In Olympic trials, tight races are decided by a photo finish. But on Saturday, June 23, runners Allyson Felix and Jeneba Tarmoh defeated the cameras to tie for third place in the Women’s 100 Meter Dash, calling the finish-line system into question. Is the technology used to judge Olympic competitions fast and precise enough to rank the top performers?
Determining who finished Olympic Trials in what place is usually straightforward. The Olympics, like other international track competitions, require the use of fully automatic time (FAT) systems. The starter’s pistol connects to a clock that starts when the gun fires. In this case, the clock determined that both runners finished in 11.068 seconds, so the order of finish was determined by an official viewing a panoramic image created by two digital line-scan cameras focused extremely narrowly on the finish line. The cameras shoot 3,000 frames per second, which are stitched together to form an image of the finish line.
Usually the photos make the order of finish obvious. But this race was different.
(Incidentally, the judging organization, USA Track and Field, had no established protocol for breaking such a tie. Judges cobbled together a tie breaker post-facto: Each athlete was offered two options: a coin toss or a runoff. If the two don’t agree, they will face a runoff.)
While much has been made of USA Track and Field's poor planning and poor response to the tie, the technology used to judge the race has gone mostly unremarked upon. Ironically the organization's president, Stephanie Hightower, in the 1984 trials was deemed the fourth-place finisher after a three-way tie for second in the 100-meter hurdles. The clock said that she and two other competitors finished in 13.13 seconds, and Hightower was awarded fourth place by what she described to the Associated Press as a photo that was blown up to such a degree that it was indecipherable.
Faster cameras do exist. The Olympus i-Speed 2 captures 33,000 images a second, an order of magnitude faster than the units employed by USA Track and Field. The i-Speed 2 was introduced in 2006, and can be purchased new for $39,000, or used for $9,500. Considering the detailed imagery of rapid events that can be captured at 10,000 frames per second, an Olympic timing system using such cameras would be a marked improvement.
However, clock resolution isn't the only variable at play in Olympic competitions. Another is distance; in this case, 100 meters. Swim races, which use touchpads to determine each swimmers’s finish time, have shown that that clock's precision is greater than that of pool construction. In theory, each lane in an Olympic pool is 50 meters long. In practice, most lanes are slightly longer or shorter, even within the same pool. While such variations may not be an issue in most competitions, at the Olympic level, when many finishes are extremely close, they are.
Omega is the official timekeeper for Olympic swim races. Following a 2008 event won by .01 seconds in a photo finish, the website Athletic Business quoted Omega's website, "Omega touch pads and starting blocks are part of an integrated timing system capable of recording times to the nearest one one-thousandth of a second. However, because it is not possible to build swimming pools in which each lane is guaranteed to be precisely the same length, Olympic and world records are still recorded to the nearest one one-hundredth of a second."
The 10,000-frames-per-second cameras used in Tuesday's trials captured the athletes' positions at intervals of 3.3 millimeters along the track. And while that can be improved – at 33,000 frames per second, each image would represent one millimeter of travel – other critical measurements are likely to be far less precise, at least for the foreseeable future.