Kelly Jaske story

She's running like 120 mpw. THAT is what is amazing.. she has obviously been working VERY hard to run the times that she is. Similar to a few men out there, that have already been mentioned in this thread.

So running for a few years and getting up to 120 mpw + natural talent, pretty good.

And by your argument, I guess we should be embarrassed about US men's distance running because Andy Wheating only took up the sport 2-3 years ago, and hey look, he qualified for the USA Olympic team.

not so far buddy wrote:

Bud, again, the IAAF tables use a linear comparison, which you CAN NOT use when comparing a normal population (men) with a non-normal population (women). Mercier appropriately accounts for this by using a rank-order method. IAAF acknowledges biological differences exists between the populations (whatever those differences may be, between or within). Even if depth is improving on the womens side, it doesn't appear to be improving much for the top 200:

I'm a little confused here. You say that it is a non-normal population of women, meaning it is not as deep (regardless of reason) as the mens population? Yet a rank order is supposed to fix this?

I have not seen either algorithm, nor studied it in depth, but to me (a lowly math/stat B.A.) it seems that the lack of depth in the womens field would, in fact, provide for an unusually fast conversion in favor of the woman. If what AN is stating is true, the part about Haile G's WR is equivalent to the top 10 performer of the marathon, then it seems this method is not correct. IF I were the one who created this method, found this result, I would go back and discover where I made that mistake. Just me.

Questions about the algorithm: does it follow the same method for all events? 100m to Marathon and field events? If so, the fact that it only takes into account the past 4 years does not show enough of the story. Womens marathoning is relatively young (along with all womens sports for that matter) compared to men's marathoning. This would help 'explain' the depth issue in the performances by showing the changes in womens times with the mens. I would love to see the algorithm to really understand it a bit better. If somebody could provide it (if available for public viewing) I would appreciate it.

She passed me at Cherry Blossom 10 miler last year. I think she ran 59 minutes and some change. I didn't even know who she was back then until I saw her picture on the lets run front page.

Here is the algorithm for the Mercier calculator (this is the algorithm for the 1999 calculator, but I believe that the algorithm was the same for the 2005 calculator, with the only difference being that data from a different four-year period were used).

http://myweb.lmu.edu/jmureika/track/Mercier/Merc99.html

As you can see, there are a number of unexplained mathematical assumptions in the calculator that I didn't mention earlier, notably these three:

1. "The next step is to assign to each weighted speed (and hence each ranking) an associated score. For the 10 performance (5 men, 5 women), these are

Rank Men Women

5th 965.8 694.2

10th 947.3 673.9

20th 931.8 657.7

50th 910.0 625.3

100th 889.7 597.2"

2. "The 11 (Vwt, Point) pairs are the subjected to a linear fit-- that is, we find the best straight-line equations

Points = A x Vwt + B

Vwt = C x Points + D

which describes the data."

3. "A 'Women's Only' scoring table is obtained by a linear rescaling of the base tables: the associated scores/performances from the men's tables are adjusted by the following equation:

Swomen = ( Smen + 370.23683 ) / 1.10218405

Here, Swomen is the adjusted women's score, and Smen the original men's."

Keep in mind that these same three assumptions are apparently used for every distance from 100 meters to the marathon. It is unclear why these arithmetic adjustments to the data would produce anything close to a useful comparison of men's and women's performances, so -- as you have correctly observed -- if you apply these arithmetic assumptions and end up with bizarre results (for example, that the top dozen marathoners are all women, or that 29 of the top 30 marathoners are women), the most sensible conclusion is that the problem is in the arithmetic assumptions of the model, and not in the training and racing of all of the male marathoners in the world.

A couple more comments:

I'm not sure why "not so far buddy" states that the male population is "normal" and the female population is "non-normal." I assume that what is meant is that the distribution of times for men in various events can be plotted on a normal or Gaussian distribution curve and the women's times cannot. I don't know of any empirical support for that proposition. (Obviously, the mean points for the respective curves will be different, and the standard deviations may be as well.) In any event, I'm not sure that the existence or absence of Gaussian characteristics in the distribution curves has much practical relevance to comparisons between performances of reasonably competitive runners, especially since the data points used in the calculator don't go beyond the top 100 performers in any given year.

