Dave Castro tried to kill me. The weapon used? 14.3 with a 14.5 chaser.
Part 1: 14.1 and 14.2
This year’s CrossFit Open was pretty crazy. Not because of the typical burning lungs, screaming quads or shredded hands. Those are everyday concerns in the world of “better today than yesterday”. No, my crazy barometer for this year’s Open was measuring the massive amount of discontent that I heard from CrossFit athletes of every skill, scale and level. From the opening bell of double-unders – literally followed by the statement “really, seriously – double-unders” by just about everyone I know – to the hellaciously nasty 14.5 torture fest (I think I started seeing leprechauns halfway through the set of 15 thrusters) the negative and unhappy commentary I heard at my gym and across the world from other friends in the sport got me to wondering.
Mathematically, how bad was the Open in 2014? More importantly, could the suck be quantified in a way that might give the community pointers on one of many possible explanations? Was it bad programming? Was there an actual effort made to “weed out” a large part of the field right out of the gate? Is CrossFit destined to become Cirque Du Soleil, where only those with ridiculously honed skills stand an iota of a chance of placing in their age, gender, region, etc.? Or, conversely could it be that a very large part of the CrossFit population is just weak? We have to consider all options – even if it means that we might have to admit that we’re not as collectively bad-ass as we think we are. Well, as you think you are – I feel pretty bad ass – in general; if I belt up, and have my Strength Wraps on, have 45 minutes to stretch and then if there aren’t any stupid muscle-ups in the workout. (Yes, I’ve seen the Darnell Sanders video – like 280 times – because it is awesome).
So what does the math say? Actually, a class by class analysis of individual scores was enlightening. I want to make a disclaimer and say that there are no compelling conclusions in these numbers. No smoking gun. But, there is certainly a lot to think about – in particular, I think the figures indicate that CrossFit headquarters might want to be thinking about whether they actually are living up to their marketing related to the “community” aspect of this sport.
First, I had to make personal assumptions about what constituted an “acceptable” performance in each of the Open workouts. These are my assumptions and mine alone. You might completely disagree with them and I can respect that. I actually feel that I set the bar low/conservative for each event (except for 14.5, which was a bear-mother to try to figure out a decent threshold for) and that the assumptions used are supportable. Second, there is definitely the potential for data inaccuracy, but not to a significant enough degree to invalidate the directional results. The reasons for this inaccuracy is that you can’t download the results easily from the Open site. It took me several days to crunch numbers that would have taken 5 minutes with an Excel dump and pivot tables.
Let’s put the “community” aura to the test first. I heard a lot of hype about the number of registered athletes this year many times on the workout broadcasts. 170,000 was a number that kept getting tossed around. Interestingly, the number of recorded scores in 14.1 was actually more than 214,000. The number of posted scores from 14.1 to 14.5 might be a good place to start if we wanted to attempt to measure what “community” looks like. We had 214,000 people pay to get into the Open and we have to assume that they had not only a financial but a personal interest in registering. So if we analyze the trend from 14.1 to 14.5 in terms of how many people actually participated – this could be an indication of the level of community support for the event, the programming and the perceptions of the worldwide body of CrossFit athletes at all levels. How big was the delta over time? Well, I think it is a troubling sign that more than 1 in every 4 athletes didn’t finish the Open.
Something else that jumps off the axis in this set of numbers is the trends related to which workouts inspired athletes to give it a go and which ones clearly had them running for the door (or not running in the door at the gym). The downward trends are certainly suggesting that 14.2, 14.4 and by a huge margin, 14.5 were very unpopular pieces of programming. One striking oddity is that apparently female competitors in the Master’s Women over 60 class were really pumped up about 14.3. You go ladies!
What about each of the individual workouts?
It seems fairly clear that the great gnashing of teeth related to double-unders was actually a lot of whining. Sorry, told you earlier that we need to hold ourselves accountable as well. The assumption for 14.1 is that the mark of reasonable performance was very conservative – that being the completing of the first 30 double-unders. 14.1 was a couplet of 30 double-unders into 15 light power snatches (75, 55, 65, 45). The vast majority of heartburn that I heard about this workout was that the double-under skill was an insurmountable task for many athletes but the number of individuals who did not move in to the snatches was actually very small. With the exception of Master’s Women over the age of 60, anywhere from 9 out of 10 to nearly 10 out of 10 (Individual Men) were able to finish that first set of double-unders.
So we came out of the gates like rolling thunder for 14.1! 214,000 athletes with a better than 95% completion rate of the dreaded double-unders. Unfortunately, the party started to get a whole lot less fun in 14.2. The second workout was a time-bonus WOD starting with a 3 minute couplet of Overhead Squats into Chest-to-Bar pull-ups. The performance assumption for this workout is pretty simple; getting into the second 3 minute round was the measure of success. So how many CrossFit athletes made that first cut? Way fewer than I think anyone would have expected – but more importantly, the workout seems to give the first indications of a fundamental programming problem.
The failure rate for 14.2 was nothing short of mind-blowing. Only Individual Males (athletes under 40) came close to a 10% failure rate, moving into at least the second 3 minute round. But the failure rate for all other male athletes up to the first scaled age group (Master’s Men 55 to 59) escalated rapidly to over 20% of all scored athletes. The story for female athletes is really eye-opening though. Not only did every age category of females in the competition clock in at a failure rate ranging between 47% and 68%; female performance was nearly always 4 times worse by category than the same male category. The red lines represent that gap. And lets not mince words on what this means – the confidence interval in the numbers clearly shows that something was consistently and programmatically wrong with this workout. If the progressive increase in failure between each male age group was less than double the next younger group – there is no other explanation for the 400% greater failure rate among female athletes. The female-related scaling was completely wrong and it had an enormous impact beyond percentages. We’re talking individual human beings here, this many:
I’ll dig further into 14.3, 14.4 and the big whopper – 14.5, in my next posting. But I want to leave my readers with an important note. This is mathematical evaluation and not a personal attack. When I use the terms “failure rate” and “washout” it isn’t an evaluation of any single person’s performance. In fact, 14.1 was a disaster for me personally – it took me three times as long to do my notoriously unlinked double-unders as it did to do the power snatches. I just want to dig a bit beyond the sweat, blisters and dry heaves to determine what we can do better in CrossFit – better today than yesterday, right?