Skip to content

Do Batting Runs Overrate Ichiro?

August 8, 2010

Non-Giants post, obviously.  This is a thought that crossed my mind a few days ago (honestly, how much of a nerd can you be if stuff like this randomly pops into your head?) and I thought I’d do a bit of investigating.

Batting Runs, also known as Linear Weights, are the preeminent method for valuing individual hitters.  If you’re familiar with this site or sabermetrics in general, you’ll know why- if not, I’ll briefly explain.  Linear Weights take the marginal run value of each event based on its impact to the run-scoring process.  We all know that a double is worth more than a single, a triple more than a double, and so on and so forth.  LWTS essentially give us an empirical value of every event, and are ideal for evaluating individual hitters.  Why?  Well, not all hitters face the same opportunities due to the lineup they’re in.  And while it is certainly true that a walk is more valuable to a leadoff hitter than, say, a #8 hitter, the player has no control over where he is placed in the lineup.  A #8 hitter for the Yankees, for example, might be the leadoff hitter for another team like the Nationals.  So we can’t look at runs and runs batted in, because they’re heavily dependent on the other players in the lineup.  By using the marginal amount of runs each event produces, we can better get an idea of how valuable a hitter’s performance is in a neutral lineup.  To get an idea of what this looks like, here is a LWTS equation for the 2009 NL (pitcher hitting excluded):

LWTS = .47*1B + .78*2B + 1.07*3B + 1.42*HR + .49*ROE + .30*NIBB + .33*HBP + .17*IBB – .29*NonKOut – .30*K

Note the bolded coefficient for singles- the run value for a single is around .47 runs.  Except that accounts for all singles.  Not all singles hold the same weight, though.  Infield singles, for example, should hold less weight than a non-infield single because it has a low advancement rate.  Turns out this is true.  A while back, I read an old article written by Mitchel Lichtman- the creator of UZR- entitled “What is Manny Worth?“, in which he investigates Manny Ramirez’s (then) future contract heading in to 2007.  He provides us with this linear weights equation:

MGL’s LWTS = .48*(1B – Inf1B) + .40*Inf1B + .77*2B + 1.07*3B + 1.40*HR + .32*BB + .51*ROE – OutValue*(AB – H)

I asked MGL in an email a while back if the run value of the infield hit generally sits around .40, and he confirmed that this is indeed the case.  Pretty cool, no?  This is something that’s usually- if not always- overlooked in most hitting performance estimators.  This made me wonder, then…how about those slap hitters that rely on infield singles to reach base?  How much does this change their offensive value as measured by linear weights?

The first player that came to mind was Ichiro.  I’m sure other players might come to yours first- maybe a Juan Pierre, a Michael Bourn, and so on and so forth.  Why check Ichiro?  Well, there’s three reasons: the first being that he’s a very prominent figure in Baseball, and the second being that I’m sort of a closet Mariners fan (my favorite AL Team are the Rays, with the Mariners trailing behind.  It was their acquisition of Justin Smoak, whom I have had a Baseball man-crush on ever since the 2008 draft, that made me jump on the fanwagon.).  The third reason is because he’s had a very long and very successful career, with 50.7 rWAR heading in to 2010 (I’ll be excluding 2010 from this analysis, by the way, because the season isn’t over yet).  So here are the league’s infield hits rate, calculated as INF1B / 1B:

The league average generally sits about 14-16% of all singles.  That’s pretty good- players by and large do a bit more damage than just slapping the ball into the 5/6 hole.  But what about a player like Ichiro, that seems to (by my own observations and by his reputation) do this rather often?  How much does this rate change?

Pretty crazy, huh?  Ichiro is actually hitting infield singles at almost twice the rate of the league average.  That’s some crazy slappage.  How does this convert into runs?  Remember, the difference between a non infield single and an infield single is .08 runs, so…that’s our run value.  We can figure Ichiro’s runs above/below the league average by comparing his expected infield singles to his actual infield singles, and multiplying by the run value.  That gives us this:

And…BAM!  There it is.  While on the seasonal level we’re not looking at a huge difference- about 2-3 runs- this definitely adds up over time.  Ichiro’s been overvalued by roughly 1.8 wins over the course of his career, and by the end of 2010, it wouldn’t surprise me if he’s surpassed the two wins mark.  This looks like a pretty fun topic to cover, so perhaps I’ll take a more in-depth look at other players at some point in the future.

