Friday, May 28, 2010
The WAR over Angel Pagan
By Mark Simon
One of my colleagues in our research department, Ryan McCrystal, and I have had an ongoing conversation the last six months over the statistic Wins Above Replacement, known in most statistical circles as WAR.
Ryan sent me a note, referencing who the leaders in the NL All-Star race would be, if WAR was the determining factor.
He was disturbed to see that the NL's starting outfielders included Angel Pagan.
Yes, that's the same guy patrolling center field for the Mets.
Ryan's a smart guy (he's written a couple of pieces for our TMI blog worth checking out), so I felt his comments were worth a closer look.
Let's put the stats aside for a second. The thought of Pagan as an All-Star, all things considered, feels like a reach ... a rather significant one.
But let's take a closer look at why the numbers are making this suggestion.
For one thing, WAR is not like batting average. There are multiple versions. There's the version on Baseball-Reference.com, there's a version on Baseball Prospectus, and there's a version on Fangraphs.
In principle, each version of WAR is the same thing. It's a measurement that attempts to take multiple aspects of a player's contribution (hitting, defense, and in some cases, baserunning) and combine them into one all-encompassing number.
Sounds like the kind of stat that would be worth having, right?
Well, we're not quite all the way there yet. There's no full-fledged agreement on the proper way to smush everything together.
In practice, the methods of WAR calculation are different. Without getting into the specifics, there are differences of opinion on how to value certain aspects of offense and most notably, defense.
The version of WAR on Baseball-Reference.com uses a means of evaluation that places a premium on players' putout and assist totals. The other sites take those into account too, but do so differently.
Baseball-Reference's and Baseball Prospectus's version of WAR say Pagan is about 10 runs above a "replacement-level" player (think: Gary Matthews Jr. or someone of that ilk).
The Fangraphs version of defensive evaluation doesn't just look at putouts and assists. It looks at where balls were hit, and how difficult they were to get outs on (the stat is referred to as Ultimate Zone Rating, or UZR). It says Pagan is only about three runs above a "replacement-level" player.
Both give credit to Pagan for something he's doing well. Pagan has four assists this season. That, in the small sample that is 2010, is an excellent total. If he maintains that rate and gets this kind of playing time all season, he'll finish with 14 or 15 assists, and you'll probably have a totally different perception of him than you do now.
Keep in mind that there are some very good center fielders -- Matt Kemp among them -- who have not done what Pagan has done defensively. WAR knocks them down a peg for that.
Getting back to comparing the WARs, while Baseball-Reference.com's WAR had Pagan as the third-best NL outfielder as of Thursday, Fangraphs rated him eighth, and Baseball Prospectus charted him ninth-best.
So one version is basically saying he's an All-Star starter. Two other versions say he's having a pretty good season.
Is that a fair assessment?
Well, let's combine some stats that we think are important to determine that.
Pagan is hitting .296 with a .361 on-base percentage and a .426 slugging percentage. He doesn't hit for power. We know that. But he makes up for that in other ways. His .787 OPS sounds rather blah, but consider where he's playing most of his home games -- Citi Field.
Factor the ballpark in and Pagan is hitting a little more than 10 percent above what an average player would do. That's actually pretty good.
And remember this: Pagan is the kind of guy who maximizes his talent by doing the little things well. As my colleague in research, Katie Sharp, said, he's the kind of player for whom you have to look beyond the back of his baseball card to appreciate.
That starts with his baserunning, which is factored into some versions of WAR (the ones at Baseball Prospectus and Baseball-Reference) more than others (Fangraphs)
* He's 8-for-11 in steal attempts (a stellar 73 percent).
* He's gone first-to-third seven times in 12 opportunities to do so. That's way above-average performance.
* And he's scored from second on a single four times in six opportunities (he held at third base twice). That's also way above-average performance.
Take all of the baserunning opportunities Pagan has had this season into account and you'll see he's taking the extra-base 55 percent of the time.
That's really good. Average players advance about 43 percent of the time. The difference between 43 percent and 55 percent is a big deal (he's not 12 percent better. You don't subtract one from the other). He's actually more than 20 percent better at baserunning than the average player.
We mentioned Pagan's defense already. The numbers say he's been pretty good, and as an added bonus, tell you he has a skill unlike a lot of other players (the assists).
So when you take an above-average batter, mix that with a way above-average baserunner, and combine that with an excellent (to this point) defender, what do you get?
You get someone about whom it's worth arguing, statistically or otherwise. Some might call Pagan an All-Star. Some might not. But he's at least someone worth checking out.
Mark Simon is a researcher for Baseball Tonight. You can follow him on Twitter at @msimonespn or e-mail him at firstname.lastname@example.org