Is there an Ideal BMI for Performance?

bmi-cover-5*Original 📷: @prestonsmithphotography

Body Mass Index, or BMI, is an almost two hundred year old concept that roughly establishes a ratio between a person’s height and body weight.  Historically considered an important health marker, current research raises questions about the efficacy of BMI as a health marker- most famously when pro boxing superstar Evander Holyfield was found to be technically “obese” according to BMI metrics.  Most glaringly BMI focuses solely on scale weight, neglecting body composition- the amount of muscle vs. fat present in a body.
fbdbd940d1606045ee2f10abc66eeb45
BMI by itself may not be a great predictor of health, but that’s not to say it’s altogether useless, it simply needs to be examined in relation to other metrics.  This article is not going to discuss the pros or cons of using BMI as a proxy for Body Fat percentage or health status. Instead, we’re interested in looking at how BMI relates to performance. Since BMI is just a ratio between weight and height for any given height, a higher BMI means a higher weight. Since we can’t do much to change our height, the practical variable of interest for us is weight.

In terms of performance in CrossFit, is there such a thing as being too heavy or too light for your height? As a coach, I have long suspected that such a dynamic exists. I’ve coached my brother Jeremy Kinnick (4x individual Games athlete) since 2007 and we’ve experimented with him competing at various weights, in search of the optimal weight for his frame. Will Jeremy, who is always going to be 5’10” (sorry bro), perform better at 185, 200, or even 215 lbs?

Given that it’s due to an increase in muscle mass, it’s assumed that being heavier should allow for more absolute strength development (look at the difference in totals between the lighter and heavier Olympic Weightlifting weight classes, for instance). While absolute strength is desirable, in CrossFit a higher body weight can be a disadvantage in many movements, e.g. pull-ups, muscle-ups, rope climbs, box jumps, handstand push-ups, running, etc.

We decided to dig into the BTWB database and see if we could use our wealth of accumulated numbers and statistics to shed light on the situation. To accomplish this we compiled the weigh-ins of over 75,000 BTWB users who have been using our service for an average of 13 months as well as those users scores on our proprietary metric of well-rounded fitness- the BTWB Fitness Level.  Comparing those data points with national data from the CDC and JAMA yielded some interesting correlations:

Highlights:

  • Average BMI for top Male BTWB users is 27.2 with a 33.5″ waist
  • Male BTWB users with a BMI of 26-28 had substantially higher average Fitness Levels than other ranges
  • Average BMI for top Female BTWB users is 24.0 with a 30.0″ waist
  • Female BTWB users with a BMI of 22-26 had substantially higher average Fitness Levels than other ranges

Quick Primer on BMI Ranges

According to the NIH, “BMI is a useful measure of overweight and obesity. BMI is an estimate of body fat and a good gauge of your risk for diseases that can occur with more body fat.” The BMI ranges are:

  • Underweight: Under 18.5
  • Normal: 18.5-24.9
  • Overweight: 25-29.9
  • Obese: 30.0 and Above

The NIH also notes that BMI “may overestimate body fat in athletes and others who have a muscular build.” This is particularly relevant for our analysis, since our subjects will likely be more muscular than the national average. It’s not surprising that we find that the higher BMI of our top male users are not due primarily to higher body fat.

We’re going to take a look at the BMI dynamics for men first, and then we’ll analyze the women. We’ll see that the BMI trends for men and women differ dramatically, especially when we look at the groups with high Fitness Levels.

Men

We’ll start our analysis by comparing the average BMI of all US males to our sample of BTWB users, separated by age group. This will give us some baseline comparisons between the two groups.

bmi1

For the 20-29 age group, the BMIs are very similar but when we look at the 30+ age groups, we start to see some significant differences. The average BMI for US males age 30-59 is 29, while the average for BTWB users in the same range is 27.5. Overall, male BTWB users have lower average BMIs than US males. This will become interesting in a moment when we find out that there is a much higher percentage of “Overweight” male BTWB users compared to US males.

