One of the things I like to do after a major competition is look at how much each event correlated to other events and the final standings. It helps draw some conclusions on whether the programming was biased towards any of the three modalities (monostructural, gymnastics, weightlifting).
What you will typically see is that a couple events are highly correlated to the final standings (I generally use anything greater than 0.70), a bunch in the 0.40 to 0.70 range and then the occasional less than 0.40. Almost never is there a negative correlation; that is, where the event results are relatively opposite of the final standings.
There are, however, quite a few negative correlations between events. For example, the women who finished in the top 10 overall had a -0.71 correlation between the Inverted Medley and Ski-Bag events, which means that athletes who did well on Inverted Medley were unlikely to do well in Ski-Bag.
As always I looked at the correlation data for the entire field this year. However, I also broke out several groups to see how, for example, the top 10 overall correlated to the events. So if you were a top 10 athlete, which event, or events, best predicted where you fell from Fittest on Earth to 10th overall.
I also looked at the Top 20 since they completed all 12 events. Then I also evaluated the Bottom 20 (or 19 in the men’s case) to see if any conclusions could be drawn by those who failed to make it to Sunday.
Here are some conclusions I’ve drawn from this year’s correlation analysis…
Top 3 Correlated Events to Overall Finish
When I run the correlation analysis, I do one for the men and one for the women. I have found that the correlations can be different between the two.
However, this year it wasn’t that different.
Helena, Intervals and Muscle-Up Logs were the most correlated events to the overall standings. They weren’t in the exact same order, but it was close.
With Helena touted as the middle of the “CrossFit” Venn diagram, it was good to see that it was highly correlated to the final standings. I was a bit surprised to see Intervals so highly correlated, but given the demand for speed on the rower and power/speed on the burpee and super-high box jump overs it does make some sense.
And then when you combine the box “get-overs” with the log get-overs on Muscle-Up Logs it starts to make more sense.
Olympic Total was the Least Correlated
For both the men and women, the Olympic Total event was the least correlated of any event to that of the final standings. There were only two events with a barbell this year – Olympic Total and Echo Thruster Final. And while there were two single modality monostructural events, Bike and Cross-Country 5K, there was only one single modality weightlifting event, Olympic Total.
Any other strength-based event was paired with gymnastics or monostructural or both. Ski-Bag had heavy squats and the SkiErg. Pig Chipper sandwiched a ton of gymnastics between the Pig flips. The Alpaca Redux added back the original seated legless rope climbs.
The table below highlights events with where either the men or women had an event correlation lower than 0.4.
|Echo Thruster Final||0.27||0.52|
The question I have from this is whether there should have been more barbells programmed into the Games this year. Obviously fitness is more than cycling a barbell, but CrossFit has always leaned heavily that direction.
In the 16 workouts programmed between the Open and Semifinals this year, 10 had a barbell. That’s 62.5% of all workouts featured a barbell.
The Games, however, had only 2 barbell events, one of which being a max lift. Could this be a trend in Adrian Bozman’s programming or will there be more barbells in the future at the Games?
Top 10 and the Olympic Total
I just said the Olympic Total was the least correlated event to the final standings of the entire field, but when you zoom into just the Top 10 athletes the story is much, much different.
So different, in fact, that it was the highest correlated event to the Top 10 standings in the men’s field. The Olympic Total had a 0.73 correlation to the Top 10 standings for the men and a 0.53 correlation for the Top 10 women.
How does one interpret this?
Well, the overall correlation is low for the Olympic Total because five of the six top finishes in the event were 19th or lower. Jeff Adler was the only athlete in the Top 10 to finish 6th or better in that event.
It makes sense though because Jack Farlow, Bronislaw Olenkowicz, Nick Mathew, Moritz Fiebig and Fabian Beneito struggled throughout the weekend, but this event was in their wheelhouse.
And when you look at the Top 10 men, the stronger you are the better you were at the Olympic Total, but also other events that required top-end strength and power.
Running is Not the Same as Biking
There were two single modality monostructural events this year, a bike “Ride” and a “Cross-Country 5K” run. While both were monostructural events, the results were not the same.
The correlation between these two events was 0.42 for the women and 0.30 for the men. Both were slightly correlated, but maybe not to the level many would have expected.
Just a quick observation, but it makes sense. Biking (and swimming) require some additional skill than that of just running…something that is trained more often than the other two.
Did both test a similar pathway, yes, but the biking required some additional skills to excel.
In general, when looking at the correlation of events to final standings this year there are not any two events that are highly correlated with each other. I interpret this to show that no two events were so similar that the same athletes did well in both or poorly in both.
The most correlated events spanned “CrossFit” and required the ability to do all three modalities well. It showed that those who were well-rounded faired better overall.
Barbell specialists had to utilize other skills to move up the leaderboard this year. With only two events if a Games athlete hasn’t been training odd objects or different ways to lift or move weight, they should start.
Overall I think the 12 events worked well together and the correlation analysis showed this.