6 Comments leave one →
  1. terpsfan101 permalink
    August 9, 2010 7:20 AM

    I don’t think batting runs overrate Ichiro. He’s getting credited with an extra 20 RAA because of his infield singles for his career. But batting runs don’t take into account GIDP/GIDP opportunities. Ichiro is 40 RAA in GIDP Runs for his career (according to Rally). They also don’t account for baserunning. Even when batting runs include SB and CS, they aren’t entirely accounting for baserunning. Rally has Ichiro at 47 RAA in baserunning runs. These include his stolen base runs. Excluding basestealing I’d guess he is 20 RAA in baserunning runs over the course of his career.

    Ichiro is not a great hitter. What makes him a great player is his combination of good hitting, great basestealing, great baserunning, great ability to not hit into DP’s, great defense, and great situational hitting. Ichiro is the kind of player that Linear Weights undervalue, because linear weights often do not account for the little things that Ichiro does well.

    • triplesalley permalink*
      August 9, 2010 8:20 AM

      You’re absolutely right. The “analysis” is rather short-sighted in that it only looks at BtRuns and ignores the situational aspect of the game. Perhaps I should have focused on a different player that ISN’T undervalued in so many different facets of the game to make my example. It’s less interesting, perhaps, but…it’d certainly be more accurate.

      Thanks for pointing that out, by the way. I should’ve chosen a better example for something like this.

  2. terpsfan101 permalink
    August 9, 2010 5:18 PM

    I’m nitpicking here, but your Linear Weights equations should not include IBB and SH. The win value of an IBB is approximately the same value as an average plate appearance for that batter. MGL says “we use run value as a proxy for win value.” IBB have a positive run value, but as a whole they have marginal win value very close to zero. The same thing applies to the sacrifice bunt. Sacrifice hits have a negative run value, but their win value is very close to zero. Another reason you don’t want to include IBB and SH in your linear weights is that they are outcomes that are out of the batter’s control. Managers decide when to issue an IBB and when to attempt a SH.

    So here is what you should do. Let’s say Pujols has 65 batting runs in 650 plate appearances (not counting his IBB). His LWTS per PA are .10. He was intentionally walked 30 times. What you do is assign a run value of .10 to his 30 intentional walks. So Pujols gets credited with 68 batting runs in 680 total plate appearances.

    You do the same thing with sacrifice hits. If Jack Wilson has -20 batting runs in 600 “non IBB and SH plate appearances” and he has 10 sacrifice hits and 10 intentional walks (batting in the 8th hole in front of the pitcher), then you assign a run value of -.03 (-20/600) to his 10 SH and 10 IBB. So Jack Wilson is credited with -21 batting runs in 620 plate appearances.

    • triplesalley permalink*
      August 9, 2010 7:10 PM

      Please don’t ever hesitate to “nitpick” or to point things out, especially on a topic such as this- your knowledge of LWTS and the like is infinitely superior to mine, so I always enjoy (and absorb) the added insight.

      I find your last two paragraphs very interesting, as I haven’t seen IBB & SH handled that way before. Why is it done that way, if you don’t mind me asking?

  3. terpsfan101 permalink
    August 9, 2010 8:48 PM

    You don’t want to completely ignore IBB and SH plate appearances. So you just use the average value of a non-SH/IBB plate appearance for that batter. In my example above, you assume that Pujols and Wilson would of performed at the same rate in their IBB and SH plate appearances as they did in their non-IBB/SH plate appearances. Their rate stats remain the same both before and after accounting for IBB and SH, but the counting numbers change. Pujols is .10 LW per PA before you account for his IBB’s and he is .10 LW per PA after you account for them. Likewise, his wOBA is the same. However, his counting LWTS change. He goes from 65 batting runs to 68 batting runs. The whole reason why you do this is that IBB and SH are win neutral events. They have a marginal win value that is very close to zero. If you used the empirical run value of an IBB and SH in a player’s LWTS, then you would have to use a custom runs to win converter for every player. By ignoring them you can directly translate runs into wins.

    The reasons why you don’t assign the actual run values for IBB and SH had to be explained to me in this thread:

    Read posts #15 through #21 and #28 through #31. You can read the whole thread if you want to. In hindsight, I said some pretty stupid things in favor of including IBB and SH in run estimators.

    • triplesalley permalink*
      August 9, 2010 9:40 PM

      That makes perfect sense, and that link has been quite helpful…thank you very much! 🙂

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your 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

%d bloggers like this: