Econometrics Applied: New Robust Measure of Repeatable Skills

As you all know, I am part of the ascending breed of baseball writers changing the way we think about the game.  My colleagues and I are illuminating the folly of results-based analysis and how we can do better.  Employing advanced statistical tools, we are carving and chipping away at variables that are out of a player's control to reveal the Davidian statue that is repeatable skills.

I hope you will pardon us for our imperfections.  Consider us more Michelangelo in the Madonna of the Steps days than hanging from a suspended scaffold in the Apostolic Palace.  We learn as we teach.  You see, as we seek to educate, we learn from each other and even from you.  These blog communities are our symposiums.  We are borrowing and sharing, sculpting and shaving, mixing and matching, lecturing and listening.

While we are imperfect, we are constantly improving.  We brought you WAR, UZR, tRA, CFBV, wOBA, to name but a few strokes of the hammer on the chisel.  We have successfully reduced the impact of ballpark and luck on our understanding of  batters, of team defense on pitching performance, of subjectivity on defensive performance. 

Nevertheless, our tape measures tell us yards, not inches.  wOBA is easily skewed when a batter hits a lot of "seeing eye ground balls" through the hole that are actually "lucky ground balls."  UZR is only as good as the person coding outcomes.  tRA, in an effort to factor out luck and defense, actually factors out a lot of information about repeatable skills in a ,burning-books-type-of censorship power-trip.  CFBV is, frankly, a little heteroscedastic for my taste.

But worry you mustn't.  I have recently developed a hefty new statistical tool that lets us examine, with the most powerful microscope yet, repeatable skill, while stripping away random noise and luck.  The tool is called wSPAR, or weighted Standing Pressure Above Replacement.

Why is wSPAR better?

wSPAR data shows an extraordinary robustness that data from other measures cannot replicate.  In other words, wSPAR stabilizes rapidly, and measurements can be conducted across smaller sample sizes with smaller standard deviations.  This is to say that if Jose Lopez posts a wSPAR that is higher than Jack Wilson's over the course of just five games, you can be reasonably confident that Lopez will be the better player (in terms of wSPAR) over the course of a season.

wSPAR also has significant advantages over other measurements in terms of immunity to variables that are out of a player's control.  Random and often lucky variables like batting average on balls in play, team defense, and home ballpark cannot tip the scales in favor of one player over another when comparing wSPARs.

How iswSPAR calculated?

wSPAR is derived, much like WAR, by establishing a baseline wSP by measuring the wSP of a replacement level player at a given position.  Then each player's wSP is simply compared to that baseline in order to calculate a positive or negative decimal that represents the difference.

If you would like to learn more about the concept or calculation of a "replacement level player," I suggest you check out a website called Fangraphs, available at www.fangraphs.com.  wSP is actually also available on www.fangraphs.com, though the writers there have yet to incorporate wSPAR totals into articles addressed to laymen.  If you would like to conduct your own research on wSPAR, visit the player pages at www.fangraphs.com.  wSP is the second number in the third column under each player's name.

Great, so what does wSPAR tell us that we don't already know?

Well, a lot actually.  For one example, let's take a recent blog post by Dave Cameron, available here.  In Cameron's post, he attempts to put together a Seattle Mariners roster for next year given reasonable constraints such as team salary and reasonable trade partners.  Cameron relies primarily on WAR (Wins Above Replacement) to construct his team.  He does not even mention wSPAR and his spreadsheet has no trace of it.

It is no surprise Cameron did not mention wSPAR.  If you consider wSPAR, Cameron's proposed team looks much worse on paper than it does through WAR's rose-tinted glasses.  I took the liberty of reproducing Cameron's spreadsheet including the players he'd like the Mariners to break camp with next year, but I simply replaced WAR with wSPAR.

As you can see, it's no wonder Cameron omitted wSPAR.  Cameron's proposal has the Mariners trading Brandon Morrow for a below-average shortstop and paying $8 million per year for a below-average second baseman when they have two cheaper and more prominent options in-house.  And I know the "Ichiro The Savior" versus "Ichiro the Albatross" debate has kept many of you up at night pondering, but I think we can all agree that this demonstrates that, especially at his salary and age, the Ichiro investment will leave fans starved for more for years.  Finally, why should the Mariners ration their resources to provide for a feast in middle relief, with the relative famine on the team at closer? 

But the wSPAR picture is not altogether bleak.  RR-S, Saunders, and Nick Hill look like intriguing young, cheap options with loads of potential for growth.  Nick Johnson is expensive, but if we can afford him, he bulks up our lineup considerably.  Tui should get his shot at either second base or third.  And Carlos Silva appears to be primed for some serious regression (in a good way).

Lessons Learned

I know that not everyone will agree with this analysis.  As I've mentioned, disagreement breeds improved understandings.  And the more I am tested by sharp young minds who do not merely accept my words as gospel, the more refined my teachings become.

I will ask one thing though.  In the spirited debate that is sure to follow, please try to back up your opinions with real authority.  Let us all strive to be rigorous in our methods and thorough in our research, as I have strived for here.  Don't simply say "We should use WAR because that is what we've always used."  Do not hurl argumentum ad hominems like "you're just a geeky parents' basement number cruncher."

Writers like Dave Cameron and I write these sorts of informative articles for no pecuniary gain whatsoever.  Our writings are merely didactic.  We welcome dissent as much as we welcome praise, but let it be reasoned dissent.  If we can follow these simple rules, the Sistine Chapel will be within reach.