Dear Video Judges: This Is Why We Can't Have Nice Things
This is in no way meant to criticize a young lady by the name of Denisse Tame from CrossFit Alpha in the Latin American Region. She’s a fine athlete I am sure, and a nice person. But Denise had the unfortunate job of pointing out just what a flagrantly sucky job we video judges are doing during the 4th Annual CrossFit Games Open.
You see, Denisse submitted a score of 198 reps in workout 14.3. That, my friends, put her at the top of the leaderboard for this workout. Having seen her name at the top of the list and with it, an attached video, I naturally logged on to give it a look. When I first looked at Denise’s video submission page, it had not been judged at all, which I found curious, in light of the fact that the score was so high. So, I started watching. Perhaps Denisse was one of those outlying deadlift monsters who would come out of nowhere and literally obliterate 14.3.
Having seen several athletes, including Stacie Tovar and Alessandra Pichelli do this workout, I knew what the first set of deadlifts might look like in order to garner even a mid-150s score, let alone a near-200. Denise was not moving at said speed. So, I sat back, started the video over, put my feet up, and started counting reps.
Part of a judge’s job is not only to verify the quality of the reps, but the quantity as well. This requires some work. You likely need to sit there with a pencil and paper and make tick marks as you watch, just like you would a live judging situation. Any judge who takes this seriously enough will keep track of the reps in addition to watching for proper range of motion. So, I started counting and as I suspected, Denisse ground to a near-standstill by the time she hit the 185lb lifts. Her range of motion was fine, although it’s possible the bar did not touch the floor once or twice, and her box steps were serviceable.
But as I said, it only took viewing the first ten reps to know without a doubt that the score had been input incorrectly. What I gathered is that Denisse meant to enter a 98 and instead, put 198. No big deal. We can all figure this one out pretty quickly.
Or can we? It turns out that fourteen people judged this video as “Workout Is Good.” Fifteen judged it “Submission Is Rejected.” Not one person fell in the middle with a “Score Needs Modification” judgment, mainly because a video has to be rejected and resubmitted in order to enter the proper score. You can’t just change your score online.
Let that sink in for a second. Fourteen people, who took the online judges’ course, clicked good on a video submission that had the world-record score, which means, clearly, they did not watch this video. We can’t see who judged this specific video, but interestingly, there is a Judges Leaderboard on the CrossFit Games website. This allows you to see all of the certified judges, as well as how many videos and how many live performances they have judged.
It only takes a quick glance at the top numbers to see that something is drastically wrong here. The highest number of performances judged is 2,696, by Jan Dragsbaek of Europe. Yes - 2,696 videos judged. Give or take, that’s 360 hours of video judging, which averages to ninety hours per week. Even conservatively estimating fifty hours per week, that means Jan is sitting in front of his laptop from sunup to sundown every single day judging Open videos. Ninety to a hundred videos per day.
(Note: in the time I have been writing this article, about two hours, Jan has judged 28 additional videos).
But getting back to Denisse Tame’s 198 on 14.3. The fact of the matter is that 14 people hit “Workout Is Good” having obviously not watched the video. The upshot of this is that none of the judging results can be trusted.
I am no actuary, but based on this one video, you have a contamination rate of 50%. Extrapolate that out to all of the videos and you have, essentially, a completely polluted pool of video results. If there are people going around just amassing “points” on the judges leaderboard by clicking “Workout Is Good,” how can any of the results be trusted? How can, for example, the average performance that bears little scrutiny be assumed to have been judged fairly when one that is extremely high profile clearly has not? Is it fair to say that there is a 50% failure rate on the video judging?
It’s fair to say, I think, that at least 50% of the “Workout Is Good” results are tainted across the board. Which means, in general, we online judges are only watching the first few reps. If they look good, we click “Workout Is Good” and move on.
It’s time to do away with video submissions. I understand the concept of wanting to include people from all over the world who may not have access to an affiliate, but frankly, with 10,000 affiliates and growing, this is much less of a problem than it was when the Open began in 2011.
As a programmer, I can tell you that it’s a fairly simple coding application that would require the entire video to be viewed before it’s voted on. The judge starts watching the video, and the voting buttons do not become clickable until the video has reached its end. Anyone who has tried to hurry his or her way through a mandatory, highly dreadful e-learning module on “document retention” knows that it’s doable. Time to hold CrossFit Judges to the standard of, at the very least, sitting through the entire video before clicking “Workout is Good.”
And Denisse: Nice job.