A Mild Protest Against the “One-Number” Stats

Mild protesters.

== And a Brief Preview of the Stats I Use ==

This will lead up to my rankings of the best hitters and pitchers of 2013, and how the Mariners compare.

Phase I

1) Baseball is paradox.

Swing hard … but don’t miss.

Wait for your pitch … but don’t get caught looking.

Throw strikes … but not meatballs.

2) The very best baseball players of all time were remarkably accomplished at consistently accomplishing self-contradictory feats.  How did Ted Williams “hit the ball on the screws” so consistently while hardly ever missing?  Well, that’s the issue, isn’t it?

3) Intuitively, baseball folk have tended to gravitate to three numbers:

  • Batting Average | Home Runs | RBIs
  • Wins | Losses | ERA
  • BA | OBP | SLG
  • OBP + SLG = OPS

4) I believe the tendency to choose three numbers relates to the inherent nature of baseball:

For hitters:

  • Don’t make outs
  • Produce offense
  • Do both at the same time

For pitchers:

  • Get outs
  • Deny offense
  • Do both at the same time

5) OPS has become the new mainstream stat because it is a relatively simple reflection of baseball truth:

  • OBP = not making outs
  • SLG = producing offense
  • OPS = doing both at the same time


But now the Power of Three is getting smothered.

“One-number” stats are on the rise — including, but not limited to, the preeminent WAR and wOBA — with the goal of refining OPS even further such that all of baseball is reduced to a single number.

Except that all of baseball is not easily reduced to a single number, because the single number does not capture the ability to resolve the paradox … to swing hard and not miss … to throw strikes but not meatballs … to avoid/get outs while producing/denying offense.

In a very rough way the “old-school” stats even got at this.  BA is a rough measure of avoiding outs; HR a rough measure of ability to hit the ball hard; RBI a very rough measure of producing offense.  Just because OPS captures it much better doesn’t mean the old coots didn’t have some idea of how things work.

To my mind, the key image is not a thermometer (higher = hotter), but the Venn Diagram. Greatness lies at the overlap of the otherwise paradoxical.

Of course, I’m exaggerating here for effect. The “slash line” with three numbers is still pretty much universal, and I don’t think anyone is claiming that we ought to look at only one number.

But if we’re picking three numbers, might there be better ones?


Phase II

1) While stats are divided between “rate stats” and “counting stats,” my sense is that all of the preeminent stats are “additive.”  That is, they add together various outcomes, but do not subtract.

  • My understanding is that the primary offensive component of WAR is wOBA.
  • My understanding of the formula for wOBA is that it adds together various results without any subtraction (other than intentional walks are subtracted from walks).

2) I believe that a central element of baseball is the difference between a strikeout and a ball-in-play.  The strikeout reduces the chances of getting on base to zero.  A ball-in-play does not.

3) Therefore, the ideal stats need to incorporate the negative cost of a strikeout.  I think this is an area that can be improved upon with the currently preeminent stats.


Phase III

1) I also believe that the difference between a single and a ball-in-play out is frequently random (although not universally random), but that the difference between an extra-base hit and a ball-in-play out is rarely random (or at least considerably less random).

2) Sometimes I use the term “random-y” to convey the notion of an event that is relatively dependent on chance.  My approach generally treats all singles as “random-y” and extra-base hits as not “random-y” (while recognizing that such is not technically true).

  • Extra wonky part: The underlying assumption is that, for most hitters, the number of “non-random-y” singles will roughly equal the number of “random-y” extra-base hits, and, therefore, we can use the number of extra-base hits as a proxy for what we are really trying to measure: balls hit with sufficient authority so as to not depend on a “random-y” element to result in a base hit.

3) As a general rule, then, I give very little weight to the offensive value of singles.  Not because they are meaningless, but because they are “random-y,” and I’m trying to measure what is not “random-y.”

4) There are hitters who can consistently produce non-random-y singles — Ichiro, to pick one — but when dealing with stats in bulk, as one has to, it is best to understand that this select class will not “register” as highly as it ought to, and deal with it separately.  So, understand: my approach will always understate the value of “singles technicians.”   We just need to go into it with our eyes open to that fact.


All of which leads to the measures that I have adopted:

  • Plate Skill Advantage, which is a measure of the ability to avoid non-random outs.
  • .900 Conversion Rate, which is a measure of the ability to produce non-random offense.
  • Composite, which combines the two above measures based on the 10-year MLB average for both.

Because the first two are pretty complicated, I have also devised a more “simplified” measure, Plate Value Index, which I have found generally ranks players in approximately the same way.

Another advantage of these stats is that they can measure hitters and pitchers.  We can just as easily measure the pitcher’s ability to get non-random outs and deny non-random offense using these same measures.

The next article will explain the numbers in more detail, and then on to the 2013 rankings.  Spoiler alert: Two of the top five hitters in baseball in 2014 2013 [oops] played for the same team, and it might not be one you’re thinking of.


6 thoughts on “A Mild Protest Against the “One-Number” Stats

  1. I think you mean “Two of the top five hitters in baseball in 2013 played for the same team”

    My apologies for being picky- it’s in my nature! That aside- I have really enjoyed the content on your new site; filling a gap that existed since Jay Yencich decided to hang up his keyboard and focus on something besides teh Mariners’ Minors. Well done!

  2. Your paradigm is similar to mine in that we both conceive of baseball as best viewed through the lens of identifiable non-random skills, and the convergence of those skills. So Ill be interested to see what your chosen numbers reflect. I would pick slightly different numbers, but I suspect this will be valuable enough to start a lot of good conversations.

  3. Pingback: The Philosophy of the Stats — Details | Mariner Brainstorm

  4. Pingback: The 20 Most Valuable Hitters of 2013: 20 through 11 | Mariner Brainstorm

Leave a Reply

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

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ 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 )


Connecting to %s