I could not resist making a nod, through the title of this article, to the great opera of John Adams "Nixon in China" on an idea of Peter Sellars (I am an unconditional fan of Sellars' work). What did spur this choice of mine was the fact that Gerry visited Copenhagen at around the same time when Nixon visited China (February 1972). Nixon's visit is considered by historians as a significant one, pushing the Soviet Union to enter an era of détente with the US.
Purdy's visit, although it did not lead to his scoring tables be officially adopted by the IAAF, had managed to convince the members of the IAAF Technical Committee that the existing scoring tables were unfair and had to be replaced. It took them more than 10 years to do this. Had they accepted to work with Purdy, combined events would have been spared a frustrating decade. The story of Purdy's visit to Copenhagen is told below in his own words without one iota of editing.
My Visit with the IAAF Technical Committee, Winter 1972
(Or From Russia without Much Love)
Copenhagen, Denmark
J. Gerry Purdy, Ph.D.
Recall April 8, 2022
Boca Raton, Florida
Edited by Jalane Meloun, Ph.D. and Basil Grammaticos, Ph.D.
Background
In 1966-68, I was training and running various road races in Southern California while working at TRW Aerospace (now Northrup Grumman) and finishing my M.S. degree in Computer Science at UCLA.
My running partner and friend was James B. “Jim” Gardner. Being engineers, we began to study running training as a science. We realized that while there were lots of coaching books to help runners train, there wasn’t any way to scientifically adapt them to each runner’s specific level of ability and situation.
Jim and I developed some models of performance, and I then developed some software from ’69-’70 that calculated the runner’s level of ability (point score) and from the level of ability generated tables for both interval training on the track and for road running training.
Jim and I put together the tables and explanation in a book called Computerized Running Training Programs [1].
It was published by Track and Field News in 1970.
During that process, we noticed that the IAAF Decathlon Scoring Tables [2] used to score the 10 different Decathlon events didn’t make any sense. The Tables were progressive for the running events, i.e. that they properly designated more points per unit change in performance as you went up the performance scale. But we found that the tables for the field events were regressive – namely that you got less points for the same unit change in performance as you went up the scale. That didn’t model physiology. Something was basically wrong.
At this time, Jim shared with me that he had developed cancer and didn’t have long to live. I was living in the San Franciso Bay Area at the time attending Stanford, so I flew to LA to visit with him back just before he passed away. It was terrible to lose such a close friend so suddenly. I told him I’d continue our work.
With the information about the problem with the IAAF Decathlon Scoring Tables, I figured I’d call the IAAF in London and let them know about the problem so that they could then fix it. I didn’t know how they were constructed. I knew a good bit about running training but not anything about the way the Olympics were managed or how the multi-event tables were created.
I found out that John Holt was Executive Director of the IAAF Technical Committee. His contact information was listed in the back of the booklet for the IAAF Scoring tables along with the IAAF phone number in London. I never have a problem taking the initiative, so I gave him a call. I can distinctly remember hearing that double tone when the phone rings in Europe vs a single tone in the U.S. After a few rings, a distinguished sounding gentleman answered. I asked if I could speak with Mr. John Holt. I identified who I was calling from Stanford University in California. He said he was John Holt and to call him John.
At this point, I recall that John responded with something that was the very opposite from what I expected him to say. I thought he’d say something like, “Well, jolly good, we’ll take it under advisement and get back to you.”
Instead, he said, “Well, why don’t you fix it?” I thought, “Huh, I don’t even know how those tables were created let alone to fix them!” I told him I’d look into it.
I was able to find out from John that the 1962 Decathlon Scoring Tables [3] were developed by Dr. Karl Ulbrich in Austria. I got his address and had a friend at Stanford who knew German translate my letter to him asking him about how he created the tables.
Surprisingly about 2-3 weeks later (think of that compared to email today), I received a nice response in German that my friend translated into English. There is plenty of reference materials including these letters available in the Stanford Computer Science Department Library.
Ulbrich explained how he developed his tables: he simply used velocity for the running events and the square root of distance and height as the metric to be equivalent to velocity for the field events [4]. While elegant mathematically, using the square root of the distance/height for the field events resulted in generating a model that was regressive instead of progressive. His table was adopted in 1962 by the IAAF (now World Athletics).
Through this historical research, I realized that I had the opportunity to do my Ph.D. thesis on finding a better way to generate the scoring table for the Decathlon in the Olympics. Quite a challenge to say the least. I wrote a dissertation proposal and got it approved. Then, I had to figure out how to develop a progressive formula from which I could develop a computer program (in FORTRAN) and create something to show to John Holt and the IAAF Technical Committee.
