New mess for masters: boogered Age-Graded Tables

Houston, we have a problem. The latest set of WMA Age-Graded Tables appears to have flaws. Gary Duncan has written me about these (see way below). And Rex Harvey, the WMA’s pointman on the tables, even admits problems with scoring the latest USATF Indoor Masters Multi-event Championships (also see below). On top of all this, Chuck Phillips has come up with his own Age-Graded Tables. Chuck is one of the pioneers of masters track. In the early 1980s, he helped produce the first versions of what became the Age-Graded Tables used worldwide to help age-groupers (and race officials) judge folks of different ages.


But he didn’t stop there. This week he wrote me about his current project — producing age-graded standards based on current age records. His new tables, he says, are more accurate than those used by World Masters Athletics. He’s not alone in thinking this. Alan Jones has his own theories. And Masters (and National T&F) Hall of Famer Bud Held has his own thoughts. Anyway, check out the Excel files at Chuck’s Web site: phillipstrackstds.tripod.com.
Chuck informed me of his latest work thusly:

The site covers the effect of aging for T&F events (track and LDR running, hurdle, racewalking and field events) for masters and youths ages 6 to 100, and includes performance standards and age-factors for those events. A narrative explanation of the standards and the effect of aging is included at “STDSTEXT” covering the origin and development of the standards and explains the purpose of the backup graphs and tables also included. The narrative explains how the standards apply to T&F masters and youths of all levels of ability and suggests ways they can be used by individuals and coaches, and how they can be used for award or recognition of noteworthy performance purposes.

When I asked Chuck how his tables differ from the WMA Age-Graded Tables, he replied:

Good question and I will answer it, but it is a long story so please bear with me. After I answer the specific question I will give you the background information you should be aware of in this regard and comment on the referenced blog.
My standards differ from the WMA Age-Graded Tables in that
(1) they are current, they are updated annually to include the latest available records and are not several years out of date as are the WMA standards (as correctly noted in the blog);
(2) my standards are based on the records of all masters ages unlike the WMA standards that are based on the 5-year age segment method, and that is an artificiality allowing weaker performances in some instances to be used in lieu of better records to determine the WMA standards;
(3) my standards for running/racewalking events (as explained in my standards in the narrative text item ”About the Standards”) are generated in a manner such that the resulting standards of neighboring events (in fact all events) are mathematically interrelated and are thus all generated to the same demanding criteria of performance level; the WMA standards are based on each event separately and in their case the standards of event “X” may be based on strong highly competitive records while event “Y” standards may be based on only the weaker records of a less competitive event, and thus their standards are flawed for comparative purposes;
(4) it takes me about one month (in November) to obtain the new and/or improved records and update my records data base and then refit standards as may be required, while updated WMA standards are bogged down a kind of international bureaucratic nightmare that requires several years (as correctly noted in the blog) for committee preparation of the initially proposed standards, going through several refits at intermediate committee levels where competing biases are hashed out to obtain the concurrence of all participating country representatives, and then finally receive the official authorization of the WMA organization; for those in the know the WMA standards are dead-on-arrival when published.
I am not trying to replace the WMA or anybody else’s standards with my own – but the sport is not being well served with the official standards now available to the masters T&F community. Therefore I have made my standards available on the web (http://phillipstrackstds.tripod.com) for anyone to use who wants to, at no cost, as my contribution to the sport and its participants.
Now for some background information so that you won’t be blindsided about my standards and you won’t think I’m holding out on you.
I was printing a booklet of running standards in the early 1980s which were seen by Rodney Charnock of the IAAF and was asked in January 1984 to help him prepare standards for field events so the IAAF could establish an “age graded scoring system for masters multi-events.” This endeavor was being followed by Al Sheahen, editor of NMN and Jim Weed of TAC.
Al contacted me and asked if I would help him establish a system of age-graded standards under the auspices of NMN. That effort bubbled along via correspondence and email and then Al asked me to participate in the effort to produce the NMN/WAVA Age-Graded Tables of 1989. I participated as an outside contributor and got an insiders view of how that process worked.
My standards were not incorporated as such and I became familiar with the shortcomings of the 5-year age segment best record method. But I hung on hoping to get NMN/WAVA to replace their 5-year method with my fit-the-records-for-every-age method but that did not happen. I participated again helping Al with the NMN/WAVA effort of the 1994 updated revision, again hoping against hope to get the old fashioned, flawed WAVA approach replaced by my kind of an improved approach – but to no avail.
A couple of years ago Linda Honikman at the USATF LDR Center asked if they could use my standards for their LDR award and recognition purposes. I said yes and provided the LDR standards to her, but the next year they were subsequently dropped as a result of the ongoing negotiations (as correctly noted in the blog) between the LDR and T&F/WMA factions of masters athletics.
I am no longer assisting any USATF, WMA, NMN or any other T&F or Masters Organization as explained in the disclaimer of my site – my standards are available on the web for anyone to use at no cost.
Thanks, Ken, for the opportunity to explain all this, and the sport does deserve something better than the WMA version of the Age-Graded Tables.