BMI2.png

Looking at waist size (another factor relating to body fat), we see that BTWB users have much slimmer waists than the general population. Both groups experience increasing waistlines as they age, but the increase is much milder for the BTWB group. So far we’ve seen that male BTWB users have generally lower BMI and smaller waist sizes compared to all US males.

BMI13.png

Here’s where it gets interesting. Among male BTWB users, we see a large concentration of BMIs in the “Overweight” range. This trend holds for BTWB men in the 20-39 age group as well as the 40-59 age group (next graphic). And if we look at Elite male BTWB users (those with a Fitness Level of 85 or higher), we see a remarkable concentration of BMIs in the “Overweight” range (over 80% of the group).

This suggests that for men, the “fitter” you become, the more likely you are to be in the “Overweight” BMI range. Very few men with very high Fitness Levels are in the “Normal Weight” (9.8%) or “Obese” (7.7%) categories .

In sum, while BTWB men have lower average BMIs overall, these BMIs are much more concentrated in the “Overweight” range compared to the general population.

BMI12.png

Even though male BTWB users are more likely to be in the “Overweight” range, this does not mean that they necessarily have excess body fat. As noted above, BMI is a poor proxy for body fat in those who are more athletic and muscular (BTWB users). Based on the small waist sizes of the BTWB users it is likely that the increase in BMI is due to increased muscle mass, rather than body fat. This concentration in the “Overweight” range suggests that you need quite a bit of muscle mass to excel in the areas tested in CrossFit.

Now let’s look at the situation from a different angle. We know the Fitness Levels of the BTWB users and we can compare BMI and waist size to Fitness Level among these users and see what relationships we find there. To start, we’ll group all male BTWB users by BMI category and see what the average Fitness Level is for men in those categories.

BMI7.png

We see that those in the “Overweight” BMI range have a substantially higher average Fitness Level than those in the “Normal” or “Obese” ranges. It appears that being “Overweight” confers a strong performance advantage for men.

The standard BMI ranges are pretty broad, so for more clarity we can zoom in closer and see if there is an “optimal” BMI range for performance.

BMI8.png

When we graph the data using these smaller buckets we see a nice smooth distribution that peaks in the 26-28 range. For men, being outside of the 26-30 BMI range carries a high cost in terms of performance.

To take the analysis even deeper, we looked at the average BMI of male BTWB users with a Fitness Level of 85 or higher and found it to be 27.2, which is consistent with the optimal range found in the graphic above. This group also had an average waist size of 33.5″.

BMI5.png

If we look at average Fitness Level by waist size we see the optimal range is between 30-34 inches. So although an “Overweight” BMI appears to confer a performance advantage, a waist larger than 34 inches does not. If one was to build a theoretically optimal CrossFit athlete it seems it would be a well-muscled man with a relatively slim waist.

Women

Turning our attention to women we see very different BMI trends when it comes to performance. We’re going to walk through a similar analysis to what we did for the men but the conclusions will not be the same.

BMI3.png

For the women we see a lot larger disparity between average BMIs. Across all age groups the BTWB women have much lower average BMIs. This difference is much more than we saw between the male groups.

BMI4.png

We also see that female BTWB users have substantially smaller average waists than the US as a whole. Both groups show waistlines increasing with age.

BMI11.png

Unlike the trend we saw for males, female BTWB users have their BMIs concentrated in the “Normal Weight” range. Females with high Fitness Levels are even more concentrated in the “Normal Weight” range. For female BTWB users in the 20-39 age group there is a very low percentage of Obesity (~10%), compared to over 30% in the US as a whole for that group.

BMI14.png

A similar trend can be seen in females aged 40-59. Over 70% of female BTWB users with a Fitness Level of 60 or higher are in the “Normal Weight” range. For the older US women, we see 13.2% fewer in the “Normal Weight” range compared the the 20-39 group. Only 13% of all female BTWB users in the 40-59 age group are in the “Obese” range, compared to 40% for the general population.

BMI9.png

When we look at the average Fitness Levels of female BTWB users from each BMI range we see that “Normal Weight” users have a slightly higher average Fitness Level than “Overweight” users. The “Obese” users have a substantially worse average Fitness Level.

