6.23.2014

Luyman's Terms #10: The Value of a Bat, Part 1

Over the history of competition, observers have been trying to separate the good from the bad, and the great from the good. We crown champions, and we wonder how those champions became to be. We realize that some competitors are simple better than others, but what is it that makes them better? In baseball, the most popular way of determining this has been to look at a players batting average.

While batting average is nice, it is very seldom a good way to measure the batter of a hitter. A double is clearly better than a single, but batting average weights them both the same. What we need is a metric that weights doubles more than singles. slugging % tries to compensate by giving doubles twice the value over single, triples three times the value, and home runs four - but is that a good way to go about it? It a double really worth 2 singles?

Lets think about that for a minute. A double puts a man on second base, but 2 singles put men on both first and second! 2 singles is clearly better than a double, so slugging % is not as good of a metric as we originally thought. What we need is a metric than can weigh 2 singles better than a double, but also have a double weigh more than a single. Enter: run values.

The concept of run values is simple: each event in baseball has an expected number of runs. As plays begin to unfold in a game, a team can expect to score a certain amount of runs depending on how the plays turned out. This notion is hardly more than common sense, but by using run values it is possible to quantify exactly how much a hit or an out is worth.

For this post (and for the rest of this series), I will use the values given in The Book by Tango, Lichtman, and Dolphin for the 1999-2002 seasons:


By counting each individual event for each player, we can determine just how much value he gave his team with his bat over the course of a season. Let's do a sample calculation with Jermaine Dye's 2002 Season:

  1. Dye hit 71 singles in 2002, good for (71*.475) = 33.725 runs
  2. Dye hit 27 doubles in 2002, good for 20.952 runs
  3. 1 Triple for 1.07 runs
  4. 52 Walks (2 intentional) for 16.508 runs
  5. He was hit by 10 pitches for 3.52 runs
  6. Add all of these run values for a total of 109.303 runs "created" by Jermaine Dye's bat
This sounds pretty good! All Jermaine Dye has to do is never get out, and he will be worth 109.303 runs to his team!. Except, Jermaine Dye will get out. In fact, he got out 365 times in 2002. We need to make another calculation:
  1. 365 outs, good for -109.135 runs
  2. 109.303 runs from hits + -109.135 runs from outs = 0.168 runs total
This means that in reality, Jermaine Dye was good for 0.168 runs with his bat, which doesn't look so good upon first glance. In my next post, I'll show you why this is actually a pretty good thing.

2 comments:

  1. "A double puts a man on second base, but 2 singles put men on both first and second! 2 singles is clearly better than a double"

    Not disputing that, but just out of curiosity what is the negative drag value you think is placed on the latter scenario given the greater susceptibility to outs? A single man on second has the option not to run when the ball is hit, but men on first and second are forced to. Do the stats speak to that, or are they simply measurements in the abstract?

    ReplyDelete
    Replies
    1. I think that's partly dependent upon the outcome of the next batter. If the batter pops the ball up, the runner's on first and second don't have to go anywhere. Of course if the batter puts the ball on the ground, they need to move. The first scenario (in most cases) will end with 0 or 1 outs, and the second (again, in most cases) will result in 1 or 2. I have no idea how to truly quantify that at the moment. I'm sure it would involve something like looking at GB/FB rates over the course of 1 or more seasons and if/how those rates change over different base/out states (i.e a man at 2nd with 1/2/3 outs vs men on 1st and 2nd with 1/2/3 outs.) That's a tad over my head at the moment, but I'm working on getting better (part of the reason I took a step back from the blog, I promise I didn't give up!). The stats I gave in this post are simply measurements in the abstract and are context neutral.

      Delete