement?

Stark

The GOAL mile was the only mile race I ever did. (Ended up running less than my 5k pace as the track had turned into an ice rink ).

Gringo78

BeepBeep67 wrote: »

Said in my best Tony's Trials voice

19 days
2 races
1 Mile (Charlesland track Oct 12th)
1 Marathon (DCM Oct 31st)

Let's put this debate to bed

Well, if you can only choose one to achieve in October, which do you want it to be?

zico10

Mellor wrote: »

2 was a typo. It should of been 3 hours.
luckily this acks up my point, maintaining it for far longer means pace isn't an issue.


A 3 hr marathon is very good.

A 15 second 100m is childs play.



I honest think you are trollign now and you failure to comprehend a simple point is incredible.

To reiterate:
Of course pace is needed hit the two targets. And of course the 5 min mile needs a quicker pace. But that alone can't be a measure for comparing. It's impossible to quantity pace and distance at the one go in absolute terms.

You simply siad, 5min mile is a quicker pace, therefore its better.
Which is nonsense and ignore the distance.

I agree that 5min mile is a better achievement, I'm disagreeign with you stupid "proof"

Well you're the mod, you should know. I'm done talking to you.

robinph

Stark wrote: »

The GOAL mile was the only mile race I ever did. (Ended up running less than my 5k pace as the track had turned into an ice rink ).

Goal mile, that's what I was thinking of.

BeepBeep67

Gringo78 wrote: »

Well, if you can only choose one to achieve in October, which do you want it to be?

`

no reason not to do both

RayCun

oldrunner wrote: »

Some of this debate reminds me of the put down by the taxi driver in the Stewart Lee story "sure you can prove anything with facts" (Homer Simpson might have said it as well) http://www.youtube.com/watch?v=4n-UGQcG3Jw

Equally, just because you write a nice equation and get numbers out the other side doesn't mean you are describing reality.

look, a formula!

Mellor

zico10 wrote: »

Well you're the mod, you should know. I'm done talking to you.

LOL, i'm not a mod here.
But i'm glad you've decide to stop talking nonsense.

Thread will be unsubscribed.

oldrunner

RayCun wrote: »

Equally, just because you write a nice equation and get numbers out the other side doesn't mean you are describing reality.

look, a formula!

That bloody science mumbo jumbo.

You are comparing a completely random made up formula designed to be a bit of fun with statistical analysis based on hundreds of thousands of actual observed data points over many years.


At this point, I say goodbye.

RayCun

oldrunner wrote: »

That bloody science mumbo jumbo.

You are comparing a completely random made up formula designed to be a bit of fun with statistical analysis based on hundreds of thousands of actual observed data points over many years.

But it isn't actually a statistical analysis, is it? You can feed as many data points into a computer as you like, and it won't produce a formula.
HP = a(b-t)^2 + c
does not magically appear from the data.
Even if we agree that the rough shape of the curve is appropriate for grading results, the values for a, b, and c have to come from somewhere.

In the end (as far as I can see from what I've found online, and no-one has said anything on the thread to contradict this) it boils down to Dr Spiriev putting different values into the formula until he arrived at some figures he liked. There's nothing shocking or wrong about this - you could say with some reason that if Dr Spiriev with all his years of watching athletics thinks 14.5 in the men's 100 metres is exactly as good as 33.9 in the women's 300 metres then he should bloody well know. If thirtyfoot agrees with him, all the better!

What I find funny is the argument? attitude? belief? that these tables are anything more than that. That they represent some sort of self-evident objective truth, that saying 14.49 in the men's 100 metres is better than 33.9 in the women's 300 metres is just as obviously true as saying that 14.49 in the men's 100 metres is better than 14.5 in the men's 100 metres.

But really, the value of the constants, and therefore the scores in the tables, are just an opinion. An educated and informed opinion, to be sure, but still just an opinion.

Pisco Sour

RayCun wrote: »

But it isn't actually a statistical analysis, is it? You can feed as many data points into a computer as you like, and it won't produce a formula.
HP = a(b-t)^2 + c
does not magically appear from the data.
Even if we agree that the rough shape of the curve is appropriate for grading results, the values for a, b, and c have to come from somewhere.

In the end (as far as I can see from what I've found online, and no-one has said anything on the thread to contradict this) it boils down to Dr Spiriev putting different values into the formula until he arrived at some figures he liked. There's nothing shocking or wrong about this - you could say with some reason that if Dr Spiriev with all his years of watching athletics thinks 14.5 in the men's 100 metres is exactly as good as 33.9 in the women's 300 metres then he should bloody well know. If thirtyfoot agrees with him, all the better!

