Archive for Defense

Photo credit: Kathy Willens/AP

Just how good is the Yankees outfield defense? The eye test paints a pretty picture, and the numbers provide a similar perspective. This morning Stephen cited a Dan Barbarisi post that further examines the defensive numbers for the Yankees’ outfield, and the returns are predictably good. As a unit the Yankees outfielders have a UZR of 20.1, or 8.7 per 150, which ranks third in all of baseball. Only Arizona and Boston lead them. The major difference among the three teams is how they accumulate these defensive numbers. Both Arizona and Boston accomplish this with range; their 37.3 and 22.8 range runs lead the league by a decently wide margin. While the Yankees do have quality range numbers, they have something that the Red Sox and Diamondbacks do not: quality outfield arms.

As a unit the Yankees’ outfielders have produced 4.4 runs above average with their arms. That ranks seventh in baseball, and just 0.3 points away from fourth. All three of the starters not only have positive arm scores this year, but all three rank in the top 20 among all MLB outfielders in arm score. Again, this passes the eyeball test at least as it concerns 2011. They’ve all had issues in the past, but it does appear that they’ve turned it around. In 2011 they’re apparently turning the corner.

Before we proceed, a word about the small sample that is the 2011 season. It is absolutely true that to gain any value from defensive metrics you need heaps of data — preferably three years’ worth. Clearly we’re not getting anything close to that by examining year-to-year improvements for each player. Yet I’m confident that we’re measuring something real — that is, something that actually happened on the field — when we’re looking at arm scores. From the FanGraphs UZR primer, arm scores are “based on the speed and location of batted balls to the outfield and how often base runners advance extra bases (advances), don’t advance the extra base (holds), or get thrown out trying to advance (kills).” While speed and location are subject to bias, the play-by-play data can give us a good idea when it comes to advances, holds, and kills. So while there is a level of noise in these data, there is also some truth, stemming from the “it happened” factor.

Since he arrived in New York, it was apparent that Nick Swisher had an arm more suited for a left fielder, or even a DH. He lollipopped throws with consistence in 2009, and the numbers bore it out; he had a -5.9 arm score, which was tied with Brad Hawpe for worst in the majors. The problems were so bad that he went to then pitching coach Dave Eiland for advice on how to better hurl a baseball. That seemingly did the trick. In 2010 he improved to -0.8 arm runs above average. This year he’s at 1.6 runs above average, which ranks 19th among MLB outfielders.

While Gardner occasionally uncorked a five-bouncer to home plate during his first two years in the outfield, he still produced generally good arm numbers. From 2008 through 2009 they went: 5.0, 2.4, 6.6. The score in 2008 and the huge jump in 2010 might have been a product of perception. Gardner doesn’t look like a guy with a quality arm, therefore coaches and base runners might be more apt to attempt the extra base. To wit, he had 12 assists last year, which ranked second among MLB outfielders. This year he has only six assists, perhaps because the league has adapted to his actual arm skill. Despite that he still has an arm score of 1.7, which ranks 15th among MLB outfielders. It suggests that he’s holding base runners, rather than killing them.

That leaves Granderson, who had mixed results in terms of arm score earlier in his career. He was actually below average in his final two years with the Tigers, but has been positive in both of his seasons with the Yankees. In fact, his 1.9 arm score from this year ranks 10th in baseball. This is due, in large part, to his eight outfield assists, which ranks 15th among outfielders. The only other year in which he’s had more than five assists was in 2007, when his arm score was at a career high 4.1. I want to say that Granderson’s arm score stems from the same bias that Gardner’s does: teams using old and unreliable information concerning Granderson. But I’m not sure there’s enough evidence there to render that any more credible than any other pet theory.

On broadcasts this season the Yankees crew has often mentioned that the outfielders, not just Swisher, have worked with Larry Rothschild on their throwing. It makes perfect sense, of course, since outfielders want to generate power with their throws just as pitchers do. While it’s an anecdote, it apparently shows up in the data as well. Whatever the case, the Yankees starting outfield is not only doing an excellent job of running down fly balls, but they’re also holding and killing base runners with efficiency. After years and years of watching one of the poorest outfield defenses in the league, it’s nice to finally see the Yankees on top.

