The defensive charts for 2008 are now available here. These provide a nice visualization of the probabilistic model of range. Chase Utley’s charts, for example, shows how well he does on ground balls toward first base.

Dan Turkenkopf takes the data from the PMR’s look at defense behind the pitcher to see how many runs defense added to or removed from a pitcher’s record per nine innings.

Billfer tries to explain the poor defense behind Nate Robertson and decides that batters were just very good at hitting balls in holes against the Tigers pitcher.

As we know, a pitcher’s ERA can be influenced by the defense behind him. This posts explores which pitchers were helped or hurt by their defenses based on how well fielders turned balls in play into outs based on how difficult they were off the bat.

Those are pretty impressive numbers for Daisuke Matsuzaka and Tim Wakefield. Not only were balls in play against them easy to field, the Red Sox did a great job of turning them into outs. For Yankees fans who are concerned about New York signing Andy Pettitte again, better defense would improve Andy’s runs allowed a great deal. CC Sabathia certainly benefitted from good defense in both Cleveland and Milwaukee, so teams looking to sign him should be prepared to send their best fielders out behind the lefty.
Brian Bannister shows how important defense is to a low strikeout pitcher. His expected DER is low, and with the Royals doing a poor job fielding behind him, his actual DER was even lower. Bannister really needs to play for a team of defensive wizards.

The survey of the positions ends with the pitchers. First the teams:

The Blue Jays not only posted an excellent team ERA, but helped themselves defensively as well. The Brewers staff depended more on the fielders behind them. In looking at the individuals, experience appears to be a key to doing well:

Jesse Litsch is young, but 2-4 are all veterans, and Maddux seems to come out near the top quite often. Mike Mussina, however, did not appear to deserve his gold glove. Remember to take this ranking with a grain of salt, since pitchers are not in the field that often compared to position players, and get many fewer chances to field balls. Luck is a much bigger factor in this group.

Dan Turkenkopf calculates runs allowed or saved for leftfielders based on their PMR stats.
Update: Here are the runs for first basemen as well.

Our survey of the positions continues with catchers. Catchers don’t field very many balls, so take these rankings with a grain of salt. First, the teams:

It looks like the Mets trade for Brian Schneider turned out to be a good one from a defensive standpoint. It also looks like rookie of the year Geovany Soto might have some things to learn behind the plate. Let’s look at the individuals:

Here’s another reason Joe Mauer gets my vote for AL MVP. He’s not only a great offensive catcher, but he fields his position well also. For those teams interested in Jason Varitek, his ranking here is certainly another strike against him.
It’s also interesting to note that Jarrod Saltalamacchia wasn’t terrible behind the plate. I know there’s much more to the position than the ability to field, but in this regard, Saltalamacchia shows a positive behind the plate.

The Probabilistic Model of Range survey continues with first basemen:

As you might expect from the ranking of the top two National League teams, Pujols and Berkman competed with the glove as well as the bat:

Year after year Albert Pujols shows his defensive skill at first base. Lance Berkman is up there, also, making the MVP argument between the two that much closer. Mark Teixeira also offers an excellent glove to go along with his fine offense.
The most surprising ranking to me, however, is Justin Morneau. Justin is still young and shouldn’t have lost a step. He’s someone worth looking at in more detail. Of course, at the very bottom is Mike Jacobs, giving Royals fans another reason to dislike the trade.

Dan Turkenkopf list PMR Runs for rightfielders.

Dan Turkenkopf continues to translate PMR stats into runs, taking on third basemen this time.