Me again:
As Rex Harvey – majordomo of the WMA tables — always reminds me: The WMA tables have only one official use: To score multi-events (like the decathlon, weight pentathlon, etc.) in major masters meets. But Hy-Tek software incorporates these tables and they’re used all the time in deciding who the best “age-graded runner” is at many road races.
The WMA tables also are the basis of Jess Brewer’s age-graded query form and a similar online lookup by England’s Howard Grubb. And M55 pentathlete Stefan Waltermann has applied the WMA tables to the indoor event.
So a competing set of tables may be interesting, but ultimately they muddy some already confused waters. Certainly, Chuck’s work should inform any future redo of the WMA tables. But that may be years away.
For now, we have bigger fish to fry.
At the moment, Rex Harvey and others are trying to figure out why different scores have been found for results of the recent USATF indoor multi-event championships, especially in Phil Raschker’s pole vault.
In mid-March, Rex wrote back:

Ken Stone was kind enough to point out where I could find Phil’s (Female age 60) Indoor Heptathlon performances. Here is my result:

60m 8.99 1162 points
LJ 4.29 930 points
SP 8.06 762 points
HJ 1.34 1119 points
Hur 10.28 1155 points
PV 2.60 1213 points
800m 3:17.29 652 points

Total points: 6993

That doesn’t seem to be a number that anyone else has come up with. Please double check my calculations by looking at world-masters-athletics.org, then laws and rules, then Appendix I for the Female Indoor Heptathlon Age Factors, and Appendix b for the IAAF scoring formula constants. Jeff (Watry), what Hy-Tek Meet Manager version were you using? If it is current, then it should score correctly. If not, then we will fix it.
Rex Harvey

And more recently, I got this note from Gary Duncan:

(Regarding) the 2006 Age Graded Tables news2006/agt2006.xls referenced from masterstrack.com/blog/archives/000785.html
Just curious why on the Data sheet the Age Factors for the 5 columns CU (F5kmRoad) through CY (F5MileRoad) are the same across each age row, and then for 4 more columns through column DC (F10Mile) for 21 and up.
This is not true for the men, nor is it consistent with the OC row or the lower age record(?) rows. Guess I could recalculate from those, or is there a corrected table somewhere?

Sigh.
I’m awaiting official responses on what needs to be fixed and by whom.

Print Friendly

March 29, 2007

10 Responses

  1. Mark Alexander - March 29, 2007

    Chuck’s point (3) is a good one: WMA 2006 factors for adjacent distances are sometimes uneven for competitive reasons.
    For example, at men’s track distances between the mile and 10000m inclusive, the “no age discount” window extends from age 23 to 27 inclusive; however, the “no age discount” window for the 1500m extends from age 23 all the way to age 33 inclusive — six more years. Surely there is more in common between performances in the 1500m/mile than between the mile/10000m.

  2. pino pilotto - March 29, 2007

    And what is it about the tables of Bernd Rehpenning, whitch was discussed on july 2006 in Poznan during the evacs 2006?
    Houston, there is many work to do…

  3. Mark Cleary - March 29, 2007

    The middle and long distance athletes have always gotten the shaft on these age graded tables.We rely heavily on age graded performance in determining our 5 year age group top athlete for our annual USATF awards–this year only one distance person on the Men’s side won a five year age group award–which shows this bias quite clearly.So too a large degree this is nothing new to age grading it’s the same old thing going on with the new tables–in my opinion the old tables were better than the new one’s–yes they need a lot more work to be equitable.

  4. Ken Stone - March 30, 2007

    One thing I didn’t mention in my blog post:
    National Masters News has been waiting for months and months to receive a final version of the Age-Graded Tables to publish as a hard-copy booklet.
    These recent revelations of problems in the tables may account for NMN not being sent the files.