Categories : Defense
Comments (52)

At the Wall Street Journal this morning, writer Daniel Barbarisi takes a look at Brett Gardner‘s range in left field. It’s subscriber-only content, but there are apparently ways to find it free it you search hard enough. Here’s something that caught my interest from the article:

He is effectively a second center fielder, ranging wide over the left side of the field in ways no other left fielder is doing. He frequently takes balls away from center fielder Curtis Granderson, when traditionally, it’s vice-versa…

Gardner teams with Granderson and Nick Swisher to create one of the best defensive outfields in baseball. Granderson is an established, rangy center fielder who has great in-line speed once he gets moving, and Swisher is an underrated and improving right fielder—his UZR is 10.7, fifth-best in baseball. And they move around significantly, adjusting for where they expect the hitter will place the ball…

The way Gardner covers ground allows the Yankees to use different defensive alignments, shifting Granderson more toward right field in some situations because they assume Gardner can cover all of left-center.

Jay Jaffe has speculated before that Gardner takes balls away from Granderson, and so it’s interesting to see Barbarisi essentially confirm this hypothesis. Like Jaffe, I wondered about Granderson’s poor UZR score since it doesn’t seem to pass the eye test and I’ve yet to find a single person who believes that Granderson is actually a poor fielder. It may simply be that Gardner’s speedy wheels and great instincts, and Granderson’s positioning, are the cause of Granderson’s subpar UZR score this year.

This is a relevant issue as it relates to Granderson’s MVP chances. In traditional categories, Granderson cleans up. He’s second in HR, first in R and RBI, and he’s stolen 24 bases. But in the advanced statistic realm of Wins Above Replacement, Granderson is held back by his poor defensive score. His -9.2 UZR rating means that he’s not as high up the Fangraphs’ WAR leaderboard as guys like Bautista, Pedroia and Ellsbury. Yet if we subbed in a value of 0 for Granderson’s UZR, still a conservative number in my estimation, his fWAR would go from 6.1 to 6.9. If we gave him last year’s value of 6.4 runs, his fWAR would go to 7.6, ahead of Pedroia, Ellsbury and Gonzalez and just a tenth of a point behind Jose Bautista. In other words, it’s possible that the case for Granderson winning the MVP should look even stronger than it currently does.

Anyway, the article is an interesting read and I recommend you take a look. There’s some cool stuff in there about how much Andruw Jones (himself a formerly-elite defender) respects Gardner’s defensive prowess, and also a fun quote about how much Gardner would love to win the Gold Glove. Parenthetically, Barbarisi has been a fantastic addition to the Wall Street Journal‘s coverage of sports. He’s been unafraid to integrate new statistics into his work without getting bogged down in explaining the stats and still maintaining the traditional feel of the newspaper sports column. If he isn’t on your radar by now, he should be.

Categories : Defense
Comments (70)

This post originally ran Saturday morning but quickly got buried by the news of Alex Rodriguez‘s torn meniscus, so we’re bumping back up because it’s really good and you should read it. Enjoy.

Recently Patrick Sullivan of Over the Monster and Baseball Analysts fame ignited a debate when he said the following: “You know who’s not as good as Dustin Pedroia? Like, not at all? Robinson Cano“. Them’s fightin’ words, pal. Sullivan later said that he dug in so stridently for fun on Twitter, but there’s an honest debate to be had here over the value of the two players. Is he right? Who is better, Cano or Pedroia? In order to answer the question, we need to evaluate all aspects to each player’s game: offense, base running and defense. We’ll run through each category, then examine the each player’s fWAR. We’ll also introduce a variation on WAR which I’ve lovingly dubbed RABWAR. Let’s get to it.

Offense: light tower power vs. the little on-base machine that could

Robinson Cano and Dustin Pedroia are both elite offensive forces at the plate. They just go about their business in differ manners. Cano is impatient. He rarely takes a base on balls, preferring to attack early in the count. As a result, he averages a walk rate of about 5% every year, a subpar showing. He makes up for this by hitting for average and for power. He’s a lifetime .308 hitter with a career slugging percentage of .492. The latter mark belies his true power skill, though. His power has been far more substantial in the past three years, and he’s slugged .520, .534 and .526 (including 2011).

For a second baseman, Cano’s power is superlative. Since 2009 his slugging percentage is .526, the highest in baseball among second baseman. The next closest is Chase Utley at .478. Cano also has the highest batting average among second baseman since 2009. Cano is the owner of a career .358 wOBA. Like his slugging, this mark is well below his totals in the past three years: .370, .389 and .375. It’s true that using 2009 as a start point is both arbitrary and favorable to Cano, but it’s also worth noting that he’s entering his physical prime. As a matter of true talent and future expectations, his 2009-2011 data would seem to be more relevant than what he did in his early 20s. This is the book on Cano: an elite hitter with poor on-base skills but who hits for average and power better than nearly anyone at his position.