BMI10.png

Dividing users into smaller buckets we see that females with a BMI of 22-26 had substantially higher average Fitness Levels than the other ranges. This gives us a lower ideal BMI range for performance compared to the Men (their optimal range was 26-30).

We also looked at the average BMI of female BTWB users with a Fitness Level of 85 or higher and found it to be 24.0, which is consistent with the optimal range found in the graphic above. This group also had an average waist size of 30.0″.

BMI6.png

Looking at waist size, we see the the highest average Fitness Levels were from women with a waist size of 26-30 inches. Women with waist sizes outside of that range had substantially lower average Fitness Levels.

Concluding Thoughts

Being in the “Overweight” BMI range confers a clear performance advantage for men, but not for women. The upper end of the “Normal weight” range appears to be better for performance for women. This is a pretty surprising result considering the average BMIs for men and women in the US are very similar. This suggests the possibility that BMIs are not gender neutral, and calls into question whether “healthy” BMI ranges should differ based on gender (currently the ranges are the same for men and women).

If you are trying to be as effective as possible in CrossFit, it might be worth your while to see how far your current BMI is from these targets. You can calculate your own BMI here: 

This can be a tool for Coaches looking to help their athletes decide if they need to bulk up or slim down in order to be more competitive. It’s not the only consideration, but it could be a helpful piece of information.

The table below was constructed using the average BMIs of the top BTWB users. It’s a quick way to see if you are heavier or lighter than the top performers.

Screen Shot 2016-07-05 at 11.44.05 PM.png

This analysis and information could also be potentially valuable for women who are struggling with body image issues but are underweight. The “ideal” weight for a 5’2″ female athlete would be 131 lbs. But for a 5’9″ female athlete it would be closer to 163 lbs. Conversely, if a 5’9″ female athlete only weighed 135 lbs (a BMI of 20), that could have a negative impact on their performance potential. The understanding that being too thin can hurt your performance can shift the focus away from fulfilling skewed societal ideals/expectations and towards a more healthy outlook.

*The data for “All US” comes from the CDC and JAMA. The data for BTWB users comes from the user-entered data in the BTWB database from 2009-2015. There were not enough level 85+ athletes in the 40-59 age group for the statistical confidence we wanted, so we used level 60+ for the 40-59 age group analysis to increase the sample size for that subgroup. All differences highlighted are statistically significant with a p-value < .0001 at a 95% CL.

About BTWB

“I started using Beyond The Whiteboard just a month or two after I started CrossFit and I have logged nearly every workout I’ve completed over the past 5 years. It has helped me tremendously to monitor my progress over the years and to set new goals. Every time a benchmark workout comes up, I go straight to Beyond The Whiteboard to see how I did last time and what notes I made about my strategy.” -Julie Foucher

“There isn’t another program out there that can track and provide information about your fitness like beyond the whiteboard.” -Dan Bailey

Join thousands of CrossFitters who are tracking their fitness journey with btwb. Start your 30 Day Free Trial.

Our Other Data Based Research

Jonathan Kinnick is a Co-Founder of BTWB as well as the owner of Crossfit Kinnick. He’s a Board Member on the CrossFit Trainer (CCFT) Certification Board and a CrossFit CF-L3 Trainer. He’s also completed the CrossFit Coaches Prep, CrossFit Competitor’s, CrossFit Olympic Lifting, CrossFit Endurance, CrossFit Mobility and CrossFit Nutrition courses. He is also a USAW Sports Performance Coach.

2 comments

  1. Scott Corbett · January 26

    That is a very cool and intriguing analysis. Another potential take away from this analysis is that the NIH has the bands calibrated incorrectly for men in regards to labels.

    Like

  2. Amanda · January 26

    I understand that height was not focused on since athletes cannot change their height but can change their weight; however, I am interested to know how bmi averages changed across height. I have heard that individuals who are either very tall or very short tend to fall outside “normal” bmi ranges vs individuals who are more average height. Was there evidence of this in your analysis? Along the same lines, are their different bmi “sweetspots” for athletes in different height ranges? I think it would be very interesting to see if the results were consistent across different heights or if tall/short people should have different goals.

    Like

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s