Sunday, December 14, 2008

Appreciating Willie Harris

Wee Willie Harris. He was a punch-line when the season started. "Voted 'least likely to hit a home run' on his high school baseball team." That sort of thing.

Well, as it turned out, Willie Harris had himself one heck of a year, easily the best of his career, and has been rewarded with a two-year contract worth $3 million. It's a steal for the Nationals.

Willie finished 2008 with with 367 at bats and a .251 / .344 / .417 line. That's nice, but for a corner outfielder it's certainly nothing to crow about. After all, the average major-league LF hit .269 / .344 / .442. It was Willie's defense that really set him apart. Whatever metric you look at, Willie was an absolute stud in the field. And combining his world-class defense with his solid bat and speed, he was one of the better left-fielders in all of baseball.

Before you decide I've gone completely off the deep end, I'm not the only one crowing about Willie's 2008. Sky Kalkman over at Beyond the Boxscore pulled together various metrics to account for both offense and defense and had Harris as the sixth best LF in baseball in '08, ahead of some guys you might have heard of like Adam Dunn, Alfonso Soriano, Pat Burrell, and Josh Willingham.

Let's look at some objective measures of our own and see what we get.

For offense, I'm going to use a Tom Tango stat you can now find on the player pages at Fangraphs called weighted runs above average (wRAA). You can read all the details and little thread about it here if you're interested in the details, but if not just think of it as the number of runs a team would gain or lose if the player being measured was replaced by an average offensive contributor. It builds on the old Bill James runs created metric, but is much better because it doesn't overvalue certain players the way the original James systems did.

For defense, it's harder to quantify the exact contribution of a player's performance in runs, but two of the better ones that do so are UZR and Rate2. Since I don't totally trust either I'm going to just take both and split the difference. UZR is also posted on the player pages at Fangraphs and gives us a projection for runs saved (or allowed) per 150 games compared to average defense. Rate2 does the same thing, but crunches the number per 100 games, so to get apples to apples I'm just multiplying the Rate2 runs above average by 1.5. Also, just FYI both stats break down performance by position, so for I'm just using the stats for the position each player played most often (LF for Willie, RF for Dukes, etc.).

Finally, to make sure we aren't weighting defense too much for the guys who didn't actually play a full season, I'm projecting out those wRAA numbers for a 600-PA season, about what a typical player hitting around the middle of the order would expect in 150 games.

These adjustments just allow us to adjust for playing time and establish some kind of apples-to-apples comparisons, but of course the obvious problem is that I'm assuming that if player X performed at Y level for 400 PAs that he'd maintain the exact same level for another 200. It's a little like early in the season every year when you hear about so-and-so who is on track for 100 homers after a hot opening week. In particular, this is relevant for Willie, because the rap on him is that he's a little guy who has tended to wear down over the season, an impression solidified by 2007 season, when Atlanta handed him a starting job only to see him fall apart in the second half to the tune of a positively blech .214 / .294 / .354 line. But still this will allow us to compare the overall offense + defense rate of production of different players while accurately weighting both aspects of the game. And given that Harris appeared in 140 games with 424 plate appearances, it's not like we're projecting an entire season out of fantasy.

With that caveat out of the way, let's get to the numbers and see just how good Willie really was. First, let's see how he stacks up against other Nationals position players who had at least 275AB in 2008. Short answer: there was no one better--not Guzman, not Dukes, no one.

UZR/150 Rate2*1.5 wRAA RAA/600 Total RAA
Harris 34.1 25.5 4.10 5.80 35.60
Dukes -1.4 3 14.90 26.77 27.57
Zimmerman 8.5 16.5 3.10 3.99 16.49
Guzman -1.7 6 6.40 6.27 8.42
Belliard -27 -15 11.40 20.30 -0.70
Kearns 23.1 7.5 -12.30 -20.67 -5.37
Milledge -10.9 -12 -1.80 -1.84 -13.29
Flores NA -19.5 -7.20 -13.33 -32.83
Lopez -12.6 -10.5 -15.80 -26.12 -37.67

Hold onto your hats, because if that surprised you, let's get a look at the very most valuable seasons by any Nationals position player since 2005.

UZR/150 Rate2*1.5 wRAA RAA/600 Total RAA
Johnson '06 11.9 6 39.4 37.64 46.59
Johnson '05 10.5 12.7 23.9 26.22 37.82
Harris '08 34.1 25.5 4.1 5.80 35.60
Zimm '06 16.4 31.5 5.3 4.40 28.35
Soriano '06 2.3 4.5 28.6 23.57 26.97
Kearns '06 10.3 13.5 14.2 13.55 25.45
Church '07 4 12 8.7 9.85 17.85
Guillen '05 11.5 1.5 10.9 10.70 17.20
Zimm '06 4.7 10.5 9.7 8.53 16.13
Young '07 -12.4 -1.5 18.2 21.50 14.55

OK, before you delete your bookmark to this page in protest of the apostasy of suggesting that Harris was better than Soriano, just step back and think about a few things. First of all, Soriano's '06 was impressive primarily because of the incredible counting stats he piled up--46 HR, 41 steals, and 41 doubles. No one's ever done that, and those are amazing achievements. But guess what? He also had a whopping 738 plate appearances, leading off all year. Only 47 hitters in the last decade saw that many. He was caught stealing 17 times. His rate stats (.351 OBP, .560 SLG, 135 OPS+) were very good, but not Hall of Fame great like those 40-40-40 numbers suggest.

