Using Win Expectancy to Gauge Watchability

The Book elaborated quite nicely on the idea of win expectancy on a given play or even pitch. The basic idea is that using the score of the game, inning, number of outs, and baserunners we can can calculate how likely either team is to pull out a win in a given game. When I say “we can calculate” what I really mean is that Fangraphs does an incredible job of tracking this in real time. If you’re unfamiliar with the concept I’d consider having a tab open with the game of your choice on next time you’re watching. You’ll be able to see how each plate appearance causes your team’s chances of winning or losing to move either up or down. Here’s an example from arguably the most important game in Rays history:

CLICK HERE TO SEE THIS MORE IN DEPTH

Each team starts with an equal chance to win, which isn’t generally the case, but this is an assumption we must make for everything else to fall into place. You can see how various events increased or decreased the chances of victory and if you click the link you’ll find the corresponding events. Here’s another “fun” game that will probably make any Rays fan’s top-10 list for favorite game of all time:

Now that you should have a rudimentary understanding of this powerful tool on to the application here. The idea is that games where your team has a 75% chance of winning or losing are basically not as much fun to watch as when it’s back and forth so that you truly don’t know if you’re going to win or lose.  Fangraphs doesn’t explicitly allow you to search for this so I compiled every offensive player’s plate appearances this year. E-mail me if you want the workbook since it took some time to compile and there’s tons of uses beyond just this post. I looked at non-pitcher plate appearances for all AL teams this year.

The first thing we want to look at is what I mentioned above. What’s the percentage that your favorite team was winning or losing big and from there what’s the percent of plate appearances where the game is still reasonably in question? Well here you go:

Winning Big% is defined as the percent of plate appearances where that team had a 75% or greater chance of winning and Losing Big% is when the team had a 25% or less chance of winning. The middle ground is when we would consider the game’s outcome still in the balance which should be the most exciting time, though I’m also a sucker for the big win. Losing big doesn’t necessarily translate to a loss as anything can happen before out number 27 (or more), but for the most part this is when fans do their most bitching.

You’ll notice the Rays have the third highest percentage of winning big at 29% trailing only the Red Sox (33%) and Tigers (32%). Conversely, the Rays are tied for the fifth lowest percentage of losing big at 20% with the Tigers, Red Sox, and Orioles (all at 16%) and Rangers (18%) spending the least amount of time getting creamed. The middle ground is what we’re hypothesizing is the most exciting time to watch. By this measure the Rays are ranked pretty low with only the Indians having a lower percentage of plate appearances with the game’s outcome less than obvious. This seems to fly in the face of my hypothesis as it seems like the Rays have been a lot of fun to watch this year with so many tight games, but comparatively, other teams have been in the middle more often.

Let’s also look at this by overall expectancy of winning or losing:

Here we’re strictly looking at the percent of plate appearances where the team has a less than 50% win expectancy (Losing, red) or a greater than 50% win expectancy (Winning, green). You’ll notice that the Rays have the fifth highest percentage of win expectancy, but they’re basically even with the Orioles, Athletics and Rangers with all around 57%. The Tigers have spent a whopping 62% of their plate appearances with a better chance of winning than losing and the Red Sox come in second at 60%. The other side of the ledger shows the Royals with a positive win expectancy only 46% of the time “besting” the Astros (46%) and White Sox (52%). It must be awfully hard to watch those teams this year when you’re expected to lose so often.

Lastly, I broke this down by ten percentage point buckets:

I’m aware that this is the wrong application of the word “decile” but I think it’s intuitive enough to lead most minds in the right direction. Basically, these are the percentages that you’re in each bucket so the Rays have spent 10% of their plate appearances between zero and ten percent likelihood of winning. This is lower than or on par with the best teams and much better than the worst. The Rays have the second highest percentage of plate appearances with a win expectancy between 90 and 100 percent likelihood. Bear in mind that if you win the game you’re going to have 100% win expectancy and vice-versa with losses being at 0%. Doing some quick math you see that the Rays spend 33% of their time with a 70% or greater likelihood of winning which is better than all but the Tigers and Red Sox.

I think this is a good read for a fan of every team, but it’s nice to see confirmation that the Rays have been a great watch this year with many games not featuring too much time wondering if we’re going to be able to pull it out. I don’t see any predictive value to this, but it’s always nice to look back on things like this and realize that even when we’re losing an individual game it could be a lot worse.

Advertisements

About Jason Hanselman

Rays fan.
This entry was posted in statistics and tagged , , , . Bookmark the permalink.

One Response to Using Win Expectancy to Gauge Watchability

  1. Pingback: Watchability Throughout Rays History |

Leave a Reply

Please log in using one of these methods to post your comment:

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