What I find funny is the argument? attitude? belief? that these tables are anything more than that. That they represent some sort of self-evident objective truth, that saying 14.49 in the men's 100 metres is better than 33.9 in the women's 300 metres is just as obviously true as saying that 14.49 in the men's 100 metres is better than 14.5 in the men's 100 metres.

But really, the value of the constants, and therefore the scores in the tables, are just an opinion. An educated and informed opinion, to be sure, but still just an opinion.

I honestly think if you actually watched the sport you wouldnt be even bothering to argue this. Anybody who follows the sport knows that these comparisons across events are fair. Not 100% perfect, but fair.

And these tables make perfectly clear on the first page that men's and women's events cant be compared, hence the different tables for men and for women.

RayCun

04072511 wrote: »

I honestly think if you actually watched the sport you wouldnt be even bothering to argue this. Anybody who follows the sport knows that the comparisons across events are fair. Not 100% perfect, but fair.

Fair, sure, but objectively true? Do you think they are anything more than an educated opinion?

(Me watching the sport doesn't come into it. Seriously. If I watch athletics for 30 years and form an opinion that agrees with the results of the tables, that is still just an opinion. If the tables are useful because they reflect the experience of experienced observers of athletics then sure, the agreement of another experienced observer is a good data point. But the claims for the validity (and exactitude!) of these tables on this thread have gone way beyond that.)

(It's also worth asking, if another experienced observer of athletics thinks, "hmm, those tables are a bit off, the points for the javelin are a bit low, the point for the hurdling events are a bit high", is there any way for that opinion to feed into the next revision of the tables? Is there any reason for this other experienced observer to even follow this up? It's not like the tables have any real world impact, is it?)

RayCun

04072511 wrote: »

And these tables make perfectly clear on the first page that men's and women's events cant be compared, hence the different tables for men and for women.

You can compare men's 100m to men's shotput, but not to women's 100m?
Why not?

Krusty_Clown

04072511 wrote: »

I honestly think if you actually watched the sport you wouldnt be even bothering to argue this. Anybody who follows the sport knows that these comparisons across events are fair. Not 100% perfect, but fair.

Ahh yes. The famous chewbacca defense. Re-worded as: If you do not agree with me, then you are wrong, because anybody who follows the sport agrees with me. I reckon most people who follow the sport couldn't give a crap about comparison tables. They are largely the domain of the sports bureaucrat (a necessary evil) and the athletics statistics jockeys.

oldrunner

RayCun wrote: »

But it isn't actually a statistical analysis, is it?

In the end (as far as I can see from what I've found online, and no-one has said anything on the thread to contradict this) it boils down to Dr Spiriev putting different values into the formula until he arrived at some figures he liked.

But really, the value of the constants, and therefore the scores in the tables, are just an opinion. An educated and informed opinion, to be sure, but still just an opinion.

I cannot leave this nonsense alone, as much as I don’t want to comment.

Here is some background into the theory of establishing inter-event comparisons (www.imnc.univ-paris7.fr/basile/TablesNSA.pdf):

