There is an interesting discussion going on over at The Book, a blog run by Mariners freelance statistical consultant Tom Tango. He goes by the name Tangotiger in internet discussions. This particular discussion/debate concerns Tango’s rebuttal of a reader questioning the consistency and usefulness of Ultimate Zone Rating (UZR), one of the advanced defensive statistics that are becoming more mainstream as we head towards 2010. I know many of you have expressed frustration with wild variances in UZR, as I have as recently as yesterday.
When trying to view Adam LaRoche’s defensive numbers at first base, starting with this past season and going backwards, I found him at -0.5, -5.9, +6.3, -3.8, -15.7, -2.2. Yeah, it’s a little all over the place. I suppose, when you consider the minuses outweigh the pluses and that, leaving a margin for error, everything is reasonably close to zero, you can consider LaRoche a bit below average — which jibes with what scouts have said for years — but still…it isn’t quite a complete picture, is it?
Unlike offensive stats, the modern defensive ones are not yet as reliable on a month-to-month — or even year-to-year — basis. Make no mistake, the offensive ones are not 100 percent reliable, either, as continued discrediting of the value of “batting average” has shown over the past few years. You can have “empty” .300 hitters a la Yuniesky Betancourt and more-complete .275 hitters who deliver value in terms of walk-drawing ability and extra-base power. So, like everything else, offensive stats continue to evolve as do the defensive ones.
But it’s a worthwhile discussion and one Mariners fans should take great interest in given how much of a priority the current team is putting into its gloves and the measurements used to gauge them. Trust me, the last thing the Mariners are looking at when it comes to gauging defense are things like fielding percentage. That’s Stone Age stuff and tells you less about what a fielder is likely to do than flipping a coin — the latter of which, unlike the former, gives you at least a 50-50 proposition of being right.
The entire mission of new defensive stats like UZR, Plus-Minus, Probablistic Model of Range (PMR), Revised Zone Rating (RZR), Out of Zone Plays (OOZ), and whatever else is spawned by them, is to tell you how good a fielder was on plays when he doesn’t touch a ball. That’s the simplest way of boiling it down. Traditional stats like fielding percentage and errors can tell you all you need to know about how a fielder did when he touched a ball. But you’ll learn nothing about their range and overall ability. I watched Orlando Hudson get charged with a ton of errors early in his career with Toronto because he was diving for balls Jose Lopez can only dream of getting close to. So, who was the better fielder? Fielding percentage won’t tell you. Even worse, it could actually mislead you into thinking a defensive pylon was better than a guy who should truly be winning a Gold Glove.
And when you put as much stock into defense as the Mariners, you’d better make reasonably sure you’re going off the right measuring charts.
Thing is, these defensive stats aren’t perfect. Even more important to remember is, they are less perfect than offensive stats.
And that means we have to be careful about putting too much faith into any one system or being too quick to put labels on certain players.
For instance, what are we to make of Adam Jones?
If we go by UZR, he was a +9.9 center fielder for the Orioles in 2008 and a -4.1 guy for them last season. So, is he good or isn’t he? We’ve all heard that the Mariners traded away a solid defensive gem — he won a Gold Glove last season, for crying out loud (we’ll get into the sham those awards have become another time) — and yet his UZR numbers claim he went from above average to below average in just one season.
Confused? You’re not alone.
We interviewed the creator of UZR, Mitchel Lichtman, back in March for a feature on modern defensive stats and he was the first to admit that you often need multiple years worth of data for his system to be reliable. The common theory espoused throughout the blogosphere is that three years worth of UZR will tell you what you need to know.
Sometimes, though, teams don’t have that much of a track record to go off. Like with Jones, who has only played two full seasons. If you were trading for him now, would you give up the players needed to acquire a Gold Glove specialist, or a slightly sub-par defender?
Other times, an injury might impact what a player did two years earlier. And as we know, judging a ballplayer based on anything they did further than three years back in a career can be a real crapshoot.
The trouble is, this is the best we’ve got for now.
And it’s still better than going off fielding percentage.
So, while you won’t understand exactly what Jones beings to the table, the prolonged track record of somebody like LaRoche tells you that signing him likely won’t bring an above average fielder to the Emerald City.
I know that doesn’t sound overly reassuring. But it’s progress. Once again, fielding percentage tells you next to nothing about a player’s defensive ability. Just about any player can make 85 or 90 percent of the plays on balls he gets to. Even I can do that. I may get to one out of every 20 balls hit my way in the outfield. but darn it, at age 41, I can still catch that ball hit right at me and be a 1.000 fielder.