Dustin Pedroia is a different animal. Like Cano, Pedroia hits for average (career .301 hitter). He’s also shown a decent amount of power with a .455 career slugging percentage, although this is well below Cano. Where he really sets himself apart is his on-base ability. Pedroia’s career walk rate is almost 10%, and this year he’s notched a 15% mark. He’s very patient at the plate and is extremely difficult to strike out, although he’s struck out more recently. Over the past 3 years, Pedroia has an on-base percentage of .376, a mark second only to Chase Utley’s .391. Overall, Pedroia has a career wOBA of .366, .08 points higher than Robinson Cano. Unlike Cano, Pedroia does not benefit from using a sample of only the past three years. His wOBA from 2009 to 2011 is .366, identical to his career average. Who’s the better overall hitter then?

As you can see, Cano has edged Pedroia out in wOBA since the start of 2009, but Pedroia has been more consistent since 2007. It’s also worth noting that Pedroia outperforms Cano slightly in wRC+, which is like a wOBA-based version of OPS+. Pedroia has a career mark of 120, and Cano’s career wRC+ is 118. In the past three years, Pedroia’s respective wRC+ marks are 113, 132 and 129. Cano’s are 121, 142 and 137.  In terms of overall offensive production, the two are very, very close. I’d like to give the category to Cano because of his tremendous upside, but his lack of a respectable walk rate means that his overall production is more likely to be the victim of the capricious whim of the BABIP dragons. This one’s a tossup.

Base running: don’t even think about it vs. the constant threat

Yankees fans know that Robinson Cano should never try to steal a base. He still tries though, and manages to swipe about 5 bases a year, giving him a career total of 26 stolen bases. He’s been caught a staggering 24 times though, meaning that his success rate is just over 50%. Pedroia is far better at stealing bases. He’s stolen 72 bases in his career and averages around 20 a year when he’s healthy. Unlike Cano, he hasn’t gotten thrown out that often – his total caught stealing  mark is 15, giving him a success rate of around 83%.

There’s more to base running than just stealing bases, though. For that we can turn to two very good base running stats, both of which attempt to quantify how many runs are contributed by a player’s advancement on the bases by considering ground, air and hit advancements. Baseball Prospectus’ version is EqBRR, short for Equivalent Base Running Runs. In addition to ground, air and hit advancements it also includes stolen bases and other advancements like wild pitches. Fangraphs’ version does not include these considerations. According to EqBRR, Robinson Cano has been worth only 1.2 runs on the base paths for his entire career, while  Dustin Pedroia has been worth 7.5 runs. This is despite the fact that Cano has played in over three hundred more games than Pedroia. It’s worth noting that Cano’s mark was negative prior to this season; he’s only in the black because he’s been worth 1.5 runs on the basepaths in 2011, bolstered by very high scores on ground and air advancement. In sum, by Baseball Prospectus’ measure Pedroia’s been worth about a half a win more than Cano on the bases.

Fangraphs’ base running stat is UBR, or Ultimate Base Running, and you can read about here. This metric grades Cano out much better than Pedroia, a surprising result. By UBR’s reckoning, Cano has been worth 4.1 runs on the base paths, while Pedroia has been worth -0.4. As mentioned, UBR does not include stolen bases, and we know that there’s a gigantic discrepancy between the two players when it comes to this factor. As such, EqBRR is probably a better indicator of base running value here, which means Pedroia gets the nod in this category.

Defense: depends on who you ask

It’d be really easy to provide the relevant UZR scores for each player and call it a day. It would also be incomplete. Astute readers know that there are some serious difficulties present in UZR and other defensive metrics. Baseball Prospectus’ Colin Wyers has been cleaning the glass like Dennis Rodman on the topic for quite some time now and has proposed an alternative, FRAA. For a primer on the issue, see this piece on the serious problems with most defensive metrics, this piece which summarizes the park-scorer and range biases problems and proposes a way forward, and this piece which examines FRAA against UZR on the topic of Derek Jeter. Colin Wyers summarizes FRAA accordingly:

Simply put, we count how many plays a player made, as well as expected plays for the average player at that position based upon a pitcher’s estimated ground-ball tendencies and the handedness of the batter. There are also adjustments for park and the base-out situations; depending on whether there are runners on base, as well as the number of outs, the shortstop may position himself differently, and we account for that in the average baselines.

The other metrics use other data to come to their estimate of expected outs—in the cases of UZR and DRS, it’s batted-ball and hit location data measured by BIS video scouts. In the cases of TZ and FRAA, it’s data collected by press box stringers working for MLB’s Gameday product.

