The most recent World Championships were an occasion for me to have a closer look at the photo-finish snapshots and I must say that I am now more than ever convinced that it is high time we drop these pesky photos as a way of determining the order of arrival. (By the way I must say that the IAAF is doing a great job when it comes to championships results. They constitute a great basis for statistical studies and the existence of the photo-finish clichés, all the way back to the Paris, 2003, championships is a substantial added value).
Before starting to criticise the use of photo-finish snapshots I must make clear that I do not believe that the judge's eye is in any way better than a photo. Our brains, as those of every decent predator, are hard-wired so as to calculate trajectories and extrapolate movement. Thus we are often tempted to believe erroneously that a runner coming fast from behind has caught up with the one in front while the photo shows clearly that this is not true. The third place in men's 400m hurdles final is an excellent example of this
No, my point here is that the analysis of photo-finish, since it is based on human judgement is error-prone and can lead to unfair decisions. In fact I already wrote about this problem in my post on the absurdity of milliseconds, but, motivated by the recent World Championships results, I would like to elaborate further on this point here.
Let me start with the photo-finish of the men's 200 m. When I watched the race I had the impression that Jobodwana had lost to Edward, which was a minor deception since I was hoping that the south-african athlete would win a medal this time. However it turned out that Jobodwana was given the bronze medal for a mere 2 millisecond difference.
Again this time Edward was coming from behind, but when I look at the photo I have trouble to find any appreciable difference (try to wipe out mentally the green guidelines and have another look at the arrival snapshot).
I do not imply that the judges do not know their job and I am certain that the ones analysing the photo-finish clichés have years of experience. Still I recall the interview of R. Jennings, photo-finish analyst extraordinaire, after the Felix-Tarnoh-gate at the 2012 US Trials. His point is that the problem lies in locating the torso and it is a really tough one. And when it comes to minute, of millisecond order, differences, the final judgement may become subjective. Have a look at the women's 100 m hurdles down at 6th position. Can you tell where Zbären's torso is really located, given that her right arm hides it almost completely?
Is she really ahead of Williams? I am not convinced.
Men's 100 m final was perhaps the most exciting event of the 2015 Championships. A substantial fraction of journalists, who do not understand much of athletics, were convinced that Gatlin would put an end to Bolt's supremacy. As we know things went quite differently. It remains that during the race our eyes (well, mine at least), were riveted on the two front-runners and I had to wait for the nth replay before looking for the bronze medalist. I was not prepared for the result. The two new great sprints talents, De Grasse and Bromell, not only did they manage to beat far more experienced and famous competitors but they tied for the bronze medal.
Of course looking at the photo-finish I cannot find any discernible difference between the two and for me the tie is 100 % justified. But that got me thinking. Suppose they had tied at first place. Would some over-jealous tie-breaker judge invent some inexistent difference in order to choose a single winner? I shudder at the thought. (And I will come back to tie-breaking in athletics in some future post).
If photo-finish is obsolete what can we do in order to have an as accurate as possible arrival judging? For me the only answer can be a fully automated one, without human intervention. Although I do not pretend to have the full solution I can imagine the timing being obtained through a transponder fixed on the body of the athlete itself. A light, pill-like, chip fixed with adhesive on the top of the athlete's sternum (the manubrium) by a special judge at the pre-departure procedure (to be retrieved just after the race by the same judge) could provide a most accurate timing, hopefully with sub-millisecond precision. Transponders are already used in car racing where the cars are hurtling down the runway at much (much) higher speeds than those of human sprinters, so why not implement this in athletics. Were such a thing to see the day I would no more object to a millisecond-difference-based classification.
Before starting to criticise the use of photo-finish snapshots I must make clear that I do not believe that the judge's eye is in any way better than a photo. Our brains, as those of every decent predator, are hard-wired so as to calculate trajectories and extrapolate movement. Thus we are often tempted to believe erroneously that a runner coming fast from behind has caught up with the one in front while the photo shows clearly that this is not true. The third place in men's 400m hurdles final is an excellent example of this
No, my point here is that the analysis of photo-finish, since it is based on human judgement is error-prone and can lead to unfair decisions. In fact I already wrote about this problem in my post on the absurdity of milliseconds, but, motivated by the recent World Championships results, I would like to elaborate further on this point here.
Let me start with the photo-finish of the men's 200 m. When I watched the race I had the impression that Jobodwana had lost to Edward, which was a minor deception since I was hoping that the south-african athlete would win a medal this time. However it turned out that Jobodwana was given the bronze medal for a mere 2 millisecond difference.
Again this time Edward was coming from behind, but when I look at the photo I have trouble to find any appreciable difference (try to wipe out mentally the green guidelines and have another look at the arrival snapshot).
I do not imply that the judges do not know their job and I am certain that the ones analysing the photo-finish clichés have years of experience. Still I recall the interview of R. Jennings, photo-finish analyst extraordinaire, after the Felix-Tarnoh-gate at the 2012 US Trials. His point is that the problem lies in locating the torso and it is a really tough one. And when it comes to minute, of millisecond order, differences, the final judgement may become subjective. Have a look at the women's 100 m hurdles down at 6th position. Can you tell where Zbären's torso is really located, given that her right arm hides it almost completely?
Is she really ahead of Williams? I am not convinced.
Men's 100 m final was perhaps the most exciting event of the 2015 Championships. A substantial fraction of journalists, who do not understand much of athletics, were convinced that Gatlin would put an end to Bolt's supremacy. As we know things went quite differently. It remains that during the race our eyes (well, mine at least), were riveted on the two front-runners and I had to wait for the nth replay before looking for the bronze medalist. I was not prepared for the result. The two new great sprints talents, De Grasse and Bromell, not only did they manage to beat far more experienced and famous competitors but they tied for the bronze medal.
Of course looking at the photo-finish I cannot find any discernible difference between the two and for me the tie is 100 % justified. But that got me thinking. Suppose they had tied at first place. Would some over-jealous tie-breaker judge invent some inexistent difference in order to choose a single winner? I shudder at the thought. (And I will come back to tie-breaking in athletics in some future post).
If photo-finish is obsolete what can we do in order to have an as accurate as possible arrival judging? For me the only answer can be a fully automated one, without human intervention. Although I do not pretend to have the full solution I can imagine the timing being obtained through a transponder fixed on the body of the athlete itself. A light, pill-like, chip fixed with adhesive on the top of the athlete's sternum (the manubrium) by a special judge at the pre-departure procedure (to be retrieved just after the race by the same judge) could provide a most accurate timing, hopefully with sub-millisecond precision. Transponders are already used in car racing where the cars are hurtling down the runway at much (much) higher speeds than those of human sprinters, so why not implement this in athletics. Were such a thing to see the day I would no more object to a millisecond-difference-based classification.