And, of course, Soriano wasn't very good in the field. I'm actually surprised that these metrics liked him as much as they did. Many baseball fans (and some GMs I know) tend to grossly undervalue defense, but ask yourself--is a run prevented by a great defensive play worth any less than a run scored from a home run?

It's also worth noting how under-appreciated Nick Johnson was. If Soriano had been re-signed and had two bad injuries like Nick did, I don't think people would be so ready to toss him overboard. But Nick at his best was decisively better than Soriano at his best. Tex is an even better hitter and defender than Nick, but the fact remains there's something puzzling about a team going to such lengths to replace its very best player.

But back to Willie.
Don't get me wrong. I wouldn't trade Willie Harris for Alfonso Soriano or anything like that. I'm not even ready to hand him an everyday job for the Nationals in 2008. This was a career year for Willie both at the plate and in the field, and I think the odds he repeats or improves on it are not all that good. But looking at just this snapshot of one season, it happened.

If you're still shaking your head that Willie couldn't have been this good, you probably were one of the many who stopped watching the Nationals while Wily Mo Pena was still out there. Treat yourself. Have a look at these clips:
  • Willie robs Ryan Church of an RBI double with a game-saving flat-out horizontal diving stab in the bottom of the ninth of a 1-0 win in Shea Stadium. This is one of the most amazing defensive plays I've ever seen.
  • Willie robs David Wright of a two-RBI double with a game-saving catch over his head in the sixth inning of a 1-0 win over the Mets.
  • Willie stabs a sinking liner in the gap off Fernando Tatis.
  • Willie dives for a shoe-string catch off Brad Hawpe.
  • Willie does it to the Mets again with an impossible catch over his head before bouncing off the wall and making a perfect relay throw to double off Jose Reyes from first base, preserving a 3-3 tie.
  • Willie takes a home run away from Corey Hart with a leaping grab at the wall.
  • Willie crashes into the wall in left to steal extra bases from Shane Victorino.
  • Willie stabs a low liner from Ichiro, making it look easy.
Willie really was the Fifth Element, as NJ commenter and player page sponsor Section 506 says. In an awful season for the Nationals, it's too bad Willie's super season has been so overlooked.


JayB said...

Your right, Harris had a very good year. You’re also right, I have deleted you as one of my bookmarks. You have to be on the field to affect the game and less than 400 ABs makes this type of projection crap just that. Stats like these are nice and fun to look at but give me Alfonso in 2006 with his league leading outfield assists and his bat. How many less outs did that arm create, that would be 22! Yes teams were dumb to run on him so much but an out was still an out and one less that was needed to win a game. In baseball on the field…..those things matter……just check how many outfield assists Milledge had last year while leading the team in innings played. That would be one, just one.

JayB said...

Ok so I put the bookmark back already, so what......What this post really should have been titled was, "Manny Acta, lier or just unqualified to Manage".

If you believe Manny, it was his decision to get Harris only 360 ABs. We all know it was Jimbo who put Pena, Lo Duca, Casto, Roger B in LF and Lopez at 2nd base instead of Harris but if you take Manny at his word.....your stats clearly show......he is not qualified for the job, OR he is less than honest with the fans. Either way the fans lost out last year.

Steven said...

Soriano's outfield assists are a good idea of why context is so important. He had 22 assists because his arm was weak enough that people wanted to test him. But it wasn't quite as weak as people thought. Come to think of it, maybe that's why his defense ratings aren't so bad when he seemed so awkward in the field to the naked eye.

You're right Manny should have played Willie more.

Will said...

To be fair to Manny, Harris did start just about every game from mid-June until the end of the season. I don't fault Acta for not starting Harris at the start of the season (really who would've expected a season like that?), but I would blame him if he hadn't continued starting Harris after he put up numbers like he did.

Steven said...

@Will--that's right, but I think he waited too long to pull the plug on WMP and rotating through Lo Duca in LF was a waste of time. I don't think it would be fair for anyone to say that he should have known to play him from Opening Day, and I think JayB is obviously using a bit of hyperbole with this "unqualified to manage" bit, but it seems like he could have found the answer maybe a week or two faster than he did. Or, if Bowden was actually writing out the lineup card, that would be damning of the organization. I hope Manny wouldn't stand for that.

JayB said...

I think the 22 Outfield Assists show the value of Alfonso that year. An out is not just an out, not all outs are the same. An out at 2nd base, 3rd base or Home which is where all those Alfonso throw were, is much better than one at first. This is why your stats need to have the reality check of watching the games.

