Tuesday, January 28, 2014

Does MLS have a bias against short players?



Edit (12:10 am pacific, 1/29/14): Additional graphs added at bottom to show USborn players in MLS relative to US population and also Netherlands-born players in Eredivisie relative to population.

This post was inspired by a story about how Enrique Cardenas felt slighted by not being drafted into MLS speculating that he was too small. There was (as always) some decent discussion on the topic in reddit.com/r/mls that discussed whether smaller, technical players are squeezed out by MLS. MLS Analyst, Matt Doyle, is someone who I generally agree with his opinions and analysis, but our initial reactions about this topic differed – he felt that small players got a fair shake, while I felt they didn’t.

MLS PLAYERS ARE RELATIVELY TALLER THAN MALES AGED 20-29 IN THE US

It seems pretty obvious just by looking at the players on the field that they tend to be a bit bigger than the average guy. As an economist/statistician, the eye-ball test usually doesn’t satisfy me. The chart below shows the percentage of people at a given number of inches for both MLS players (listed as “active” on mlssoccer.com as of 2pm eastern 01/28/2014) and the US male population aged 20-29. It is absolutely clear that MLS players are taller than the general population. 


The height at which MLS players overtake the general population in terms of percentage of the total sample is from 5’11 and 6’0. My mind immediately thinks there’s two potential factors at play in this particular break. The first, is that MLS teams might fudge a player’s height by a tiny bit to list some players as 6’0 when they are really just under 5’11.5.  The second, is that MLS scouts have a mental propensity to think of guys under 6’ as “small” in the same way that people tend to think something priced $4.99 seems significantly cheaper than if it were priced $5.00.

NCAA PLAYERS ARE RELATIVELY TALLER THAN MALES AGED 20-29 IN THE US

Reddit user "wacksoff" suggested that the small players get weeded out earlier in the US development system than MLS. I looked at the height of all players for teams in the final 32 of the 2013 NCAA division 1 tournament and compared them in the same manner as I did MLS players in the above section. We find the same general trend of NCAA players being taller than the general population and there is again a big difference between 5’11 and 6’0. So, it seems pretty clear that the smaller players do tend to drop out before they get a shot at MLS.


MLS PLAYERS ARE RELATIVELY TALLER THAN NCAA PLAYERS

Alright, so we’ve established that MLS isn’t completely at fault for height bias. Do they share at least some of the blame? We can’t draw an official conclusion based on the data I used. This is largely because many players continue to grow into their early 20’s, so surely some of the shift from shorter players to taller players is simply due to the physical growth of those individual players. However, we can’t say for certain whether the entire shift is due to that or if MLS does indeed tend to have a bias in favor of taller players.



PLAYERS UNDER 5’9 ARE DISCOUNTED HEAVILY IN MLS

So how screwed are small players by their size? For this section I borrowed something from economics called a location quotient that is most often used to identify industries of importance for an area and is also very useful for studying demographics. It is a ratio of ratios that basically measures the density of one category in a particular area to the density of the same category in a different area. So, what we are measuring here is how much more (or less) dense players of a certain height are relative to the total population. Values less than one say that the density is lower and values greater than one say the density is higher. Players that are between 5’9 and 5’11 are pretty close to having similar densities within MLS and the general population, which means they don’t get any different treatment based on their size. Players that are 5’7 or 5’8 are at a massive disadvantage because of their height, with ratios of 0.35 and 0.61 respectively. That means a player who is 5’7 is about 3 times less likely to be an MLS player than would be expected based on the distribution of the overall population.



IS THIS A BAD THING?

This is a question impossible to answer based on the data I’ve used. Assuming technical and tactical ability are distributed normally across the US population, it makes some sense that there’s some bias toward bigger players since they should be able to simply do more things like push small players off the ball and win more headers. The extent of that bias is the real meat of the problem. Personally, I think the US soccer system discounts smaller players too heavily and instead focuses on physical attributes rather than soccer ability. This is based on my personal experience as a competitive boys U14 coach and women’s university coach. I plan to further explore the idea of US soccer discounting small players in a post to follow.

Note: US height data from the US Center for National Health Statistics

Additional interesting tidbits related to the topic:
  • There was almost no correlation between average team height and points earned.
  • The tallest team was Columbus at 71.96" on average
  • The shortest team was Philadelphia at 70.41"on average
  • Keepers were the tallest on average at 73.94" followed by defenders (72.13"), forwards (71.57"), and midfielders (70.01")
The data for "additional interesting tidbits" are for players with 200+ minutes played in 2013.

COMPARISON TO NETHERLANDS

A couple people on reddit wanted to see some comparisons to other leagues around the world. I only had time to do one tonight (finding and entering this data is time consuming).

This first graph shows MLS players who were born in the US compared to the US male population aged 20-29. Clearly, the graph shows that MLS players tend to be taller than the population, which is to be expected from our findings above. The second graph is Eredivisie players who were born in The Netherlands compared to the Netherlands male population aged 20-30. We find that Eredivisie players are actually shorter than the population, in stark contrast to in the US. I will be looking at more leagues as I have time, but this small sample size supports the hypothesis that small players are at a disadvantage in the US compared to large players in the US and when compared to small players in other countries (so far only proven true for Holland). This does not tell us when exactly they get pushed out of the game, but the NCAA data above suggests it is largely sometime in the NCAA recruiting process, high school, or even earlier.




No comments:

Post a Comment