To curve or not to curve? I was at a Nats game a few weeks ago and I noticed that a lot of players don’t curve the brim of their hat anymore. This seemed especially true among the younger players.
I made a mental note to revisit the issue later.
And well, here we are.
Obviously I don’t want want to get too anecdotal, so I’ve gone through a few 40-man rosters and recorded the curves of players’ hats. I did this using a Chrome addon called Page Ruler that lets you draw rectangles on your screen and count the number of pixels on the x- and y-axes.
Using portraits from the MLB website, I counted pixels for three teams: the Nationals, the Orioles, and The Yankees. I chose these teams because I was an Orioles fan before the Nats stadium was built, and the Yankees because I wanted a third team and they were the first team to come to mind.
But wait! Some players seem to pull their heads back in their portraits, which makes their brim appear less curved than it is. (Imagine if you were looking straight up: my curve measuring method would return a 0px value.)
To correct for this, I measured the height of the logo on each player’s hat. At least one player per team was looking straight at the camera, so I used their value to represent the “true” height of the logo. Dividing all other players’ logo heights by this “true” height let me estimate the percent-of-head-incline for each player. I then divided the raw brim-curve value by this percent to get an adjusted brim-curve value. Now all players’ brim-curves are on an even playing field. Great success!
With this adjusted value, I used a regression model to predict brim-curve with player age. Lo and behold, player age is a statistically significant predictor of brim-curve (p=.012; Cohen’s d=.45).
So now we know: the older a player gets, the more curved their hat.
Extrapolate this effect and we get…
Thanks for reading!
Want more data visualization? Check out my other posts at: https://vizthis.wordpress.com/