So we have two different metrics both attempting to quantify defensive value, just in different ways. How do the two second-baseman, Cano and Pedroia, stack up against each other using UZR and FRAA? We’ll start with Cano:

Wowza. UZR hates Cano’s performance with the white hot intensity of a supernova, grading him out at -39.3 runs above average at second base. It’s given him a negative value for every year but 2007, although the worst scores came early in his career. The overwhelming majority of Cano’s poor UZR mark comes from his range. He grades out at nearly average in terms of double play and error runs above average, but has a -36.4 runs above average mark for range. Unlike UZR, FRAA is a huge fan, grading him at 31.2 runs above average. This is a difference of over 70 runs and clearly raises big questions. Other defensive metrics aren’t as harsh on Cano as UZR is, but none are as positive as FRAA. Where you come down on Cano’s defense, then, is likely informed by your own subjective evaluation from watching him. I’d split the difference. Cano certainly doesn’t strike me as a lousy defender, he gets to plenty of balls and turns a double play smoother than anyone. At the same time, I wouldn’t call him an elite defender. He simply doesn’t strike me as being cut from the same elite defensive cloth as someone like Adrian Beltre or Mark Ellis.

Like Cano, UZR and FRAA also see Pedroia differently. He grades out superbly by UZR’s standards, clocking in at 32.5 runs above average for his career, but looks far worse according to FRAA, scoring -1.2 runs above average. From a subjective standpoint, I’d argue that Pedroia is a very good defender. Whether he’s as good as UZR purports him to be is difficult to say. There are serious issues surrounding defensive metrics, so declaring a winner in this category is difficult. In this situation it’s wise to follow the advice of Tom Tango, who recommends we assume that all sides have something to add and take the midpoint. In that case, this category goes to Pedroia if only because of how poorly UZR grades Cano.

Conclusion: the final countdown

What WAR gives us is a systematic, consistent framework to value the accomplishments of players.  The good thing about a framework is that each person is free to create his own implementation.  Not all houses are built the same, but they all follow the same principle.  That’s what WAR gives us.” – Tom Tango.

Fangraphs’ WAR, which uses UBR for baserunning and UZR for defense, grades the two players accordingly:

By this standard, Pedroia is the clear winner. Give Pedroia some 1200 more plate appearances, and he would lead Cano by a wide margin. But as we know, fWAR relies on Fangraphs’ UBR and UZR. So let’s swap out UBR and UZR for Baseball Prospectus’ EqBRR and FRAA, respectively. We’ll call this little SABR-demon spawn RABWAR.

Here Cano is the clear winner, thanks largely to the difference in the way their defense is scored. So who is better: Cano or Pedroia? The offense is a tossup, the base running goes to Pedroia and the defense is a toss-up leaning towards Pedroia. At the end of the day, whether you pick Pedroia or Cano will likely hinge on which defensive metric you prefer, or which team you prefer. Cano and Pedroia are both incredibly talented second baseman and it’s tough to see any daylight between their two respective statistical profiles. In this sense, the claim that Cano is not “nearly as good” as Pedroia simply doesn’t ring true. If I was forced to pick between the two and was able to erase their prior team affiliations from my mind I’d likely go with Pedroia, in no small part because of my preference for his approach at the plate. It’s a very difficult choice though, unless I’m allowed to pick from the other division rival and take Ben Zobrist. Now there’s a second baseman.

Special thanks to Joe Pawlikowski and Moshe Mandel for their contributions to this piece.

Categories : Defense, Offense
Comments (135)
May
12

The Defense Question

Posted by: | Comments (17)

Just to make sure this is perfectly clear right up front: defensive statistics for the 2011 season are a long, long, looong way from being reliable. The sample size, 30-something games for all teams, is just way to small for the data to mean anything right now. That unreliability is part of the reason why none of the advanced metrics agree on the Yankees’ defensive performance.

The table above, screen cap’d from FanGraphs, shows the advanced defensive numbers for each American League team. I recommend clicking for a larger view, one that you can actually read. The Yankees are third in the league with a +8.6 UZR, meaning they’ve saved more than eight-and-a-half runs better than expected through 34 games. Aside from the outfielder’s arms (ARM), they’ve been better than average on double plays (DPR), at ranging for the ball (RngR), and when it comes to making errors (ErrR). With positive production in three of the four components, it’s no surprise their UZR ranks so high. Then why is their +3.9 UZR/150 just seventh in the league, right in the middle of the pack? It’s just a sample size issue, these numbers are very volatile right now.

John Dewan’s famed +/- system (DRS in the chart) says the Yankees are a dozen runs below average right now, fourth worst in the AL. Revised Zone Rating (RZR), which measures how many balls hit to a player’s zone were converted into outs, says the Yankees are the second best defensive club in the AL at .858. RZR is used in conjunction with Out of Zone Players (OOZ), which is the number of balls hit outside of a player’s zone that he turned into outs, but the Yankees are just sixth in the league in OOZ. DER is defensive efficiency ratio, and although it’s not in the chart, it’s easily calculated. It’s just 1-BABIP, telling you how many balls in play are converted into outs overall, regardless of where they’re hit. The Yankees are middle of the pack (ranking eighth in the AL, to be exact) at 0.716 DER.