I'm also not sure why "not so far buddy" states that the IAAF tables "use a linear comparison." I assume that the contention is that the IAAF tables simply multiply the men's times by a constant to arrive at equivalent times for women. I don't know where "not so far buddy" got that information. (As to whether such a conversion method would be reasonable, I'm not taking any position one way or the other.)

We talked about this on a previous thread. The women's population is more positively skewed to the right, signaling something more going on (biologically?) than a "lack of depth". Even if depth is improving (ex. there were slightly more women Olympic Trials qualifiers in 2008, with a tougher standard of 2:47), it doesn't appear to be improving much at the top (with the spread of the annual top 200 women being at least 12 1/2 min. the past 8 years, vs. at least 6 1/2 min. for the annual top 200 men). If anything, even with improving depth, the women's population of elites/subelites is staying positively skewed. With few women at the top, such a large spread amongst them, and a general population of elites/subelites that is positively skewed, it makes sense that the very top women (top 15 women all-time, excluding repeat performances) appear as anomalies, compared to the more homogenous men's population. Not that Paula's time doesn't count, but it certainly is more of an outlier (approaching the women's physical limits) compared to Geb's time, being 3 min. off the predicted men's limit (as shown by the stats of the Stanford Biologist, along with the Mercier calculator).

The IAAF tables state they were statistically developed and points assigned (based on 3 years of world rankings, 2005-2007) to compare performances "within" each gender, rather than "between" the populations. As such, your assumption that the points are equivalent between the genders may not hold true. For example, if you use the IAAF tables to compare the humans limits found by the Stanford Biologist (which is comparable to Mercier), 2:14:59 for a woman equals 1294 points, whereas 2:00:47 for a man equals 1344 points. So that's saying 1294 points for a woman = 1344 for a man.

If you're unclear on the Mercier algorithm, you can always try to message the author, Daniel Mercier:

I applaud them for attempting to explain their model, while the IAAF tables lack this transparency.

1. What evidence do you have that the performances of male marathoners can be approximated by a normal or Gaussian distribution, and that the population of female marathoners cannot? You have offered no evidence to support your assumption. You assert that the women's distribution is "more positively skewed to the right," but unless you are merely making the trivial observation that the mean for the women's distribution curve is further "to the right" -- that is, slower -- you haven't provided any empirical support for your assertion. Even if there is a larger "spread" among women's performances at the top end, that doesn't mean that the distribution of female performances is less accurately approximated by a normal or Gaussian distribution curve (as you should know, having "taught the course"). Moreover, you still haven't explained why such a difference in the distribution of men's and women's marathon times would make the Mercier calculator any more reliable, or why it would make the IAAF tables less reliable. Indeed, there is nothing in the Mercier calculator that suggests that it would account for any asymmetries or other deviations from a Gaussian distribution.

2. You have absolutely no evidence that Paula R.'s performances are closer to female physiological limits than Geb's performances are to male physiological limits. I don't know who "the Stanford Biologist" is, but I can say with great confidence that anyone who claims to know the physiological limits of male and female marathoners isn't much of a scientist. I've seen many goofball Ph.Ds and M.D.s make such assertions over the decades. Sixty years ago, an earlier generation of these goofballs were telling the world why no man would ever break four minutes in the mile and why no woman would ever be able to run a marathon. Thirty years ago, I remember listening to goofball Ph.Ds and M.D.s explain why no one would ever jump farther than Bob Beamon did, why Lasse Viren could only have accomplished what he did in the '76 Olympics by placing himself in a state of ketosis, and all kinds of other nonsense. Able scientists don't do that sort of stuff -- unless they have an incentive, like money or publicity, for saying highly unscientific things.

3. Why would I want to talk to Mercier or anyone else involved in that project about how they came up with their goofy little calculator? People come up with goofy formulas all the time. You "applaud" the people behind the Mercier calculator for "attempting to explain their model," but they actually don't attempt to "explain" their model at all. Rather, they simply reveal some of the arithmetic calculations that they used to come up with scores. They don't explain why they think that those scores bear any resemblance to the real world. Until these people give a reasonable explanation for why they think that the top dozen marathoners in human history have all been women, and why 29 out of the top 30 marathoners in human history have all been women, it's reasonable for the rest of us to assume that they have simply devised a patently absurd model for measuring athletic performance.