The survery of range continues with leftfielders. The following table shows how the thirty teams fared at the position:

As in rightfield, there doesn’t seem to be a huge correlation between doing well in left and winning. The Royals displayed the best defense at the position, while the Phillies came out at the bottom of the pack.
The list of individuals in left shows that very few teams employ a regular at the position:

Ryan Braun is the first player on the list on the field in left for over 3000 balls in play. Some of this was caused by injuries (Soriano, Matsui), but for the most part, managers mix and match at the position. The move to left was clearly the right one for Braun.
The other rankings of note belong to Manny Ramirez and Jason Bay. Manny actually did better than Jason in 2008. I’m going to need to break down the two by team to see how much the parks might have made a difference. Bay certainly looked better than Manny watching him play for the Red Sox.
All those late inning substitutions Charlie Manuel made for Pat Burrell looked proper, also. Pat ranks as the worst leftfielder in baseball in 2008, so it’s no wonder Charlie wanted a better glove in left when the Phillies had the lead late.

The Probabilistic Model of Range reports continue with rightfielders. First, the team data:

It seems rightfielder defense didn’t have that much influence on playoff teams. Five of the eight post-season teams finished in the bottom half of the majors. Here’s a look at the individuals:

Denard Span not only improved the Twins leadoff slot, he also did a great job tracking down balls in rightfield. While I’m not surprised to see older players like Ken Griffey and Bobby Abreu near the bottom of the list, I didn’t expect to see Gary Matthews, Jr. there.
Ichiro Suzuki also adds some interest. He came out near the top in center, but in the middle in right. It’s a bit of a mystery why he does better in center than he does in right.

The Blue Jays and Cardinals made a challenge trade at the start of the season, swapping Scott Rolen and Troy Glaus. Defensively, at least, the Blue Jays came out on top. Here are the Probabilistic Model of Range team rankings for third base:

I’m impressed that the Braves rank so high. Chipper Jones isn’t known for his defense at third, but he played well this season.

Adrian Beltre comes through here as an interesting player. With the run up in salaries the last few seasons, his $12 million contract for 2009 is pretty good. It’s the same amount Mike Lowell will make next season. The two are also very close in terms of win shares. If a team is looking to upgrade their defense at third base, Beltre is a great pickup. The Mariners may be looking to move his contract. If the surgery he underwent in September helps with his hitting, he could be a very good pickup.
Evan Longoria certainly was a big part of the Rays defensive improvement. Unexpected major leaguers Blake DeWitt and Jack Hannahan also provided much needed defense. In addition, Andy Marte hasn’t hit but he did field the position well in limited playing time.
At the other end of the spectrum, Casey Blake appears to be as overrated defensively as offensively. Alex Gordon and Mark Reynolds look like they won’t have long careers at the position, given their poor play at a young age.

Mike Emeigh took me up on my offer to watch the best and worst plays of a second baseman. I sent Mike the highest probability plays Dan Uggla didn’t make, and the lowest probability plays Uggla did turn into outs. Here’s his report:

Of Uggla’s eight low-probability plays made, four of them were plays on which the Marlins had an infield shift on – three by Ryan Howard, one by Carlos Delgado – and in each case the ball was hit directly to where Uggla was playing; it’s very likely that any 2B would have made those plays in that position, and I don’t know thow much credit you want to give Uggla for being there. Two more plays – the 8/15 play against Alfonso Soriano and the 7/10 play against James Loney – also appear to be primarily due to positioning. On the former, Uggla was playing Soriano fairly far up the middle and had a good angle on his popup; on the latter, Uggla was pulled over fairly close to 1B. The other two plays were good, far-ranging plays – Uggla going far to his left to throw out Brian Schneider on 5/26, and ranging well to the left side of 2B in shallow center to grab a Mark Reynolds flare on 5/20 – although on the latter positioning also played a role, as Uggla was playing Reynolds up the middle and the CF was playing fairly deep. It was a good play, but one that a good CF or SS might have been able to make, also.
The six high-probability plays that Uggla should have made but didn’t:

• 6/17: routing GB by Ichiro, just booted it
• 8/6: hard-hit “at ’em” GB by Jimmy Rollins, basically one of those you either catch or don’t catch
• 5/30: GB by Pedro Feliz that took a bad hop as it got to Uggla; he barehanded it across his body, which threw off his timing
• 4/18: GB by Ryan Zimmerman toward the middle; Uggla made the play but didn’t get enough on the throw, which Mike Jacobs should have scooped anyway
• 9/21: With Jamie Moyer on 1B, Jimmy Rollins hit a slow grounder to the right side. Uggla stopped to avoid a collision with Moyer, which made him have to hurry the play when he did get to the ball. Had he kept coming Moyer would likely have collided with him, which would have been interference on Moyer
• 7/28: Endy Chavez hit a GB which took a funky hop as it got to Uggla, who booted it

Mike brings up something I’ve discussed before. Range is probably a poor word for what we’re studying here. Range isn’t just the ability to move a long distance to field a ball. It also includes the ability to position yourself (or have someone position you) so you don’t need to move very far. Uggla (and Utley) put themselves into position to field low probability balls without having too many high probability outs sneak through their vacated normal positions. Someday, we’ll measure range directly.

The third position presented in this year’s Probabilistic Model of Range study belongs to centerfielders. First, the overall team numbers:

It was a good year for the Mets and the Rays. I find it interesting that the Rangers rank so low, as a number of people noted Josh Hamilton’s defense this season. The next table will show where he ranks as an individual:

It was a very good year to be a centerfielder named Carlos. B.J. Upton, however, gets the nod as the best everyday DF. Looking at individuals, it becomes apparent why the Cardinals rated so poorly at the position. Skip Shumaker and Rick Ankiel were equally below average.
As for Josh Hamilton, he ranks 35th out of 44 fielders in the study. He’s going to be worth exploring in more detail, since I suspect people who watch him give him better marks.

Before I get to the centerfielders, a question arises sometimes that I’d like to address. I’m sometimes asked when a fielder does well, especially an outfielder on fly balls, what about ball hogs? So for outfielders, I’d like to take a look at where balls get hogged, and who does the hogging.
The first graph shows the percentage of plays made on each Probabilistic Model of Range vector, each vector representing about five degrees. Data is all outs made by major league outfielders in 2008. Leftfield is represented by low number vectors, rightfield by high number vectors. Straight-away centerfield is vector 36 (click for a larger image).

There are two things to notice from this graph. The first is that there are very few vectors in which ball hogging might occur. There are only six, in fact. Second, centerfielder hog more balls from leftfielder than they do from rightfielders. This makes some sense, since most hitters are right-handed, meaning centerfielders are going to be shaded toward left most of the time.
The other thing I want to point out is that balls are hogged in places where fewer outs get recorded (click for a larger image):

So, it’s tough for an outfielder to get a huge boost by ball hogging. They don’t stray that far into another’s territory, and when they do there are fewer outs to be gathered in anyway.
This also gives us a tool to use to look at individual teams if a question of ball hogging comes up.

Beyond the Boxscore translates PMR for shortstops into runs.

The following table shows how team second basemen ranked according to the Probabilistic Model of Range:

The Marlins number one at second base? That certainly flies in the face of Dan Uggla’s performance in the All-Star Game. It’s not that surprising, however, to see the Dodgers with the aging Jeff Kent coming in last. On to the individual players:

One thing I need to look at more closely is why Dan Uggla does so well. In the previous post on shortstops, a couple of commenters wanted more proof that this system actually works. I was a bit suprised by Akinori Iwamura rating so low, so I thought I would look at his poorest plays to see if they made sense. Of his four worst plays, all with a probablility of .889 or higher of being turned, two were errors hit right at him. One was a grounder to his right when he was playing too far left (poor positioning) and one was just bad judgement on a double play ball.
To compare, I looked at Utley’s best play, since he was the best regular at the position. All three of his best plays were balls to the right of first base that got by Howard off the bats of left handers. In each case, Utley ranged into the outfield to field the ball and throw out the batter at first, twice I believe to the pitcher covering. He made those plays because Howard couldn’t, but he was positioned so well he was in the right place to cover for Ryan.
The other thing I noticed is that toughest plays Utley made were much tougher than the best plays Iwamura executed. At the other end, easiest balls in play that Iwamura failed to turn into outs were much easier than Utley’s worse plays.
If anyone would like to review video on MLB.com for a particular player, I’ll be happy to send you the dates and innings of their best and worst plays.
In case you want to check my work, Iwamura’s worst plays were on 9/7, 3rd inning, 8/20, 9th inning, 4/25, 9th inning, 7/30, 5th inning. Utley’s best plays were on 7/23, 1st inning, 7/1, 1st inning and 8/3, 3rd inning.

Dan Turkenkopf at Beyond the Boxscore is translating PMR into runs so I don’t need to. Thanks, Dan.

The first position to examine is the most important of the fielders working behind the pitcher, the shortstop. As a reference, the first table looks at the position on a team-wide basis:

Notice the Mets are last. While recent articles mention Jeter as the worst fielder in the majors, in 2008 he wasn’t the worst shortstop in New York:

Jose Reyes converted 24 fewer balls into outs than expected while Jeter was just down four. It was actually one of Derek’s better years.
Omar Vizquel remains impressive at an advanced age. He didn’t play a whole season, so he didn’t get a chance to wear down, but if any team is looking for a great late inning defensive replacement, Omar is it.
Mike Aviles came up as a great find for the Royals in terms of his batting, but he also performed well with the glove.
Troy Tulowitzki went from first to almost worst in 2008. His injuries seemed to limit his range.
Jimmy Rollins won the Gold Glove today, but J.J. Hardy deserved it based on this data.
Feel free to comment and criticize.

Baseball Info Solutions sent me their fielding data, and that means it’s time to start presenting the 2008 Probabilistic Model of Range. If you’re new to this, you can find explanations in this archive. Basically, for each fieldable (non inside the park home runs) ball put in play, six parameters are used to determine how difficult it was to field the ball. A probability of turning the ball into an out is calculated, and those probabilities are summed. That gives us expected batted balls turned into outs. We turn that into a predicted DER (defensive efficiency record), compare that to the actual DER and calculate a ranking.
The model is based primarily on visiting player data, smoothed, using distance on fly balls and a hard hit indicator on ground balls. Only 2008 data was used to construct the model.
Note that a team can post a poor DER during the season, but do well in this model if the balls put into play were extremely difficult to field. This helps the Braves rank second.

The Rays turned in the best combination of good pitching and good defense. Their .699 predicted DER was second to the Mets. Unlike the Mets, however, the Rays fielded more balls than expected, giving the best DER, but only the third best Index. The Blue Jays turned in a tremendous defensive season, a big reason their pitching staff did so well in ERA in 2008.
The bottom of this chart is very interesting. From the Padres down, teams 23-30 all turned out to be very poor teams with the exception of the Yankees. Defense didn’t necessarily help a team win, as the Phillies were pretty middle of the road, but it certainly seemed to indicate a pretty bad team.
Note that last season, the Devil Rays were at the very bottom of the list. They improved both their predicted DER and their ability to turn batted balls into outs. That was enough to lower their runs allowed from 944 to 671 and make them American League champions.

Gideon Gil attended the AAAS symposia yesterday and gave both SAFE and PMR a nice writeup.

On Saturday morning I’ll make a presentation at the American Association for the Advancement of Science (AAAS) meeting on the Probabilistic Model of Range. The symposia is called New Techniques in the Evaluation and Prediction of Baseball Performance and meets in the Hynes Convention Center, Second Level, Room 202 at 8:30 AM. Alan Schwarz and Shane Jensen will also present.

The 2007 Defensive Charts are up. These provide a visualization of the Probabilistic Model of Range data based on position and batted ball type. Enjoy!