Six different stats (though it’s really four since UZR and UZR/150 go together, as do RZR and OOZ) giving us five different approximations about the team’s overall defense. Some say it’s great, others say it’s terrible, others say it’s middle of pack. But like I said before, it’s just way too early in the season for these numbers to have any meaning. So the question to you is this: what is your take on the Yankees’ overall defense this year? Is it good, bad, average, or something else? I think it’s been slightly better than average mostly thanks to Brett Gardner, Curtis Granderson, Robinson Cano, Mark Teixeira, and Russell Martin (who’s reminded everyone what defense behind the plate actually looks like), but that’s just me. What about you?

Categories : Defense
Comments (17)

You might remember that a few weeks ago the Yankees blew a 4-0 lead against the Twins thanks in part to a bloop three-run double that eluded a sliding Nick Swisher, and you surely remember that Brent Lillibridge robbed the Yankees of two game winning hits just two nights ago. Mark Simon at ESPN took a look at those two balls in play using hang time, and found that within the last year, hits like the bloop double have fallen in for a hit approximately 56% of the time. The ball hung up in the air for 4.4 seconds, but it was just perfectly located. Swish wasn’t the only one unable to get there.

The ball Robinson Cano hit for the final out on Tuesday? That sucker had a hang time of just 2.5 seconds, and there have been 61 balls hit to that spot with that little hang time in the last year. Lillibridge was only the third outfielder to turn that ball into the out. That’s a .951 BABIP, but somehow the Yankees unlucked into small piece of the pie. Go figure.

Categories : Asides, Defense
Comments (10)
Mar
21

The Flip

Posted by: | Comments (17)

I thought this was pretty cool.

(via BtB)

Categories : Defense
Comments (17)
Mar
15

Shifting against Adrian

Posted by: | Comments (16)

(AP Photo/Charles Krupa)

Although he has yet to sign a long-term contract extension, it’s pretty much a foregone conclusion that Adrian Gonzalez will be a Boston Red Sox (Sock?) for a very long time. That’s bad news for Yankees fans, because we’re going to be stuck seeing one of the game’s very best hitters 18+ times a year for the better part of the next decade. I don’t think he’ll be the next David Ortiz (a.k.a. a stone-cold Yankees killer), but I don’t think anyone would be surprised if he turns into that guy.

Gonzalez has a reputation as an opposite field hitter and deservedly so. He’s hit 48 homers to the opposite field over the last three seasons, six more than anyone else in baseball. His .535 wOBA to the opposite field is second only to Ryan Howard’s (.606) during that time. Of course Gonzalez can certainly pull the ball as well, posting a .374 wOBA on balls hit in that direction since 2008. He’s just more productive when he’s driving the ball the other way.

The chart to the right shows the percentage of ground balls Gonzalez has hit to each field over the last three seasons. This is slightly different than the data found on FanGraphs’ splits page because their stuff is expressed as a percentage of the balls hit to that specific field. My rates are based on all balls put in play, regardless of field. Capisce? Capisce.

Anyway, the chart shows that nearly one out of every four balls Adrian has put in play since 2008 has been a grounder to the right side. Another 13.2% have been grounders back up-the-middle, so that’s more than 36% of his balls in play being hit on the ground between the shortstop’s approximate location and first base. For every 25 balls he’s put in play, just one is a grounder to the opposite field. Those are some drastic percentages, something the Yankees can exploit by using an infield shift.

On the flip side, here’s similar data for Gonzalez’s fly balls. Almost two out of every ten balls in play over the last three years has been a fly ball to left, plus another 13.1% to center. That means 31.1% of his balls in play have been a fly ball to left-center-ish field. Between that and the ground ball tendencies mentioned in the last paragraph, two out of every three balls Adrian puts in play are either a fly ball to left/center or a grounder to the right side/up-the-middle. Pretty good odds.

I’m not sure how teams have been playing Gonzalez over the last few seasons, but there appears to be a very real advantage to be gained by employing some … let’s call it “optimized positioning.” Based on Adrian’s batted ball tendencies outlined above, I’d align my fielders like so when he’s at the plate…

Obviously this isn’t terribly precise, so don’t kill me over the artwork. Rather, it’s just a general defensive alignment based on what Adrian is likely to do on a given ball in play. The brown splotches are non-catcher fielders by the way, and the original spray chart comes from Texas Leaguers. The quick-and-dirty explanation is that I would shade the center fielder over towards left-center while employing the full-blown infield shift. That means the second baseman on the right field grass, the shortstop playing where the second baseman usually is, and the third baseman playing short. If a ball is squibbed down the third base line, so be it. It’ll happen every so often, but you’re trading that occasional extra hit for (theoretically) several additional outs elsewhere.