A professor in the Operations Research Department was interested in my work and agreed to help me. At first, I thought some kind of exponential term would do it, but I found when running the calculations that a single exponential formula would either work well at the high end of the scale but not at the low end or the opposite: that it would work well at the lower end but not the higher end.
The professor suggested we use a linear term for the low end and an exponential term for the high end which gave now my ‘famous’ formula that is the basis for what we are using in our new much improved approach now:
Where -
p is the point score or level of ability which we now call the TraxScore
x is the performance mark (velocity for running and height or distance for field events)
a,b,c are constants
z is the zero offset or the value below which is deemed to not be a performance.
I did some fiddling and determined that the zero offset should be around 2 m/sec for running and a set of arbitrary but rational values for the field events.
Next, I then had to define three points along the curve to solve the equation for three constants. I talked to Bert Nelson (then Publisher at Track and Field News) who provided me with access to statistics from their publications. In the end, I had about 100+ performances that were defined a priori as:
• High (1400 points) – The world records at the time with some hand adjustments.
• Middle (1100 points) – The average of the 50 top performances in the world.
• Low (500 points) – The average of the 50 top master’s performances in the 50-60 ages.
The values for the 500, 1100 and 1400 point levels are included in both [3] and [6].
In retrospect, it’s better to have this rough estimate of what a good model should present than not to have anything. Better (as we are now doing) to have large datasets that allow the scoring function to best model the behavior of the entire population for all the major events. Then, the calculation of the TraxScore for odd distances will follow along the path of the known population from low to high performances.
I then used the least squares subroutine available from the computer center and ran the solution over and over until the difference between the constants calculated for each event was miniscule. Voilà! I had a solution that worked for all 10 of the Decathlon events plus a number of additional road running events such as the 5K, 10K, Half Marathon and Full Marathon.
My program kept growing and growing as I had to program printouts that were attractive that was very difficult to do in FORTRAN! My program grew to over 10,000 lines of code. It was all originally prepared on punched cards and took up boxes to get stored on the IBM 2311 disc packs. I vowed I would never write another computer program again!
One humorous vignette: I was called in to talk with the director of the Stanford computer center. He told me I exceeded the total storage that was allowed for graduate students: 5MB ... not GB but MB! I explained what I was doing, and he was nice and said he’d approve my having allotment of 10MB but to get it done by June and take it down when I was finished. I did.
I was giving John Holt updates via letters and occasional phone calls. He was delighted that I had made so much progress. He then invited me to present my new tables to the IAAF Technical Committee that was meeting in Copenhagen, Denmark in February 1972 before the Olympic Games that summer in Munich, Germany.
Well, I was graduating in May 1972, so I put together my tables into a concise write up (that later was published in Medicine & Science in Sports [4-6]) and sent the writeup to John who then made copies for all the Technical Committee members to read before I showed up. Other relevant articles include [7-9].
The IAAF Technical Committee Meeting in Copenhagen
I can’t remember the name of the hotel where the meeting was held, but it was fairly close to Tivoli Gardens which I went by to see where it was located while there even though it was closed for the winter.
John hosted a welcome reception for everyone who flew in from all over the world. I had a chance to talk to a number of the IAAF Technical Committee members and tell them why I was there.
The meeting the next day was rather uneventful. I waited outside the room as they didn’t want me to hear some of the things they were discussing. I was brought in during a break and presented to them using overhead foils (remember them?). My research and development were received quite well, and the questions they asked were all reasonable. They appreciated my work to correct the definite problem with then approved scoring tables. The meeting lasted around an hour.
At the break, I was gathering my things when a rather stout lady in a printed dress approached me with a badge indicating she was from the USSR. She said – with a strong Russian accent – something along the lines of, “Dr. Purdy, the leader of our delegation would like to meet you in the hotel lobby for coffee at 4 pm before our dinner this evening.” I told her that would be delighted to meet with him. Then, she said something that surprised me, “If you don’t mind, could you bring me a book in English of a popular novel in America?” I sort of pulled back a bit mentally and she continued, “I want to improve my English as a translator.”
I told her I would try to satisfy her request. It was clear she didn’t have easy access to English books even though she could likely find one locally in a nearby bookstore. (I remember thinking at the time that while she could go buy a book in English, she wanted me to select something that was popular in order to better understand American culture.)
I spent the day looking around trying to find a US novel in English (vs. a translation into Danish). The first one I found was The Carpetbaggers by Hold Robbins. I felt this just wasn’t the right story to give her. I also thought of giving her Margaret Mitchell’s Gone with the Wind but also felt it wouldn’t be appropriate glorifying the Civil War and segregation in the south. I think I settled on Love Story by Erich Segal and found it in a nearby bookstore.