4. Your comments about the IAAF tables indicate that you still don't understand them, but at least you're now admitting that you don't. Previously, you were asserting that the tables simply applied some sort of "linear" comparison between men's and women's performances; now, you're criticizing the tables because they lack the "transparency" of the arithmetic calculations of the Mercier model. What is clear from the IAAF's explanation of its tables is that they are based on actual performances, that they are updated and refined every year, and that they do not use a single goofy arithmetic formula for all athletic events, from the women's javelin to the men's 50K walk.

5. Finally -- and probably most notably -- you still haven't said why you regard the Mercier calculator as the "only valid" comparison of men's and women's performances. Do you have the slightest idea of the reasoning behind the arithmetic calculations and assumptions of the Mercier model? Do you actually defend the results obtained by applying the calculator? Do you really believe that the Mercier calculator provides accurate comparisons between, for example, performances in the men's high jump and the women's marathon?

Daburt wrote:

Or is the underlying story that anyone can run, because it takes no skill only natural talent? Sort of makes all your hard work a waste of time...hmmmm?

You're absolutely right, anyone can run. Only 90% can run fast, however.

Avocados Number: Moreover, you still haven't explained why such a difference in the distribution of men's and women's marathon times would make the Mercier calculator any more reliable, or why it would make the IAAF tables less reliable....

Are you kidding? I have said why many times, on this thread and in exact response to you on that other thread. No more spoon feeding, if you're not going to take the time to read my posts or the links provided. It appears "you think" because you have two degrees from MIT that you are somehow above the efforts of others. Objectivity will always trump what "you think".

not so far buddy wrote:

Again, Mercier is the only valid tool to compare men and women times.

The Mercier calculator is crap. As an experiment I just entered my steeple PR and it told me that I am equivalent to a man running: 8:19 for the steeple, 2:08 for the marathon, 1:01 for the 1/2 marathon. Let me just tell you that there is no way in hell that my steeple PR is equivalent to those times. The Iaaf calculator gives me: 8:40 for the steeple, 2:14 for the marathon, 1:04 for the 1/2 marathon. That is much more reasonable.

Yep she finished 18th, 59:45

http://www.cherryblossom.org/results/2008/women.htm

who stole steeplechick wrote:

The Mercier calculator is crap. As an experiment I just entered my steeple PR and it told me that I am equivalent to a man running: 8:19 for the steeple, 2:08 for the marathon, 1:01 for the 1/2 marathon. Let me just tell you that there is no way in hell that my steeple PR is equivalent to those times. The Iaaf calculator gives me: 8:40 for the steeple, 2:14 for the marathon, 1:04 for the 1/2 marathon. That is much more reasonable.

The steeple is such a relatively new event that it couldn't be expected for there to be enough data using 2001-2004 world rankings. While the IAAF doesn't recommend using their calculator to compare genders (their tables weren't statistically developed for this), if you use their 2008 tables, your PR of 9:57 equates to 8:35 for a man. It could be assumed that when the Mercier calculator is updated again, it will provide a more reasonable, and valid, comparison.

It could be assumed that when the Mercier calculator is updated again, it will provide a more reasonable, and valid, comparison.[/quote]

So you at least acknowledge that (i) in its current form and (ii) for the steeplechase, the calculator is invalid or "crap" (as steeplechick put it)?

I have to disagree with point 1, my old friend. A positively skewed distribution, by definition, means that there is a larger spread among performances at the top (or fast) end. It also means that the assumptions of normality (mean, median, and mode are equal, equality of variance to the right (faster) and left (slower) in terms of distance on the x axis) have not been met. It does NOT mean that the mean for the curve is farther to the right, and, if it did, the "right" would mean faster, not slower, times. So your below statement is false, and this is not a point that can be argued--it is statistical fact, and is taught in an intro to statistics course:

"Even if there is a larger "spread" among women's performances at the top end, that doesn't mean that the distribution of female performances is less accurately approximated by a normal or Gaussian distribution curve."

Now I'm not stating that the men's distribution does follow a normal curve, but it is most certainly closer, as there is less variation among the top times for the men than there is for the women, which is why the very top times for the women are not well predicted by the Mercier calculator. This point could be proven using any statistical program--one could enter in every marathon time ever run by a man in one data set, and every time ever run by a woman in another data set, and then test for assumptions of normality. The picture is complicated by the fact that our current discussion has only to do with the top 1-2% of each distribution, so we are dealing with only the tail end, and violating assumptions of normality out in these regions has a HUGE affect on attempts to apply calculations to equate the genders.