The Yankees do have a nice advantage in the outfield when it comes to combating Gonzalez’s opposite field tendencies because Brett Gardner and Curtis Granderson can flat out get after balls. Nick Swisher, the least rangy of the outfield trio, will see the fewest fly balls off Adrian’s bat come his way. Perhaps Grandy’s and Gardy’s speed means they won’t have the shade towards left so dramatically.

Would the Yankees do something like this? Who the hell knows. I would expect to see the infield shift employed before the outfield shift, though I suppose there’s a chance they’d do both simultaneously. The important thing is that the pitcher keeps the ball down in the zone. If he doesn’t and ends up elevating a ball to Gonzalez, it’s not going to matter where the fielders are positioned.

Categories : Analysis, Defense
Comments (16)

(Kathy Willens/AP)

When the Yankees signed Mark Teixeira in December of 2008, it meant two upgrades for the Yankees. The first came on offense, where Teixeira’s bat would represent an improvement over the aging Jason Giambi. In his final two Yankees seasons Giambi’s numbers dropped a bit, and he came to bat only 868 times in those two seasons. Teixeira would bring not only a superior bat, but also durability. But the most significant upgrade came on defense. Giambi was known as a statue before he even signed his $120 million contract. Teixeira was considered one of the game’s premier defensive first basemen. I can’t count the number of times I said, “I sure is nice to have a real first baseman” in 2009.

Defensive metrics did not agree with what Teixeira’s reputation, and what our eyes, told us. In 2009 Teixeira produced a 0.9 UZR, which ranked him 12th in the majors. That might not have been as ridiculous sounding had Miguel Cabrera not finished with a 3.5 UZR, fifth in the majors. Much as our eyes can deceive us, I don’t think that they deceive us to the level it would require for Cabrera to be a better defensive first baseman than Teixeira. After the 2009 season I recall a lot of ill feelings towards UZR, because of Teixeira’s situation specifically. The stats did not match what our eyes told us, and so we blamed the stats.

In 2010 UZR ranked Teixeira a bit worse. He finished with a -2.9 UZR, 14th in the majors. There might have been a number of good defensive first basemen ahead of him, but it’s doubtful that he finished more than a win worse than, for instance, Ike Davis. Maybe Teixeira isn’t the league’s best defensive first basemen, but after watching him for over 150 games in each of the last two years, and watching him frequently enough during his pre-Yankees seasons, I’m fairly confident that he ranks in the top five.

While UZR is still a widely used defensive metric, it does contain flaws. Almost all defensive metrics will, since we’re still figuring out how to best quantify defense. Perhaps the most aggressive in the pursuit of fielding knowledge is Baseball Prospectus’s Colin Wyers. He has spearheaded BP’s effort to create a more effective defensive stat, and after reading a number of his columns on the topic I see his point. With observation stats such as UZR and DRS there can exist significant range bias. Total Zone, the fielding stat used on Baseball Reference, takes the observation out and instead uses the play-by-play logs to determine defensive value. It’s here that Teixeira excels.

FanGraphs just added Total Zone (with location) data for the 2010 season, so we can see where he ranks compared to his peers. Surprisingly to UZR, but unsurprisingly to Yankees fans, Teixeira finished with a 13 TZL, which ranks him second in the majors. The only first baseman to finish better was Daric Barton, and we know he’s a top-notch first baseman. In 2009 he had a 10.1 TZL, which ranked fifth. But instead of sitting behind Miguel Cabrera (-4.4), he was behind only Albert Pujols, Adrian Gonzalez, Todd Helton, and Lyle Overbay. That’s a list a bit easier to accept.

This isn’t necessarily an indictment of UZR. After all, the correlation between UZR and TZL in 2010 was .75, so they’re pretty close to one another. What it makes me wonder more than anything is why UZR views Tex so differently. Barton, for instance, led the league in both TZL and UZR. Of the players with worse than -1 UZR, all but two — Tex and Todd Helton — also had a negative TZL. What about Tex’s game causes UZR to rate him so poorly relative to what we see? I don’t have an answer, but I do hope that this sheds a little light on current defensive metrics. Maybe UZR isn’t flawed for everyone. Maybe its biases affect different players in different ways.