In fairness, too, UZR is coming under increased scrutiny because it is one of the only advanced metrics the average fan can access daily (on FanGraphs.com). UZR is bringing advanced defensive stats to the masses and that’s a good thing. The problem is, some fans — even those well-versed in sabermetrics — have a tendency to view these numbers the same way they would offensive ones, forgetting that they are less-than-reliable at times.
Some guys have been labelled as defensive “busts” because of negative numbers, even multiple years in a row, when they may have been victims of the home ballparks they play at. We see this debate going on internally in baseball circles right now with Jason Bay and his poor UZR numbers at Fenway Park. While UZR does adjust for park factors, even Lichtman will tell you the Fenway stuff is largely guesswork because of the bizarre Green Monster in left field and its unique height and proximity to home plate.
So, while baseball insiders will pretty much agree that Bay is no Endy Chavez as a left fielder, there is little agreement on how bad he actually is because of the way Fenway skews defensive stats for left fielders. And it may not seem important to you, but it is very important in gauging how valuable a player is. A slightly below average Bay in the field could wind up being worth $16.5 million per year because of the value his bat brings. Whereas a terrible defender in Bay could never offset that with his bat at that price.
I’ve told you before, that the Mariners fell into the former camp on Bay. They felt his defense — after making internal inquiries with the Red Sox — was not as bad as UZR showed and that he would not be the fielding liability some suspected. The Mariners did not feel he was worth $16.5 million per season like the Mets do. But Seattle had a number and would have acquired Bay had he agreed to it.
We all heard talk last year at this time about how the Phillies were fools to give Raul Ibanez a three-year, $31.5 million deal because of his terrible left field defense in Seattle. Ibanez scored a -12.1 in 2008 and a -20.9 in 2007 on the UZR scale.
But then, moving to a more fielding friendly ballpark, which maximized his defensive skills, Ibanez posted a +8.0 in Philadelphia last season. Not exactly a bust in the field, was he? At least not according to UZR. Yes, that’s a wild fluctuation.
According to FanGraphs, the combined offensive and defensive production Ibanez gave the Phillies was worth $19.1 million — nearly two-thirds of the entire three-year value of his contract in just one season alone. So, those who criticized the Phils for their Ibanez deal — including some of the brightest minds in sabermetrics — were, at least, according to the numbers, dead wrong. He gave them nearly double the value they paid for in 2009.
Sure, he could regress next season, who knows?
But those who trumpet the value of Adrian Beltre and what he delivered in Seattle over the length of his contract can’t suddenly turn around and say the same logic doesn’t apply to Ibanez. Maybe Ibanez was just a one-year Philly fluke and will be a disaster in 2010 and 2011. Or, maybe not. Maybe those who were quick to label him become of his UZR numbers were way off-base.
Stats maven John Dewan of Baseball Info Solutions, creator of defensive systems like RZR and Plus-Minus, warned me last March that Ibanez was a better defender than the numbers were giving him credit for and that he’d do well in Philly. So far, at least. Dewan is looking to be correct.
The lesson with Ibanez, who has criticized the inaccuracies of defensive systems before, is to not be too quick to label someone or criticize a potential trade or signing based off of UZR alone. Or any one defensive system for that matter.
And I won’t do it here either. If Ibanez regresses to a minus-20 defender the next two years, he might prove a Philadelphia disaster yet. Thing is, we just don’t know.
Best option I can recommend is to do what the Mariners do. Tony Blengino, the head of the team’s stats department (and technically Tango’s boss) does not subscribe to any one defensive formula. Blengino enjoys looking at the Out of Zone Plays stat (available free, along with RZR on The Hardball Times website) when gauging defense. But he takes a multitude of stats into consideration. The Mariners also supplement that info with the advice of scouts and have their own, in-house stats tools they go by as well.
This is a long way of saying that UZR remains on the cutting edge of what folks are trying to do with defensive numbers. It is also one of the few tools the average fan can access free of charge to get a snapshot of what defense a player brings to the table.
But it is only a snapshot. Only one tool. A better tool than fans had in the 1980s with fielding percentage and errors. And one that will get better as we head into 2010 and gain insights we didn’t have before — as long as we don’t misuse it by believing it to be the absolute, final authority on all things defense.