  5. Jeff Watry - March 30, 2007

    Ken and all – wish it where a simple answer – First the laptop which I had used for the last four years – including scoring all the multi’s crashed – ever try and find the original software cd to reload Hytek and then go do all the updates – to make it more interesting is Hytek also came out with a different version I think in early January – the tables changed last year after the indoor season – there we even problems last outdoor with people not running the latest version and tables – to make Phil’s case more interesting is the 27″ versus 30″ – for a while there was a different factor for each height – all said and done if I had been running the latest version – on the laptop I had along we would have been fine – We did check that – Rex, the Hytek people and myself once we found differences – which tables or factors should be used is not my call – I just need to make sure I use what is specified – I was out of date – not Rex or WMA –

  6. Bud Held - March 30, 2007

    Age grading is a really complicated process. Records are always changing and the continued changing of weights and specifications in the throws adds to the problem. I took a quick look and Phillips numbers and while some seem pretty good, others are not so good.
    An example of Phillips inconsistency is in the W65 60 m and the W65 100 m. Phillips gives the 100 m (100%) mark as 13.85. That is a pretty good number, but then he gives the 60 m mark as 9.29. This is way off. A 13.85 100 m actually converts to an 8.90 60 m, The WMA numbers here are much more consistent, 14.19 for the 100 m and 9.25 for the 60 m. Philips ladies long jump marks also seem a little low. His W65 long jump mark is 4.54 m, but the world record is already 4.64 m.
    There is an inherent bias in tables that are based on current masters world records. Some events have been much more heavily constested than others. World records in the heavily contested events are harder to approach. Since many awards are given on the basis of age graded percentages, certain groups will be favored. Women’s field events, especially over 65, and most men’s events over 80 are favored. If Don Pellmann had done the decathlon when he turned 90 he would likely have scored more than 12,000 points. Don is a great athlete, but 12,000 points in the decathlon is just not reasonable.
    One solution to this problem is to set up tables based on open class world records and modified by appropriate rates of age related decline. I have done some work on this but find that it is not very popular because in some instances, current world records barely reach 80%.
    I think we need a group of two of three computer geeks to work on this problem. I’ve got taxes to do now.
    Bud Held

  7. Chuck Phillips - March 31, 2007

    ken,
    There is an extremely important difference between how the phillips standards and the WMA Age-Graded standards are generated ?¢‚Ǩ‚Äú my standards use a single two-point equation that fits the records of all masters ages (e.g., 35-100) for each event to determine the age-versus-time aspect of that event?¢‚Ǩ‚Ñ¢s standards; while the WMA Age-Graded standards employ a combination of cobbled together 5-year-best-record segments to determine the standards for an event, and this is done separately for each event with little regard for the standards of neighboring events. The significance of the differences are germane to the blog comments.
    With the WMA standards curve fitting method, the standards of adjoined 5-year age segments are each manually adjusted to provide for a reasonable continuity of the progression of standards from one 5-year age segment to the next, but the beauty of this method is that the progression of standards can be arbitrarily stepped around to pick up weak but importantly sponsored records for inclusion as standards while still having the age 35-40 and age 95-100 segments also be at their desired performance levels. Use of this method is two-fold in that it is simple for curve fitting purposes (no complicated mathematical equations to mess with here, anyone who can balance their checkbook can do this kind of curve fitting), and in addition it allows the curve fitting to be arbitrarily manipulated so that biases or special interests can be accommodated such as favoring either young , medium or oldest masters at the expense the others, favoring either sprint, middle or long distances at the expense of the others, or favoring the records of specific individuals, and so forth.
    Now to be fair, the phillips curve fitting method could also be manipulated to accommodate a special interest, but various steps are taken to prevent that from happening unknowingly or by design. Mathematically, the two-point curve fit over the entire age range prevents any meandering around of the standards between those of the resultant standards for age 35 and age 100 ?¢‚Ǩ‚Äú if you don?¢‚Ǩ‚Ñ¢t like the standards resulting for the intermediate ages, the standard values for either or both of the age 35 and age 100 standards must be changed. In addition, several other aspects of the evolving standards are looked at to see if the standards are suitable ?¢‚Ǩ‚Äú the resultant distribution of age-factor values are looked at, the resultant graphs of running rate curves are looked at to see if there are excursions preventing the curves from having a smooth continuous shape, and the evolving standards for all of the events combined, plotted in a single graph, are looked at to verify that all are of a reasonably similar shape and are reasonably distributed among themselves to prevent favoring any one age group or any one event.
    And as previously mentioned, the phillips standards are mathematically generated in such a manner (as explained in the ?¢‚Ǩ?ìAbout the Standards?¢‚Ǩ¬ù item in the phillips web site) that the standards for all neighboring events are interrelated to take into account the reality that some events have better/stronger records than their neighboring events; while unfortunately that is not the case for the WMA age-graded table standards, a flaw that renders the standards unsuitable for any and all performance comparison purposes.
    I agree with Bud Held that the problem he mentions should worked on.
    Chuck