A final point: the current discussion originated from someone's doubt about Kelly Jaske's "equivalent" men's half-marathon time. Because Jaske's time is quite far away from the top times ever recorded for the women's half-marathon, and is instead nestled in a slew of other similar times, it is likely to be better predicted by the Mercier calculator, due to the fact that it will be less affected by the "problem" of the positively skewed distribution.

Karl Popper wrote:

It could be assumed that when the Mercier calculator is updated again, it will provide a more reasonable, and valid, comparison.

So you at least acknowledge that (i) in its current form and (ii) for the steeplechase, the calculator is invalid or "crap" (as steeplechick put it)?[/quote]

The Mercier statistical model is valid because it appropriately uses rank-order, non-parametric statistics. It's the data set that was used that provides an inaccurate comparison, as there wasn't near as much data on the women's side as there is now. If "who stole steeplechick" ran her time 11 years ago, she would have been the world record holder. If she ran her time in 2004, she would had had the 44th best time for the year. If she ran her time last year, that ranked her 273rd best in 2008.

The IAAF tables are completely invalid for comparing the genders, as they weren't statistically developed to compare a normal with a non-normal population.

I think you may have misunderstood the context of the discussion about the "spread" between times at the top end. We were talking about the spread between top women's times when compared with the spread between top men's times, not the "spread" between times at the top end of the women's curve when compared with the spread between times at the slow end of the women's curve. You can't tell whether one distribution is more closely approximated by a Gaussian distribution curve than the other without knowing, among other things, what the distributions are on the other end -- that is, the "slow" end. If the "spread" between times at the slow end of the women's distribution curve is wider than the "spread" between times at the "slow" end of the men's distribution curve, then both distributions may be equally approximated by a Gaussian distribution, although the standard deviations may be different.

As for which direction is "right" or "left," it depends on what the parameter is being used for the x-axis. Is it speed, or is it time, or is it something else?

Of course, all of this talk about normal and Gaussian distributions is a bit of a diversion from the main point, which "not so fast buddy" continually dances around -- the Mercier calculator is a silly little toy that produces nothing remotely close to a reasonable assessment of athletic performances. Nobody on this thread has yet come up with any cogent defense of the Mercier calculator.

This thread cracks me up. It has been a week and you clowns are still arguing over the rules. You remind me of the person that has to explain the rules for 20 minutes for a 15 minute game.

Kelly ran a great race and beat several Olympians in the process. I have my beliefs on how good that was and I have a stronger belief that there is a huge upswing in american women. But that is unimportant and is not intended to take away from the accomplishment of any male athlete.

I did not misunderstand the context. I am simply stating this: if a distribution curve is positively skewed, then it is not normal. And I think we would agree that the men's distribution more closely approximates normal than does the women's distribution, although perhaps neither of them meets requirements for normality necessary for the application of parametric statistics.

STOP IT, AND TALK ABOUT JASKE.

not so far buddy wrote:

It appears "you think" because you have two degrees from MIT that you are somehow above the efforts of others.

Wow, that's a pretty big chip sitting on your shoulder.

You're still dodging the question about why you have asserted that the Mercier calculator is the "only valid" tool for comparing men's and women's times. I haven't heard anyone -- not even you -- defend the bizarre results obtained by application of the Mercier calculator. And you don't appear to have the slightest idea about how the Mercier calculator arrived at its various mathematical assumptions. So if the results suck, and the mathematical underpinnings of the model are incoherent, I'm wondering why you continue to defend this silly little toy.

Usually, when relatively intelligent people defend really stupid positions, it's because they have some particular interest in doing so. If you have no connection with the Mercier calculator, why are you defending such a silly model? Does it provide results that make you feel better about your own athletic performances? Are you just being contrary-minded? Did you simply take the creators' self-serving assertion of the validity of their model at face value, and now find it difficult to back away from that position for fear of losing face on an anonymous message board? I generally try (not always successfully) to avoid attacking the motives of those who take positions contrary to my own, because (among other things) I think that it's more productive and less hurtful to everyone if we focus on the merits of the various positions and arguments, but your continued defense of this goofy little model, without any understanding of its mathematical assumptions and without any defense of the ridiculous results obtained by its application, suggest that you are not particularly focused on the merits of the model.