For the Decathlon fans

I was expecting this article, since my Décapassion friend, Frédéric Gousset, was present during the Décastar at Talence. The article has now appeared on their blog along with superb photos. 

If you have followed the event and you wish to re-live it, just follow this link. And if you have missed it, that's an even more important reason to go and visit their blog and enjoy the excellent article on the 2025 Décastar. (And don't worry if you don't know french. Today's translations are near perfect). 

PS In my post on the "One-day decathlon" I talked about the idea of 4-misses-out  (or 5 or 6) for the vertical jumps. In fact, the more I think about this the more I like it. It's a perfect mechanism to mitigate the no-height risk. Suppose an athlete starts at 4.80 m in pole vault and has three misses. He cannot go higher but he still can try at a lower height. If the organisers have proposed an initial height at, say, 4.40 m and increments of 10 cm afterwards he can use his remaining misses at, say, 4.50 m. If he passes, he can even try at some superior height (but still lower than the 4.80 m where he had three misses). Had such a rule been adopted early enough, the 1992 olympic title would have most probably changed owner.

Hell froze over: I am writing an article on triathlon

Those who read regularly my blog know that I do not like the triathlon. I find the formula, where the stopwatch is running continuously, somewhat unnatural, in particular for short events like the olympic one. The athletes must manage the change of attire and equipment, between swimming and biking and again between the later and running, losing the least possible time. In an event of, roughly, 50 km where the best athletes are taking well below 2 hours, the loss of one minute during equipment change can easily modify the outcome. But, of course, when compared with the supposedly "modern" pentathlon the triathlon is infinitely more interesting. 

Now, to be fair, my objection concerning the transition time from one specialty to the other is no more valid when it comes to the original triathlon formula, the Ironman. In an event that takes 7-8 hours to complete, the transition time is negligible. Talking about the Ironman as the "original" triathlon needs some explanation. It is true that the first competition over the Ironman distances (3.8 km swimming, 180 km of bike and running a marathon) took place in 1974 and the name was coined in 1978 with the creation of the Hawaii triathlon. However the very idea of a swim-bike-run event can be traced back to France and the beginning of the 20th century with  the famous « Course des Débrouillards ». The names « Les trois sports » and « La course des Touche-à-tout » were also used and sometimes the swimming was replaced by canoeing. (Were I a fan of triathlon, I would have inserted a « cocorico » here).

How did I end up writing about the triathlon? As you may remember I published recently two articles on Ultra-running and the Speed Project, both having to do with running events that stretch over several days. While researching for the latter post I stumbled upon an article, mentioning the fact that some canadian lady athlete had just completed a Deca Ultra-triathlon in 11 and a half days. I read the detail and I could not believe my eyes. The event consists in 10 Ironmans lumped together: 38 km of swimming, 1800 km of biking and 10 marathons. I decided to learn a little bit more about these crazy events and I found out that the "deca" is not the last word. What started in 1985 as a double Ironman (or was it the 1983 Ultraman that launched the ultra-triathlon events?) has been steadily expanding with, today, 20x and 30x events. And the two formulae do exist, a continuous one and a "one per day". 

When I saw the performances of the triathletes in events spanning up to 40 days I was immediately reminded of my work in collaboration with J. Meloun and G. Purdy (A mathematical model for scoring athletic performances, Math. and Sports 5 (2023) 1) where we analysed the running velocities for distances going from 20 m to 1600 km. As explained in my post on running velocities, for multi-day events the additional slow-down mechanism is due to the fact that the athlete must take time to sleep, eat and take care of other bodily functions. And, in fact, for races with duration above 12 hours, a non-negligible amount of time is devoted to non-running moments. In our paper mentioned above, we had presented a fit of the velocity as a function of the distance for various ranges of distances. The one that is interesting here is one covering the distances from 100 to 1600 km, for durations from roughly 6 hours to over 10 days. In the figure below I present the data together with the best fit of the form 

leading to an exponent γ close to 0.35.

Next I turn to the ultra-triathlon data, plot the velocity (in the appropriate units) as a function of the repetitions of the Ironman and fit the date with the same formula and with the same exponent γ. (The choice of units affects the value of A but not that of g).

We can remark that the fit is quite satisfactory. Thus we may conclude that the ultra-marathoners and the ultra-triathletes face the same difficulties linked to the excessive duration of the event and the fact that they must reserve time for sleep and other body's needs. It is also telling that in the case of the triathlon the fit over-estimates the velocity for the 30x event. Since the event requires more than a month to be completed it is natural that long-term fatigue sets in with as a result a further velocity degradation. One should compare this to the fit we presented in the article mentioned above for the velocity as a function of the distance for distances from 1 km to the marathon. There, the exponent gamma is much smaller, around 0.07. Ij this case the velocity degrades very slowly, in a mechanism remaining predominantly aerobic over the whole span of distances, the decrease being due mainly to the onset of fatigue. 

