There has been a lot of discussion on social media and in the fantasy football community the past several years about Body Mass Index (BMI) in fantasy football prospects – specifically wide receivers. The analytical data suggests that bigger WRs are more dominant and better fantasy assets. These WRs have a higher likelihood of succeeding in the NFL and therefore yielding more fantasy football points.
This season there is one prospect in particular that has stirred up the ole BMI argument again. The aforementioned player is the 2020 Heisman Winner from Alabama, DeVonta Smith. While it might seem crazy to you that people would be doubting Smith after the season he just had, it would appear that a chink in his armor is his small/slim stature. In other words, his low BMI. While Justin Jefferson, Ceedee Lamb and AJ Brown might not be in Smith’s range of outcomes, I show why he compares well to players like Adam Thielen and Calvin Ridley.
What is BMI?
First, let’s establish what BMI stands for, what it measures, and what it does not. Body Mass Index (BMI) is designed to be a simple way to determine someone’s mass/body fat based on their weight.
Formula:
weight (kg) / [height (m)]2 or 703 * [weight(lbs)]/[height(in)²]
The number that this formula yields is used to determine if someone is underweight, appropriate weight, or obese based on others in their demographic.
That’s essentially it. Weight and height. BMI does not tell you what % of their weight is muscle or fat. It does not tell you how strong the person is. It doesn’t tell you if they are athletic or physically in good shape. You could see where there might be some limitations with this measurement.
How did we get here?
As more focus was put into the metrics in fantasy football, analysts started to find what attributes/production was useful for predicting top prospects. For NFL WRs a high BMI seemed to correlate to fantasy production. A prime example is the sheer number of top-24 WRs the past three seasons who have a large BMI (>26). 
On average ~86% of the WRs that finish in the top-24 have a BMI over 26. It’s easy to see why this higher BMI craze has caught on.
However, when you look at the sheer volume of players in the NFL that actually have a BMI >26 over that same time period, one could also infer that the sample is skewed.
*Data includes WRs who produced any stats during their respective season.
When I limited those numbers to just fantasy-relevant players (WRs who finished in the top-100) the data set was slightly more skewed.
This data suggests that the league is saturated with WRs who have a BMI greater than 26. Roughly 30% of fantasy-relevant WRs have a BMI under 26. Forecasting that we should see that the sample to top-24 WRs would represent the whole population. Although not exact, nearly 20% of WRs who finish in the top-24 have a BMI under 26. The fantasy football community should be finding what these “smaller” WRs have in common and what allows them to succeed in the NFL.
Additionally, anyone in the analytic fantasy football community would tell you that one single measurement or data point is not great at predicting future prospects’ success in the NFL, but rather multiple factors should be taken into consideration. With that said, some data points are better than others. Peter Howard’s (@pahowdy on Twitter) database shows that BMI is terrible at pending future WR success, with r² values of 0.002 for both average expected point in year one and average PPR points in years 1-3 (meaning there is virtually no correlation between BMI and fantasy production).
For anyone interested in more of Howard’s work, I highly recommend you check out his Patreon. It’s only $1 a month and I can guarantee no one is giving you the amount of information for less money than @pahowdy.
Bigger = Durable Right?
If BMI is not really good at indicating which WRs will succeed at the next level and the majority of WRs with a BMI over 26 surely it’s good for something? Maybe durability. Are these bigger players more durable? Short answer: sort of.
To look at the effect BMI has on injuries I pulled in Undroppables math/analytic’s expert @BpoFSU – Brian O’Connell. We looked at wide receivers since 2018 that have had at least a 20% target share, to weed out irrelevant WRs that would have skewed the data.
(Nerd alert. These next few paragraphs are about the specifics of our findings and for full transparency I included them. If you just care about the results, scroll to “Results”).
Players were either injured or not, giving me binomial values of 0 or 1. To figure if BMI was related to injuries, I used logistic regression which allowed me to discover trends in the data. The two main outputs of logistic regression included the odds ratio (listed as “Odds” in the data set) and the Wald statistic with a corresponding p-value. The odds ratio tells us the increase or decrease in the probability of injury due to a one unit increase in BMI. For example, in our data set the odds ratio for any injury is 1.073. This means for every 1 unit increase in BMI, the likelihood of a receiver being injured increases by 1.073. If the odds ratio is below 1 then the probability will decrease as BMI increases. As for the Wald statistic, think of that as logistic regression’s version of a z-score, which allows us to find the p-value for our regression. A p-value is the probability that this data is due to random chance so the lower the p-value, the better. We decided for this study that the significance level would be 0.10 and any values that are lower than that would be considered statistically significant.
Results: All Wide Receivers (since 2018)
The likelihood of all types of injuries increases with a larger BMI by 1.073 per one unit increase of BMI. While injuries overall were not statistically significant, skeletal injuries (such as bone bruises and fractures were (p = 0.073). They increased by 1.351 suggesting that larger BMIs lead to more skeletal injuries.