If, IF…….if jimbo was making out the lineups.......Estrada starting catcher over Flores in July.......come on....wake up...there is a reason Manny Acta was given the job without any experience and why so many much more talented managers refused to even interview.....

dcbatgirl said...

Why are you assuming everyone will disagree w/you?

Even without the documentation you provided, Willie's intangibles should be apparent to anyone paying attention.

Count me among those who will be happy to see Willie back next year.

Steven said...

@dcbatgirl--I dunno. My therapist says that it might have something to do with being a child of divorce and inconsistent displays of love as a child. Maybe that's TMI ;-)

OK but seriously I think the reason I think people will be skeptical is just that defense isn't as easy to measure and isn't hardly ever even mentioned in the mainstream press. If you're at every game and paying attention, you really see the difference. But while every homer shows up in the boxscore, it's hard to notice when a player is truly having an off-the-charts magnificent season in the field as it's playing out unless you're there, because those plays don't get tallied in the boxscore, and the best measures of defense (UZR, Rate2, PMR, Dewan's plus-minus) are hard to understand and don't generally come out in real time.

Soriano's offensive contributions were clear for all to see. The value of Harris's glove, cumulatively as great as Soriano's bat, is harder to perceive. So I think people will be naturally skeptical that Harris was really this good overall.

But just because the value of defense is harder to measure doesn't mean it's any less important. His glove was magnificent, and won several games for this team.

Hendo said...

I'm no fan of using a rebuilding team to extend the life of modestly skilled over-30 utilitymen. The Nats and you should take a bow for seeing that Harris is more than that.

His considerable tangibles aside, an intangible element of his value was pointed out during a Nats broadcast that I was listening to while I was driving back to Maryland from a P-Nats game late one Saturday night last summer, when either Slowes or Jageler -- I forget which -- said that Harris coaches himself at the plate. I'd always been a Willie fan for reasons I hadn't understood, but when I heard that, the Eureka! moment was so vivid I like to drove off the side of Route 1.

Put more succinctly, Willie Harris is the kind of guy you'd be scrambling around for if you didn't have him. Good on the Nats for keeping him in the fold.

James Bjork said...

As a scientist by training, I appreciate the new, improved metrics for player impact above the old stats on my Topps cards growing up (that Boswell still clings to apparently). Therefore I am not surprised that there is an increasing appreciation of how wins are determined by runs scored RELATIVE to runs given up.

That said, the implications of your number crunching, even if you overstate Harris' net worth by 20-30%, are staggering. I find it hard to believe that we got such plus ability for such a pittance!

I'm still left wondering if this is for real.

On another front, one still has to consider that from a popularity standpoint, perception IS reality, just as in politics. A Pat Burrell type will nevertheless do more to increase the 9000 viewers on MASN (and fannies in the stands) more than highlight reel catches.

Will said...

James, well said. Just note that defensive stats can fluctuate significantly one season to the other. Players considered great defensively have been known to go from the best to average back to the best in three seasons. For example, Carl Crawford, one of the best LFs in the game, went from 14.5 runs above average to 4.7 to 28.6 over the past 3 years. Curtis Granderson went from 10.4 defensive runs above average last year to -11.0 runs below average this year.

The defensive metrics are still a little rough, but have improved dramatically just over the past couple years. That being said, Harris has been excellent defensively whenever he's been given ample playing time (his UZR/150 was 20.5 runs above average in 2007). On top of his defensive ability, he has immense value in his versatility. He played five different positions last year.

Harper said...

Willie is like the Bizarro Redding. At first we were like "Oh this guy is awful!" then during the course of the season he didn't do anything specifically to shaker that feeling. But at seasons end in review we were like "Hey, this guy had a pretty decent year".

Now if only it didn't take his best offensive season of his career at age 30 to make this happen...

Steven said...


I agree with you both. This isn't what I expected when I started. I drafted a post which described Willie's season as "one of the best seasons by a Nats outfielder not named Soriano."

But then I punched the numbers into my spreadsheet and found that Harris was even better. I went through and double-checked all the numbers because I didn't really believe it. But I can't think of a reason why my methodology isn't sound and there aren't any typos.

I agree about the inherent slippery-ness of the defense numbers, which I think is a function of sample size, not that it's BS. That's why I pulled both the UZR and rate2 numbers, rather than one or the other. But note that Willie comes out ahead whether you use UZR or rate2. There's a lot of evidence that Willie was not just good but absolutely marvelous in the field, and that performance translates into actual runs saved that are every bit as valuable as the runs created by Soriano's homers.

I do think it's important to note that we're talking about their rates of production, not absolute production. I don't think that's taking anything away from Willie to just note that we're talking about rates of production. Over 2/3 of a season he was more productive than Soriano was over a full season. But it's not the same, and there's also a lot of evidence to suggest that Willie isn't likely to maintain this rate of production.