When I set out to write this article I was a triathlon-skeptic. After having read several articles on the origins of the discipline and on its "ultra" forms I am less so. But, let's be frank, I am nowhere near becoming a fan. One reason for this that the triathletes I am meeting are invariably as...les, considering themselves as "ironmen" despite their obvious mediocrity. People who run know very well that it's not because they managed to finish some marathons that they may consider themselves on par with Eliud Kipchoge. This is a lesson that triathletes have yet to learn.

Night of records in Eugene

I was not planning to write something on the Eugene Diamond League. But when I saw those superb images of Béatrice Chebet celebrating Kipyegon's 1500 record more than her own in the 5000 m I felt that I had to do something. 

Chebet and Kipyegon

But let us start at the beginning. And I will start with a missed record, that of Winfred Yavi in the 3000 m steeple. She ran in 8:45.25 at less than a second form the world record of Béatrice Chepkoech. This is the second time that Yavi gives the record (that goes back to 2018) a scare. And since she is just over 25 years old she has several good years ahead of her that would probably allow her to bring the record bellow 8:40. (But she should beware of Faith Cherotich and Peruth Chemutai who are just 21 and 25 respectively and who ran in 8:48.71 and 8:51.77 finishing behind Yavi. In fact Chemutai was second behind Yavi last year in Rome when Yavi ran her best 8:44.39, finishing in 8:48.03). Those who follow my blog know that I am a great admirer of Béatrice Chepkoech (I insist on her given name because there is a second Chepkoech, Jackline, running also the 3000 m steeple, with a 8:57.35 from 2023 and who is not quite 22 years old. She looked at some point as the successor to her homonyme. Unfortunately her last two years have been really below-par: her 2024 best was 9:19 and this year she has not done better than 9:30). However when Yavi made her appearance I immediately saw that we had there a pure gem. She did take her time to reach the summit (and there was a time, like in my report on theTokyo Olympics, when I was voicing my disappointment for her 10th place in the final). And thus whenever the two, Chepkoech and Yavi, ran together I knew I was going to be frustrated, since only one of the two could win. 

Béatrice Chebet confirmed (it's not as if anybody had a doubt) her position as the queen of middle distances. After having broken the 29 min barrier in the 10000 m, she did break another magical barrier, that of 14 min in the 5000 m. The men's 15 min record was improved in 1904 and the 14 min fell in 1942 when G. Hägg ran 13:58.2. His record was improved by such illustrious names as Zatopek and Kuts in the mid fifties. Things are even more spectacular in the 10000 m. It took Zatopek from 1949 to 1954 and five successive records before he managed to finish under 29 minutes. Seventy years later a woman manages to do as well as Zatopek! Paola Pigni was the first woman to run under 16 minutes in 1969 and Ingrid Kristiansen broke the 15 minutes barrier in 1984. Then it took another 40 years for the next barrier to fall. In the case of the 10000 m  Mary Decker Slaney was the first to run under 32 minutes in 1982, Kristiansen went under 31 just three years later but the 30 minutes barrier fell under the highly questionable chinese 1993 boom, thus muddling the statistics. Chebet, with last year's 28:54.14, set the record straight (pun intended).

I did not like the idea of the sub-4 mile but I do like the 'Nike Fly Suit'

I was somewhat skeptic concerning Faith Kipyegon's chances at a world record. In particular after the crazy, publicity-motivated, attempt at a below-4 mile. (Let's get real. Kipyegon's record is a great 4:07.64. Asking for a 3% improvement in one go is ridiculous. And one could see that Kipyegon, running already all out in the first three laps, was struggling on the stretch opposite to the finish. She managed a 4:06.42, but I am convinced that she could have done equally well in a normal race with just one or two female pacers and a sensible race strategy. And in Eugene she adopted the best possible one. She followed the pacer, S. Hurta-Klecker a 4:01 runner, who did a great job, up to 1100 m and then took off. While in Paris she faltered over the last two hundred meters, in Eugene she was flying. Even J. Hull, who is admittedly the second best 1500 m runner, could not follow her opponent's devastating sprint. Kipyegon finished in 3:48.68 and the celebrations followed. And Kipyegon redeemed herself after the Paris failure. 

PS No, I am not going to comment on the article "Could a female athlete run a 4-minute mile with improved aerodynamic drafting?" by da Silva and collaborators (Roy. Soc. Open Sci. 12: 241564).  It is easy to fiddle with parameters and obtain the result you wish when it is just a 3% effect. Obtaining that on the track is a totally different matter. But let us not be negative and just rejoice at these two superb records we watched in Eugene.