Categories : Defense
Comments (43)

Earlier today we went back and looked at the five biggest hits of the Yankees’ season using WPA, so now we’re going to go back and look at the five biggest outs/defensive plays of the campaign. You’re not going to see the same kind of huge win probability swings just because it’s very difficult for one defensive play to increase a team’s chances of winning that much. Sometimes though, getting that one out can be a whole lot more stressful than getting that one big hit. I hate to ruin the surprise, but the greatest closer of all team will be featured prominently…

(AP Photo/Andy King)

May 26th: Mo gets Denard Span to bang into a double play
Things weren’t going so well for the Yankees in late-May. Their inaugural trip to Target Field was coming on the heels of five losses in six games, and the team was struggling to score runs. Although this game started on the 25th, it was actually completed on the 26th because of a rain-induced suspension of play in the fifth inning. The Yankees were leading one-zip on a Derek Jeter solo homer when Mariano Rivera came to the mound in the ninth. He sat down J.J. Hardy to lead off the inning, but then walked pinch hitter Jim Thome, who was immediately replaced by pinch runner Alexi Casilla. Denard Span was at the plate as the winning run, but Casilla never attempted to steal second and get himself into scoring position. Mo got Span to ground the ball to second, resulting in a game ending 4-6-3 double play. The WPA of this play was 0.22.

August 11th: Mo gets Josh Hamilton (video)
The state of Texas was not kind to the Yankees in 2010, and in fact this game took place after David Murphy’s walk-off single the night before. With the Yankees up by a run to start the ninth, Elvis Andrus gave the Rangers some hope with a leadoff triple, putting the tying run 90 feet away with the heart of the order coming up. Michael Young popped up the first pitch of his at-bat to shallow right, too shallow for Andrus to score. Texas would have still been able to tie the game by making an out at this point, and they had the eventual MVP coming to the plate. To make matters worse, Rivera fell behind Hamilton 2-0. Mo gave him his trademark cutter, but Hamilton tapped the ball back to the pitcher, again keeping Andrus anchored to third. Because it took the opportunity to score a run on an out away, the WPA swing of Hamilton’s at-bat was 0.25. Vlad Guerrero grounded out to end the game two pitches later, but Hamilton’s out was key.

May 26th: Andy gets Joe Mauer to bounce into a twin killing (video)
A few hours after Span grounded into his twin killing, Andy Pettitte got the reigning MVP to do the same. The score was tied at two in the eighth inning, but Andy was still out there since his pitch count was barely over 80 (83 to be exact). Backup catcher Drew Butera led the inning off with a double, and he then moved over to third to when Alex Rodriguez botched a Span sacrifice bunt attempt. Runners were at the corners with no outs, and the meat of Minnesota’s lineup was coming to the dish. Orlando Hudson lined a pitch back to Andy for the first out, but like I said, it was just the first out. Pettitte fell behind in the count to Mauer, putting him one ball away from a bases loaded situation. Instead, Mauer made weak contact on a slider away, resulting in a garden variety 6-4-3 double and one amazing fist pump from the old man. Mauer’s GIDP resulted in a WPA swing of 0.25.

Oh yeah, and he got an at-bat. (AP Photo/Ross D. Franklin)

June 23rd: Mariano’s Mona Lisa (video)
Okay, I confess, this isn’t just one out, it’s three consecutive outs. But they all happened in the same inning, and they each resulted in one of the highest individual defensive WPA swings of the season. I figured it was only right to lump them together, since together they represent the mastery of Mariano Rivera.

The Yankees were in Arizona and up a run in the tenth inning after Curtis Granderson‘s solo homer, but things started to get tenuous rather quickly. Stephen Drew led off the inning with a single, then ended up at third after The Justin Upton doubled. With the winning run at second and a base open, Mo intentionally walked Miguel Montero (2-for-3 in the game and 13-for-33 with five doubles and two homers in his previous eight games) to create the force at any base with still no outs. This is when the master went to work.

The first out of the inning was completely harmless as Chris Young popped up a 1-1 pitch behind the plate, with Frankie Cervelli making the catch. Adam LaRoche followed Young, and after another 1-1 count, the first baseman popped the pitch up to third base, invoking the infield fly rule. Two men down, but the bases were still loaded. Mark Reynolds ended 54.7% of his plate appearances with a walk, a homer, or a strikeout in 2010, but only two of those would have been helped him in this spot. Instead, he went with door number three. Rivera got Reynolds swinging at a pitch off the plate, stranding all three runners and preserving the win. The WPA swings were rather remarkable: 0.20 (Young), 0.27 (LaRoche), and 0.28 (Reynolds). The last two were the third and second biggest outs of the season, respectively, and they came back-to-back. Oh, and that was Mo’s second inning of work on the night. Insanity.