Results: Rookies Only (since 2018)
This sample is significantly smaller (n = 58) but still gave us some good results.This trend is not even close to being statistically significant (p = 0.55). This indicates there are other factors that contribute to injuries besides BMI. The trend of lower BMI leading to a higher probability of injuries is also the case for soft tissue injuries and concussions but again they are also not significant. Skeletal issues and their relation to BMI in rookies have a very similar odds ratio to the whole sample size mentioned earlier. While the p-value is not statistically significant, this is something that might warrant further investigation as there are hints to a trend with skeletal injuries and BMI.
Focusing on DeVonta Smith, this data suggests that Smith has a 4.64% chance of suffering a skeletal injury over the first three years of his career and a 2.13% chance of that type of injury during his rookie season. On average the time missed for a skeletal injury in the NFL was 5.7 games, slightly larger than the 5.2 games for soft tissue injuries.
Analytical Data
I mentioned before about focusing on attributes of successful WRs with sub-26 BMIs. I went back as far as 2016 to find the WRs who had a BMI under 26 and were able to finish the season as a WR1 or WR2. Next to their name is the rate that they finished as a WR1 or WR2 over their career in PPR leagues:
- Adam Thielen – WR1: 43%
- Marvin Jones – WR1: 11%, WR2: 11% (total: 22%)
- Robby Anderson – WR2: 40%
- DJ Chark – WR2: 33.3%
- John Brown – WR2: 14%
- Calvin Ridley – WR1: 33.%, WR2 33.% (total: 66.6%)
- Emmanuel Sanders – WR1: 9%, WR2: 27% (total: 36%)
- AJ Green – WR1: 40%, WR2: 20% (total: 60%)
- Tyrell Williams – WR: 20%
This is a list of nine players. A small list yes, but as I pointed out before, there are not many WRs in the league with BMIs under 26.
In an effort to try and forecast DeVonta Smith’s chances of success I analyzed what these WRs had in common in college along with measurements taken at the NFL combine. I used seven metrics: 40 Yard Dash Time, SPARQ-x score (a measure of athleticism via Playerprofiler.com), Breakout Age (BOA), College Yards Per Reception (YPR), College Dominator Rating, College Target Share, and Best College Yardage Share. Below are the averages for this group:
Every WR listed above met at least 4/7 criteria with Calvin Ridley and AJ Green being the only two that were a perfect 7/7. One could argue that the closer you get to hitting those seven marks the better fantasy asset Ridley is currently a converted dynasty fantasy football WR and AJ was a top-5 play at his position for most the last decade.
DeVonta Smith Profile
Using those seven metrics above, we can try and forecast DeVonta’s Smith success rate. Without a combine/pro day score, we won’t have a 40 time or SPARQ-x score but Smith has 4/5 college metrics with the exception of BOA.
Another thing people are worried about with DeVonta Smith is his ability to handle press coverage. Because he is small, analysts are concerned that DBs will be able to jam Smith at the line of scrimmage and disrupt his routes. For this issue I turned to The Undroppables very own film guru, @2on1FFB – Tommy Mo to look at how DeVonta handled press coverage at Alabama.
Tommy’s video breakdown of DeVonta Smith vs. Press Coverage showed that in college teams were not pressing Smith very much. The reasoning behind this was because DBs were not very good at it when it came to facing off with Smith. Most college DBs anticipated getting beat by Smith or were afraid of his speed. Even when they lined up close to the line of scrimmage against Smith, they quickly retreated. The few times defenders did try to get hands on Smith, he used his feet, route-running skills and speed to separate. This is likely the rationale behind coordinators electing not to press him. NFL DBs are definitely more skilled and will likely try to jam Smith, but in college at least these attempts proved futile.
Summary
There are holes in DeVonta’s Smith’s profile. He does not check all the boxes. He had a late breakout age, and we will have to wait and see what his final numbers look like when he weighs in and runs at his pro-day. I understand why the skeptics are worried. Smith is an outlier, and we typically try to avoid those guys in fantasy football. But as I pointed out, the NFL is saturated with larger WRs which bias the data. My research shows that there is not a correlation between smaller BMIs, and fantasy production. If anything bigger WRs have a propensity to skeletal type injuries. There are success stories for WRs with a sub-26 BMI and Smith’s college profile fits that mold. If you have a top-5 rookie pick in your dynasty leagues this might not be the player I would select. However, getting the next Thielen, Ridley, or Sanders is a valuable cornerstone for dynasty rosters.
So does BMI really matter? Kind of. But that title isn’t as catchy.
Special Thank You
I would be remiss if I did not thank the people who helped me with this content. A special thanks to The Undroppables Team – specifically @2on1FFB for the film breakdown and @BpoFSU for the countless hours sifting through data and helping me incorporate it into this piece. I also want to thank Fantasy Football Astronauts for the film study resource that came in handy along with PlayerProfiler and @pahowdy for a lot of the data used in this article.
Stay tuned these next couple of weeks for The Undroppables Rookie Profiles which are coming soon!
Link to data used: WR BMI Data