“Competition, typically for scarcely available resources, is hardwired in all living beings. In human society the drive for competition has led to various ritualised activities the most prominent of which are sports. While recreation sports express the need for physical exercise and are essentially of ludic character, competition sports focus on the comparison of the individual (or a team) to others within a precise setting. The notion of comparison is even more deeply ingrained in sports with quantitative character, where the performance can be objectively assessed. The advantage of quantitative, codified, sports is that they allow universal comparisons: the performance of two athletes can be compared even though they have never competed against each other.
On the other hand it would be illusory to extend these comparisons to noncontemporary athletes: too many parameters change over time (material conditions, rules, scientific support etc.) to make diachronic comparisons meaningful. Still, in sports where continuity is ensured through a sustained activity, the existence of records and the monitoring of the performances is quite useful since it, almost always, involves short time steps, typically of the order of a few years.
While quantitative measurements allow comparisons of the performance of competitors within the same discipline, the situation becomes more complicated when different specialities are involved. Clearly such a comparison is meaningful. Everybody would agree that the performance of a world champion in some event is better than that of an inexperienced competitor in a different event. But there is a possibility to make this comparison more precise and to distinguish performances on which only specialists could a priori pronounce themselves. The need for this comparison goes back to the competition trait we have stressed at the beginning of this section. But there exists also another, more practical, necessity, that of combined events. Since the most ancient times, the quest for the best all-around athlete has led to the proposal of combined- or multi- events [1]. While the classification of the ancient greek pentathlon was based on ranks in the four events with the last, wrestling, event being the tiebreak, the modern approach was much more quantitative. With the introduction of decathlon tables [2], the performances in athletics were converted to a number of points (typically between 0 and 1000) allowing thus a fine scoring for combined events and greatly facilitating inter-events comparisons.
In this paper we shall examine the question of scoring from two points of view. The first focuses on the physical (dynamical, physiological) basis for scoring [3]. The second is inspired by the theories of D. Harder (his "apples to oranges" approach [4]) which furnishes a basis for inter-sport comparisons. We shall compare Harder's tables to the current IAAF ones [5] and establish their close parallel.
As pointed out in the introduction, there exists a psycho- and socio- logical need for scoring. But what is more important is the necessity of efficient and fair scoring for combined events [6]. In sports like athletics where scoring tables were introduced a century ago one can assert that a certain level of maturity has been attained.
How does one go about setting up scoring tables? Two things are needed. First, decide on the correspondence between performance and points within a given discipline and second on the correspondence of performances accross disciplines. The commonly used approach relies heavily on statistics [7] and, provided a large data basis exists, gives fairly reliable results. In what follows we shall address these two points, the first on the basis of physical arguments and the second with the help of the theory of D. Harder. Scoring tables are obviously not written in stone and have to be revised regularly (not too often, lest they destabilise the combined events community). In particular, whenever a major change, due to equipment evolution or rule change, intervenes, the scoring tables must rapidly adapt in order to reestablish a fair judgment.
D. Harder has developed his approach to scoring in order to compare athletic achievements in different sports. The key to this comparison is that you "… compare the number of athletes who reach any given level …" (proportional to the number of athletes competing in that sport, of course). The quantitative basis of Harder's method is the following. A mark of 100 points is attributed to some performance if a fraction of 0.5 of the population can realise a score equal or better than this. For 200 points only a fraction of 0.05 of the population can do better than this performance. The next 100 points, i.e. 300, correspond to a performance realised by just 0.005 of the population and so on up to 1000 points where only a fraction 5.10-10 of the population can realise the corresponding performance.
We shall not pursue our analysis of Harder's work here. It contains a monumental amount of data collection and statistical analysis combined with a solid knowledge of sports and an abundance of common sense. Moreover the correlation of Harder's table for athletic events with the IAAF table is perfect.”

Some comments:

1. This forum is littered with threads that relate to comparisons (including this one).
2. By definition the participants on this forum are competitive.
3. The IAAF tables are not ‘a matter of opinion” (unless, of course, you believe that the science of statistical analysis is completely bogus). For each individual event they are based on a statistical measurement of tens of thousands of data points. So, each event is measured and then placed on a scale with a range that is common to all other events. Having completed the individual rankings per event, each event can be compared by using this common scale. See above for explanation of the theory on how this is done. To understand the maths involved, please take some statistical courses.
4. inter-event scoring is essential for the decathlon and heptathlon. While similar methodology is used to arrive at the points for these events it is not the same because there are different base principles required - see http://www.oocities.org/mdetting/sports/decathlon-points-history.html for a discussion on the evolution of decathlon scoring systems.
5. IAAF scoring tables are in practical use – examples (1) for World Rankings (see http://www.irishrunner.com/iaafrank.html to understand how the rankings work, including their relationships to the IAAF tables) (2) to determine ‘athletes of the meet’ and to distribute prize money (see http://www.iol.co.za/sport/iaaf-statistician-confirms-asa-errors-1.617540?ot=inmsa.ArticlePrintPageLayout.ot for an example of a mistake in their use and http://www.athleticsireland.ie/content/?p=19646 for their use in the Cork Graded series.