The one-day decathlon

This article is directly inspired by a discussion I had with my Décapassion friends, Frédéric and Pierre Gousset. We were talking about one-hour decathlon and then they mentioned the Iron Deca organised in Arles roughly ten years ago. I don't know how long this event survived (probably just two-three years) but it piqued my curiosity and I decided to follow the trail. The Gousset had devoted an article to this event entitled "how long should a decathlon take?" where they were mentioning an article by ukrainian coach, A. Fatieiev, published in New Studies in Athletics and who was addressing the question whether a one-day decathlon was possible.  Having read Fatieiev's article I decide to present my ideas in this post.

But let's start at the beginning. The Stade Olympique Arlésien organised, in 2013-15, what they called the Iron Déca, a meeting where the whole decathlon (heptathlon for the women) was held in one day instead of two. The organisers were planning two sessions of 2h30 separated by 4 hours of rest. Given the tight schedule, the competition was limited to groups of 8 athletes. Preparing the soil marks for the jumps had to be done before the competition. But, most important, no rule modification concerning the jumps and throws were introduced. So, it was a perfectly normal decathlon, condensed in just one day. 

The only result I could find is that of the 2013 winner, Bastien Auzeil, with 7396. His best "normal" decathlon performance that year was 8022 points, situating the one-day performance at 92% of the best. This is, and by far, not exceptional. Both the world and the french best one-hour decathlon performances (Zmelik, 7897 and Blondel, 7715 points) are at roughly 92% of their maximum, of the same season (Zmelik, 8627 and Blondel, 8387 points). And when one thinks about the out-of-this-world 96% performance of Dadic in a one-hour heptathlon, 6325 to 6552, the performance of Auzeil looks somewhat paltry. 

While writing this article I did revisit some of my old articles on decathlon variants. And I ended up invariably at the article on C. Beach's amazing 10 minute decathlon. (If you are a decathlon fan and haven't read it, take a minute to do so).

What is the problem of the one-day decathlon? This is best summarised in the words of M. Avilov, who won the olympic title in Munich, 1972, with a world record of 8454 points: "without the 400 and 1500 m I could easily do a decathlon in one day". And while one can argue that when it comes to the 1500 m one can use his last reserves, the 400 m poses a major difficulty. There is simply no way one can go all out in the 400 m and recover completely in the 4 hours resting period so as to attack the remaining events in top condition. Just to realise the impact of the 400 m it suffices to look at the performances of C. Beach in the 30-minute decathlon (where he scored 6242 points to be compared to his 8084 best). With a personal best of 46.72 s in the 400 m he ran in 53.26 s during the 30-min, followed by 16.96 s in the 110 m hurdles where he has a personal best of 14.23 s. And Beach is not just anybody: he is exceptionally resistant with a 3:59.13 personal best in a decathlon 1500 m (and an even better 1:47.36 in the 800 m). So, even with a 4-hour rest, the one-day decathletes must run a cautious 400 m and, even thus, their 110 m hurdles may be below par.

But let us go back to Fatieiev's proposals. He is worried by the fact that the decathlon, spread over two days, does not attract the interest of the un-informed spectator. Well, let us be serious: if somebody does not care about a two-day decathlon, would he be attracted to a one-day one? I beg to differ. Fatieiev proposes to reduce the number of attempts for throws and long jump to two instead of three. I don't see any appreciable time savings due to this. The proposal of Iron Déca organisers, to prepare the marks for jumps in advance, makes more sense. The one point where Fatieiev is correct is that a disproportionate time is spent in the two vertical jumps. But there is not much one can do about this. In a one-hour decathlon, where the stopwatch for the 1500 m starts at precisely 60 min, whether one is present at the start or not, the athletes have their own strategies for the vertical jumps. But what Fatieiev suggests is to limit the attempts to one at every next height or to allow for three attempts in total. In their article, the Gousset, comment that something like this would convert Athletics to Circus Games. I totally agree with them. Decathlon is the king of events in athletics (an event where, to date, women are banned from, due to World Athletics spinelessness). A two-day decathlon is a perfect event.  A one-hour one, albeit totally different, is equally perfect. A one-day decathlon a futile exercise in style. 

PS I have just watched the European Athletics team championships and I discovered a new rule applied to vertical jumps: an athlete is out after 4 misses (while at every height the regulatory 3 misses apply). Perhaps a 4 (or 6, or any number obtained after a statistical analysis of existing results) misses rule could be used to speed up the decathlon vertical jumps. This would make more sense than limting the number of attempts.