(AP Photo/Chris O'Meara)

Sept. 14th: Mo to Golson to A-Rod (video)
Despite all of his success, Mo needed a little help in making the biggest defensive play of the 2010 season. Jorge Posada had given the Yankees a one run lead with a long solo homer in the top of the tenth, but Carl Crawford led off the bottom half with a single. As expected, he stole second base but only after Evan Longoria flew out to deep center. The tying run was in scoring position, and a blown lead would have been rather demoralizing since the Yanks were coming off four straight losses and seven in their last eight games.

Matt Joyce had proven to be a thorn in New York’s side earlier in the season, and this time all he needed was a little bloop or a seeing eye single to knot things up. Rivera ran the count full, and on the sixth pitch of the encounter Joyce lifted a routine fly ball to Greg Golson in right. Crawford tagged up and once the ball settled into Golson’s glove, he took off for third. Unfortunately for him, the outfielder decided to give everyone a free look at the gun show. He threw Crawford out the third, ending the game in perhaps the most unexpected way possible. The combination of the fly out and throw out at third resulted in a WPA swing of 0.29. That throw still amazes me.

Categories : Defense, Pitching
Comments (19)

(AP Photo/Kathy Willens)

Brian Cashman has been running the Yankees since 1998, but it wasn’t until the after the 2005 season that he gained autonomy and full control of the baseball operations. Ownership was constantly dipping its toe in the baseball ops pool before then, and the Tampa faction of executives and team officials were meddling as well. The three-year contract Cashman signed after 2005 changed all that, but he was still stuck with the same team. Jason Giambi was just four years into a seven year contract and bitter old Gary Sheffield was still around. Carl Pavano was still under contract, ditto Bernie Williams and Jaret Wright. As much as he probably wanted to, Cashman couldn’t just flip a switch to make these guys go away.

It took a few years for Cash to get rid of those guys and replace them with players he wanted, but by the end of the 2008 season the process was pretty much complete. Sheff, Bernie, and Wright were long gone, and the contracts of Pavano and Giambi had mercifully expired at long last. That allowed Cashman to seek younger players at several positions, and he did just that by acquiring Nick Swisher and signing Mark Teixeira. Robbie Cano had established himself as no worse than a legit everyday second baseman with the potential for more, and one of the outfield spots was going to Melky Cabrera or Brett Gardner, whoever happened to be playing better at the time. And that was just that one offseason.

As a result of all the new blood, the team’s defense improved. It was hard not to, frankly. The Yankees were probably the worst defensive team of the decade up to that point, and bringing in just average defensive players would have been a big time help. The table on the right compares the team’s defense from 2006 through 2008 (a period starting when Cashman got his autonomy and ending with a host of albatross contracts expiring) to the defense they’ve played since. I used UZR/150 instead of straight UZR because we’re dealing with a three-year sample vs. a two-year sample, so the rate stat makes more sense. In fact, they always do, but I digress.

The scary part is that this data does not include the 2005 Yankees, which may have been the worst defensive team in the history of baseball. If we include them, we’re looking at an improvement of 70.4 runs saved per 150 defensive games, a simply staggering amount. As it stands, the Yankees made what amounts to a five win improvement defensively thanks to the moves made in recent years, all because a few more of  batted balls are converted into outs on a nightly basis.

Just three of the team’s eight fielders on Opening Day 2006 were playing the same position on Opening Day 2009, though we really should consider it four because of Alex Rodriguez. He was on the shelf in early ’09 recovering from his hip procedure, but he obviously would have been at the hot corner in Game One if healthy. The other holdovers were Jorge Posada, Derek Jeter, and Cano. The entire outfield alignment changed during that three year period as did the first baseman.

The change continues as well. The presumed 2011 Opening Day defensive lineup figures to have just four players playing the same position that they did on Opening Day 2009, and that includes A-Rod. The entire outfield alignment changed again, going from Johnny Damon-Brett Gardner-Xavier Nady (LF-CF-RF) to Gardner-Curtis Granderson-Nick Swisher. Jorge Posada has been replaced behind the plate by Russell Martin. Just the infield remains intact, as they will for at least the next three seasons.

The cool part is that the Yankees made all this defensive improvement without sacrificing offense. In fact, they actually got better with the bats. They led baseball with a .353 wOBA from 2006 through 2008, an offense that was 15% better than league average according to wRC+. That improved to a .356 wOBA in 2009 and 2010, 19% better than average. Younger, more athletic players led to better defense and even improved what was already the game’s best offense, who’d a thunk it?

Categories : Defense
Comments (49)