Around 4 pm that afternoon, I went to the lobby to meet the head of the delegation from the USSR along with his translator. He was a bit stoic – formal and let his translator do the introductions. Somehow it came over me that this woman was more than a translator for this distinguished emissary. Just a gut feeling. He wore a suit and tie. I was just a Ph.D. student from Stanford, but out of respect I did wear a suit and tie.
So, we finally got down to talking about the research papers I had written about my new model of human performance and generation of scoring tables. He would talk in a sentence or two in Russian, and she would then translate that into English. It was rather slow going as she talked much slower than him thus extending the time of the dialog.
He thanked me for my sending the research papers to him. Without missing a beat, she interpreted and said, “Your scoring method is really quite good. But we have our own scoring system that we are submitting for review and consideration.” I thought, “Well, this is interesting. Maybe I can learn something about how others approach solving the scoring table problem.”
I then asked him, “Why that’s very exciting. How is your system determined? May I see of your research papers?” Off to the translator. He nodded as in taking in what I had asked. He then responded with a bit of high brow authoritarian emphasis. He responded slowly directly to me in broken English, “Our system very similar to yours. We changed few things. We not give you any information about our system. We submit directly to the IAAF with our name on it.”
I clearly remember being dumfounded. Here I was a very innocent young American talking to someone of high position in Athletics from the USSR. I then thought, “Holy shit, he’s taking my system and putting his name on it!” I sat there not sure what to say next. I felt like having been run over by a steamroller.
There wasn’t the miniscule of a smile on his face to indicate that he was just joking like you might think would happen in the U.S. You know, like suddenly smiling and saying, “Just kidding.” Nope, he was dead serious.
I wondered if there was anything else to discuss. I also wondered why he just told me that. If he was really just going to put his name on my work, why in the world did he mention that? It just didn’t make any sense. But that stern expression when he told me was one of superiority and not caring what he was doing. [Seems very similar to the way Putin's Russia decided to invade Ukraine with impunity. Culture lesson.]
I may have said something innocuous but likely simply thanked him for his time. I then handed my book gift wrapped in a bag to the translator, “Here’s a small gift for setting up this meeting.”
But wait, there’s more!
I found John Holt just before dinner and related to him what had happened. As I told him the story, a smile developed on his face as he was dying to tell me something funny. Once I finished, he responded with a smile still on his face, “Gerry, we’ve known that fellow for many years. He always acts that way, and we just don’t pay any attention to him. Don’t worry, he’s not going to put his name on your work and get away with it.” Phew!
John hosted a magnificent closing dinner for the IAAF members. They were very kind to include me. As a young Ph.D. student and soon-to-be graduate, I was almost awestruck by it: the menu was printed on silk fabric with the Olympic name and logo at the top. The dinner was served all at once with metal warming covers and then removed concurrently, and – best of all – dessert was baked Alaska with sparklers on top and served to us with the lights dimmed in the room. What a great experience!
Great ending to a fascinating story.
P.S. John later told me that everyone agreed with my principles for generating future scoring tables. They set up a committee to figure out how to generate the next official table based on my work. It took them 12 years as the next version was adopted in 1984.
References
1. Purdy, J. Gerry, Computerized Running Training Programs, Track and Field Press, Palo Alto, CA 1970.
2. IAAF, Scoring table for the men’s track and field events, IAAF, London, England 1962. (Note: IAAF is now renamed World Athletics).
3. Purdy, J. Gerry, “The Application of Computers to Model Physiologic Performance in Track and Field,” Ph.D. thesis, Stanford University, June 1972.
4. Purdy, J. Gerry, “Computer Generated Track and Field Scoring Tables: I. Historical Development,” Medicine & Science in Sports, Vol. 6, No. 4, pp. 287-294, 1974.
5. Purdy, J. Gerry, “Computer Generated Track and Field Scoring Tables: II. Theoretical Foundation & Development of a Model,” Medicine & Science in Sports, Vol. 7, No. 2, pp. 111-115, 1975.
6. Purdy, J. Gerry, “Computer Generated Track and Field Scoring Tables: III. Model Evaluation & Analysis, “Medicine & Science in Sports, Vol. 9, No. 4, pp. 1212-218, 1977.
7. Grammaticos, Basil and Purdy, J. Gerry, Meloun, Jalane, “Distribution of performances and scoring in athletics,” Math. and Sports 3 (2022) 1.
8. Grammaticos, Basil, “The physical basis of scoring the athletic performance,” New Study. Athl. 22:3 (2007) 47.
9. Grammaticos, Basil, Purdy, J. Gerry and Meloun, Jalane, “Scoring running performances,” draft, March 2022.
No comments:
Post a Comment