Note: This is one of many questions I’m attempting to answer by analyzing the 2013 CrossFit Games Open. For more details on the project, check out this post
here. This is the first (of hopefully) many posts on this topic.
Imagine you’re building a super-star team of men for CrossFit. How tall should your candidates be? Or, more specifically, is there a height range that you should immediately rule out?
- Shoot for 5’8″ to 5’11” tall for the “super-team.” If you have to choose between someone a little too short or a little too tall, go for the shorter man. (Sorry, tall people!)
* Crossfit, where the Open is concerned, attracted competitors who are taller than the general population on average by over 1 inch.
- The most competitive Crossfit athletes are on average shorter than the general population by about 2 inches.
The Data: Height & Crossfit Men
To start, here’s a random sample of all men who completed all 5 Open workouts (sample size 14,853):
Data Summary of Men’s Sample from CFG 2013
The average height of all competitors worldwide is about what we’d expect — 179cm (5’10”) and the distribution looks normal.
(The lumpiness in the distribution is due to rounding. Since competitors entered their height in inches and we converted to centimeters, some values don’t show up as often.)
What we do know is data about the height of the
US Population, so we can look at how the US competitors compare to that. The distribution of the subset of competitors in the USA aged 20-29 looks like this:
The average of all competitors is 179cm (5’10.75″). The US average height is 5’9.5″ for men 20-29. Crossfit for the Open is taller than average.
But, a skew starts to emerge towards below average height as we limit our sample to the top performers. Below is all men in the Top 500 in total reps completed for the open in the USA who reported their height in their profile:
Height of Top 500 Men by Total Reps in the CFG 2013 OpenThe Top 500 are a full centimeter shorter than the average.
The Top 50 get even shorter. The average of the top 50 in the USA goes down to 175cm (5’8.75″):
In the Top 50 performers in the Open in the USA, only two are over 6′ tall.
Performance vs. Height
Let’s plot the heights into buckets versus average rank (lower being better). This graph is the all men in the top 40,000 overall:
The average best rankings happen right around the 175-183cm (5’9″-5’11” range). Being a little shorter or a little taller doesn’t seem to hurt so much. But, when you start to push over 6′ tall, things get worse.
(The N/A means the people reported heights outside of our graphing range.)
Here’s another view of height versus average ranks. This one includes a “heat map” and average showing the count of places in the top 40000:
Another view here is the total reps for the entire open versus height in cm.
(Using total reps and not a more advanced statistical method is because of laziness. It’s my belief that any difference is inconsequential given the size of the data set.):</span
Total Reps Completed vs. Height for a Random Sample of ~14k CFG 2013 ParticipantsAn interesting insight here is how tall people vary more around the mean for their height than shorter people (as demonstrated by the width of the gray area around the line and the color distribution).
Being too short for Crossfit is not a killer for someone aspiring to top-tier performance. Being too tall is. As we exceed 6’3″ (190.5 cm), the number of high scores drops off. These would be represented by black square in the heat map in the top right of the graph. Once we exceed 6’4″, the number of elite competitors drops to nothing in our sample
(Sorry, Aja Barto).
But, how much does being too tall or too short hurt a person? My hypothesis is that the relationship is non-linear (i.e. things get progressively worse as people get taller). At the present time, this is beyond my data skills. If we glom on a Simple Linear Regression, it does give us a model (though not a strong one.)
And going up to 180cm:
Men Less than 180cm Tall from a Sample of 14,000 Men of All Ages CFG 2013
Does the same apply for the best of the best CF competitors?
In this sample, we’ve taken the top 15% of performers worldwide from 10-54 in Age. (At 55, the standards changed.)
- It is probable that the skew towards being shorter is more pronounced than even this data shows. Given the data on height is self-reported and the social belief that “taller is better”, many competitors are likely to have exaggerated their height.
- The design of the Open workouts was done to make them as accessible to as many competitors as possible. This, inherently, made the workouts often favor taller people. Many of the exercises that shorter people love like ring dips or handstand push ups did not make appearances in the Open.
- Do our results jive with what we know from the sport? Yes, it would seem to. The movement standards for CrossFit are highly dependent upon a person’s body build. If you’re doing a barbell movement from shoulder to overhead, the bar will need to travel the length of your arm. In addition, it’s well-known that the best movers in Olympic Lifting are those with long torsos and short-limbs. This is a characteristic usually not found in the tallest of people. Additionally, looking at the top competitors in the Games, very few of the best performers are over 6′ tall.