In summary, you have ‘an opinion’ that the IAAF tables are based on opinions but your opinion is objectively and demonstrably wrong. In the words of Mick McCarthy "Opinions are like backsides - we've all got them but it's not always wise to air them in public"

RayCun

oldrunner wrote: »

Some comments:

1. This forum is littered with threads that relate to comparisons (including this one).
2. By definition the participants on this forum are competitive.
3. The IAAF tables are not ‘a matter of opinion” (unless, of course, you believe that the science of statistical analysis is completely bogus). For each individual event they are based on a statistical measurement of tens of thousands of data points. So, each event is measured and then placed on a scale with a range that is common to all other events. Having completed the individual rankings per event, each event can be compared by using this common scale. See above for explanation of the theory on how this is done. To understand the maths involved, please take some statistical courses.
4. inter-event scoring is essential for the decathlon and heptathlon. While similar methodology is used to arrive at the points for these events it is not the same because there are different base principles required - see http://www.oocities.org/mdetting/sports/decathlon-points-history.html for a discussion on the evolution of decathlon scoring systems.
5. IAAF scoring tables are in practical use – examples (1) for World Rankings (see http://www.irishrunner.com/iaafrank.html to understand how the rankings work, including their relationships to the IAAF tables) (2) to determine ‘athletes of the meet’ and to distribute prize money (see http://www.iol.co.za/sport/iaaf-statistician-confirms-asa-errors-1.617540?ot=inmsa.ArticlePrintPageLayout.ot for an example of a mistake in their use and http://www.athleticsireland.ie/content/?p=19646 for their use in the Cork Graded series.

1. Yes. Most often comparisons within a single event (the 'Best of...' thread is divided into different distance races), or sport (the 1,000 miles thread). The Swim/bike/Run thread contains three different sports, listed separately AFAIK. I could be wrong, but I think the VDot thread is the only one that compares race performance at different distances.
2. Not everyone, but most of us, sure.
3. Yes, the IAAF tables are based on the input of thousands of data points. But as I said above, you can feed as much data as you like into your computer, and it will not produce the formula HP = a(b-t)^2 + c
'Statistical analysis' does not tell you the values for a, b, and c in different events. It will tell you that if you choose a particular set of values, it will calculate that this performance has the same score as that performance. But the data doesn't tell you that those performances should have the same score. That is a subjective assessment.
Sure, that uses a lot of data. And of course, the people looking at the data are experienced observers of athletics. But if the tables are compiled by someone plugging different values of a, b, and c into the equation, until they get results that they think look right - and nobody has suggested another way in which those values were derived, just repeated the magic words 'statistical analysis' - then the tables are a matter of opinion.

The quote above describes Harder's approach, which is an interesting contrast. Because you really could say that - once you set the initial criteria, of awarding points based on the percentage of athletes who reach that standard - you are just plugging in numbers. You enter all the performances everywhere, and you locate one performance on the scale. But there is no tinkering with constants.
(Which is not to say that this makes Harder's approach the right one to use to compare performances, because the initial criteria are also open to question. Some events have high participation, some have a low participation. If you are in the top 10% of a low-participation event, is that really exactly as good as being in the top 10% of a high-participation event? It's a moot point.)

4. I mentioned inter-event scoring yesterday, because it is interestingly different. There really is a need for scoring tables, and there is no pretence that these scoring tables will magically appear if you enter enough data. The principles for creating scoring tables have to be discussed and agreed, and this is not value-neutral. (When you construct the scoring, there will be some point at which a competitor who is outstanding in some events and pretty good at the rest will beat the competitor who is very good at everything. Where should that be? How brilliant do you have to be to outscore the generalist? The answer to that question can't be found in the data.)
5. These tables are used to distribute prize money? That is shocking.
So, is there a process available for people to question the method of calculating the tables and request changes?

donothoponpop

RayCun wrote: »

But if the tables are compiled by someone plugging different values of a, b, and c into the equation, until they get results that they think look right - and nobody has suggested another way in which those values were derived, just repeated the magic words 'statistical analysis' - then the tables are a matter of opinion.

That's twice you've belittled the use of statistics. You're no more going to understand what those constants means, than you would e=mc^2, without context. And even then, there's going to loads of regression curves, lines of best fit, least squares estimators, confidence intervals... (I've been cramming that stuff all week;)). It's not something a lay person can understand.

The common theme in the tables is the best people in the world striving to go higher, faster, longer, and enough of them doing it for long enough for there to be masses of results. The fact that they all happen in the same stadium, would be enough for people to query which was the performance of the night? I reckon that's fair enough. The more I read about Spiriev and all-athletics.com, the more I appreciate his ranking results. If the tables are a matter of opinion, they're a matter of the most considered opinion around.

