Game Theory

Creating the best softball team with heatmapping

Don’t be a goofus

I tend to play a fair amount of softball in the summer. For those unfamiliar, it’s essentially the same thing as baseball except its slow pitch, uses a larger and heavier ball, and is mainly used as a medium for the players to hang out and drink beer. You see, unlike other sports, softball is one of the few games that can be played by the less physically in shape and still be considered athletic. It’s a great time!

With this in mind, there’s a lot of downtime. If you’re in the outfield, it’s a fair amount of waiting until a pop fly heads your way. At that moment, the player gets in a five-second sprint. When that’s over, it’s back to standing around again. On the other hand, if you’re in the infield, it’s similar except instead of a five-second sprint, it’s more of a two-second moment of panic (while protecting your face) and getting rid of the ball as quickly as possible before it becomes something you have to deal with. At the end of the day, it’s just an elaborate game of Hot Potato.

Myself, I’m pretty small and fast so when I’m standing around, it’s usually because somebody put me left-center—a popular area for fly balls. And between my sprinting, I began thinking: based on the average flight pattern of a fly ball verse the positioning of our defense, how can we best plan our defense to minimize the amount of sprinting needed (or maximize laziness, if you want to look at it that way)?

That night, I hopped on to Illustrator and made myself a map.

For the next few weeks, I watched nine other softball games. Every time a batter hit a ball, I would note five things:

  • Where did the ball land?
  • Was it a single, double, triple, home run, or out?
  • Was the hit a grounder, line drive, or pop fly?
  • Was the batter left handed or right handed?
  • Was the batter male or female?

I also decided that, if I was tracking the information anyway, I might as well follow the scientific method and propose a hypothesis. And even though my original goal was the determine where outfielders are best placed, there was another question I had that this data could potentially answer. With this in mind, I had two hypothesis ready for validation.

The average position of outflieders shouldn’t be stationary. Depending on the batter, the positions should change.

Fast ground balls are twice as effective as line drives.

My hypotheses were made! All that was left was to collect and interpret the data.

The data

  • Total hits = 406
  • Total men hits = 308
  • Total  women hits = 98
  • Green = single, double, triple, or home run
  • Red = out or fielder choice

Notes

  • Dots are where the ball landed. The graphs do not represent speed or rolling distance, just where the ball made initial ground contact.
  • Not enough data on lefties
  • Not enough data on women’s triples or home runs
  • This data has been pulled from Tuesdays at Comstock, Thursdays at Comstock, and Fridays at Howarth. They’re all bottom-level leagues but that’s why I have more data on men than women.