  8. Jess Brewer - April 1, 2007

    Sounds (from what I just read) like Chuck Phillips has done some thoughtful and meticulous work on this perennial problem. I’d like very much to know the details of his algorithm. If he’d like me to offer it as an alternative to the WMA AGT I’d be happy to do so, but I gather he already has it in his Website — I’ll add a link thereto if he wants. We’re on the same side here: all I ever wanted to do was make it easy for people to compare their performances to those of different age groups, and I used what I could get my hands on. I completely agree that the WMA tables are distorted by the different samples for popular and unpopular events; but in a way this is beneficial, because it tends to encourage more people to do the unpopular events, where they tend to get a better evaluation than they deserve — for example, this works in my favour [long hurdles].
    But when we really want to know the objective truth, I can see that an algorithm like what I imagine Chuck’s to be would be the way to go. Will he share the formula? [Note: I'm not interested in "politics" -- a word derived from the Greek "poly", meaning "many", and "ticks", meaning "small bloodsucking insects".]

  9. Chuck Phillips - April 3, 2007

    Ken ?¢‚Ǩ‚Äú
    Jess Brewer?¢‚Ǩ‚Ñ¢s comment is insightful and spot on. As suggested, it would be fine (and probably beneficial) for my website to be one of his links ?¢‚Ǩ‚Äú some comparative results (mine as an alternative to the WMA AGT) might be interesting.
    The algorithmic considerations are pretty esoteric for many masters participants but as Jess notes they critical to turning out a suitable product ?¢‚Ǩ‚Äú I have tried to explain (in the ?¢‚Ǩ?ìAbout the Standards?¢‚Ǩ¬ù item in the website) in simple terms what most of the suitability-tests are and why it is important that they be done. Suitability tests include age-factor distribution plot tests, plotted running rate tests, distribution and symmetry tests for the plotted standards, with all test plots being those of both (1) standards-by-event plotted across all ages, and (2) standards-by-specific-age plotted across all events. It would be quite instructive to take WMA AGT standards and prepare the same suitability test graphs for those standards.
    Chuck

  10. Ken Stone - April 3, 2007

    After reading my blog bio, Chuck Phillips sent me his own biography. He hopes it establishes his credentials as a statmiester.
    Chuck wrote:
    In 1947, when San Diego had both a morning Union and an afternoon Tribune, I lowered the Hoover High mile record to 4:36.4.
    While at UCLA in a Navy program I was a walk-on mediocre miler and at the NCAA championships in 1952 I was silver medalist in the 3000 meter steeplechase. I received a bachelor degree in mathematics from UCLA and went on to Pensacola, Florida for training as a naval aviator.
    I commenced my career flying on and off aircraft carriers for 4 years and then attended the Naval Postgraduate School where I received my master of science degree in electrical engineering.
    I was then ordered to a developmental test and evaluation squadron serving as a project test pilot and project engineer. I was then ordered to the navy special project office where I was a technical engineer during the development of the Polaris missile program.
    I became the CO of an aircraft carrier based squadron and then served in several different program offices developing and evaluating navy weapon systems.
    In 1976 while serving as Commanding Officer of the Naval Air Station at Lakehurst, NJ, I embarked on my hobby of studying the aging effect on running ability (I was a masters age runner at clubs in NJ and DC for several years).
    I retired as a captain in 1983 and commenced a 14-year stint as an engineering consultant (pejoratively referred to as beltway bandits in DC) and my final effort in that regard found me on a 1-year contract with NASA at Houston and the Kennedy Space Center as an engineering consultant on the International Space Station ?¢‚Ǩ‚Äú and was thereafter referred to as your fully-discredited, proverbial rocket scientist.
    I?¢‚Ǩ‚Ñ¢ve had considerable experience in the art of scientific curve plotting over the years, and now at age 78, I believe that the performance standards that I have developed and brought to fruition are a useful tool that brings credit (and not discredit) to the age based standards concept in the masters t&f community.

Leave a Reply