RayCun

donothoponpop wrote: »

That's twice you've belittled the use of statistics. You're no more going to understand what those constants means, than you would e=mc^2, without context.

But m and c are directly measurable values, they are independent of the equation.
If a, b, and c are derived from something else, I withdraw my objection
My problem is that I think the points score is generated by those constants, and the constants are derived from the points score.

donothoponpop wrote: »

The more I read about Spiriev and all-athletics.com, the more I appreciate his ranking results. If the tables are a matter of opinion, they're a matter of the most considered opinion around.

And I do appreciate that, honestly. As I said above, "if Dr Spiriev with all his years of watching athletics thinks 14.5 in the men's 100 metres is exactly as good as 33.9 in the women's 300 metres then he should bloody well know"
What I think happened is that Spiriev collected a lot of figures, thought long and hard about the best formula to use to calculate points, plugged in values, studied the outputs, tweaked the values, studied the outputs, tweaked the values again and again, so that he came up with a set of scores he was happy with. Which is not a trivial job, by any means.

But it does look like Spiriev was reasoning backwards from his conclusion. He started with the opinion that these performances were all roughly as good as each other (and so was this set, and this set) and found a formula and set of constants that would express that, rather than working out what the formula and constants must be (for some other reason) and taking whatever values fell out.

Which isn't the worst thing in the world, obviously. The argument that a sub-5 mile is better than a sub-3 marathon because some of the most experienced observers in the athletics world say so is pretty strong. It's just a little different in kind to because the tables say so.

oldrunner

RayCun wrote: »

What I think happened is that Spiriev collected a lot of figures, thought long and hard about the best formula to use to calculate points, plugged in values, studied the outputs, tweaked the values, studied the outputs, tweaked the values again and again, so that he came up with a set of scores he was happy with. Which is not a trivial job, by any means.

But it does look like Spiriev was reasoning backwards from his conclusion. He started with the opinion that these performances were all roughly as good as each other (and so was this set, and this set) and found a formula and set of constants that would express that, rather than working out what the formula and constants must be (for some other reason) and taking whatever values fell out.

See http://en.wikipedia.org/wiki/Linear_equation and http://en.wikipedia.org/wiki/Statistical_analysis for some introductory reading.
He didn't invent a formula - he used an appropriate linear equation - he then derived the constants from the observed actual data (times or distances). He then used the formula (including the constants) to calculate the full universe of points for the relevant range of times or distances. By using a common scale, you can compare perfomances across disciplines.

This is Statistics 101.

you have no interest in comparing performances - others do.

RayCun

oldrunner wrote: »

he then derived the constants from the observed actual data (times or distances).

Well, no, because time is already an element in the equation.
To put it another way, when you draw your graph for calculating points in the 100m, points are one axis and time is the other. So to define the shape of the line itself, you must be adding some additional information.

Harder used the proportion of athletes achieving each result as his variable. Perhaps Spiriev was doing something similar, but nobody has said yet just what that something was...

rom

I did a intervals today and ran one for 1.5k at 17km/h on a treadmill. A mile would be just .1 of a K longer which I could do no problem. Now as a 5 min mile is 19.2k/h for 1.6km . Ok it was a treadmill so no big deal but I won't consider myself only new to running. I am training for DCM and would be happy with sub 4 but maybe I am more suited to sprinting ?

Krusty_Clown

rom wrote: »

I did a intervals today and ran one for 1.5k at 17km/h on a treadmill. A mile would be just .1 of a K longer which I could do no problem. Now as a 5 min mile is 19.2k/h for 1.6km . Ok it was a treadmill so no big deal but I won't consider myself only new to running. I am training for DCM and would be happy with sub 4 but maybe I am more suited to sprinting ?

I'd try and replicate it on the road or a track, before you make any life-changing decisions about what direction your running career should take.

rom

Krusty_Clown wrote: »

I'd try and replicate it on the road or a track, before you make any life-changing decisions about what direction your running career should take.

Totally agree as that is why i said threadmill. Just a sign that my 5k time should be able to improve a lot if i work at it.

Krusty_Clown

rom wrote: »

Totally agree as that is why i said threadmill. Just a sign that my 5k time should be able to improve a lot if i work at it.

Seriously though, take it to a track. Perhaps you do have a natural bent towards shorter stuff. Who knows? Only one way to find out.
Here's a chart that proposes equivalent treadmill times versus road-time. You should have the tread-mill at a 2% incline to approximate road pace, because of the dearth of wind resistance.