PNC

As you can imagine I've been spending some time in Pittsburgh over the last few weeks and so just a couple of photos today. The first looks back at PNC park from the Roberto Clemente bridge and the second is of the Willie Stargell statue that was unveiled in 2001 and sits on the Federal street (left field) side of the ballpark.

Heading to Steel Town

Today my column on Baseball Prospectus did not feature any new analysis or the usually endless tables overflowing with numbers and headed with three-letter acronyms. No, today marked the 100th and (at least for now) final Schrodinger's Bat column, marked with the title "Opus 100".

The column does not require a subscription and so you can read it for yourself but in the final section I briefly discuss the reason that this centennial edition is the last in the series.

To borrow from that bit...I'm leaving BP to join the front office of the Pittsburgh Pirates to become their Director of Baseball Systems Development. In that capacity I'll be assisting the excellent staff--including Kyle Stark, Bryan Minniti, Greg Smith, and Joe Delli Carri, under the direction of General Manager Neal Huntington--in building systems to support and inform the decision-making process of the baseball operations staff. All of those individuals mentioned, and many others, have made me feel more than welcome, and I'm thrilled to start the process of integrating the array of quantitative and qualitative information in a way that makes both even more instructive.

While I'm very excited at the new challenges and opportunities, I'm also a little sad as I say goodbye to Compassion International, where I've worked as a software architect for the last three years. Beyond the successes of the team in building a service oriented infrastructure that I've been blessed to be a part of, working at Compassion has truly been a wonderful experience. The dedication, intelligence, and vision of the organization and its employees to release children from poverty in Jesus' name is truly inspiring. It really is a great place to work and hey, I hear they might be looking for someone :)

And Even More Refinements in SFR

As some readers will know I've been writing the last couple weeks here and on Baseball Prospectus about a play by play fielding system termed Simple Fielding Runs or SFR.

After last week's column I had a chance to tweak the system yet again, and made the following changes.

Took into consideration the additional context of whether first base was occupied for all infielders. It turns out that for first baseman on ground balls hit by lefties the percentage of runners who reach (hit or error) goes from 16% to 24% when first is occupied, and from 20% to 29% for right-handers. For second baseman in the same scenarios it goes from 27% to 29% and 30% to 34%, for shortstops, interestingly, the trend is the opposite as it goes from 34% to 30% and 32% to 31%, and for third baseman it's pretty steady at 27% to 27% and 24% to 26%. Although a smaller percentage of fielded balls, the differences are significant for bunts when there is a runner on first since the vast majority of such attempts are sacrifice bunts instead of bunt hit attempts which have a higher success rate. The end result, I'm hoping, is a higher correlation with UZR for first baseman. I have not, however, taken into consideration the ability of first baseman to catch throws from infielders.

Changed the partitioning rules for which balls are assigned to first and third baseman. In the past I was partitioning all balls in the shared areas of responsibility except for middle infielders all line drives that resulted in extra base hits. Now I also exclude bunts from the calculation as well and assign all extra base hits on ground balls to left field to the third baseman and all extra base hits on grounders to right to the first baseman on the assumption that they're hit down the line. I also no longer partition any fly balls or line drives to the outfield since by doing so we didn't seem to be adding much information to what we already knew. Finally, in the last column on the subject I noted that I used a 50/50 split on ground balls up the middle. After the article was submitted, however, I realized that there is a simple way to partition balls between the short and second area of responsibility following the same principle outlined in the column. That change has been made.

While I haven't re-run the correlations with UZR I did want to show a few results and so here are all infielders who were assigned 15 or more balls in 2007 in the National League Central. You'll notice that the range is smaller with Milwaukee's infield not looking quite so terrible as in previous versions of the system.

Team    Name                  POS   Balls     SFR
CHN     Carlos Zambrano         1      33     0.6
CHN     Jason Marquis           1      30    -0.4
CHN     Rich Hill               1      26    -1.7
CHN     Ted Lilly               1      30    -1.8
CHN     Sean Marshall           1      17    -2.3
CHN     Jason Kendall           2      17    -0.5
CHN     Michael Barrett         2      23    -0.9
CHN     Daryle Ward             3      21    -1.5
CHN     Derrek Lee              3     294    -1.8
CHN     Ryan Theriot            4     102     1.3
CHN     Ronny Cedeno            4      18     0.9
CHN     Mike Fontenot           4     200     0.1
CHN     Mark DeRosa             4     295    -0.1
CHN     Aramis Ramirez          5     406     7.1
CHN     Mark DeRosa             5     103     4.9
CHN     Cesar Izturis           6     189     2.3
CHN     Ryan Theriot            6     405     2.1
CHN     Ronny Cedeno            6      48     2.0
-------------------------------------------------
CIN     Bobby Livingston        1      17     1.4
CIN     Aaron Harang            1      25     0.8
CIN     Bronson Arroyo          1      32    -0.4
CIN     Matt Belisle            1      18    -0.9
CIN     Javier Valentin         2      21     0.0
CIN     David Ross              2      49    -0.5
CIN     Scott Hatteberg         3     167     1.2
CIN     Jeff Conine             3      91     0.1
CIN     Jorge Cantu             3      17    -0.9
CIN     Joey Votto              3      32    -1.1
CIN     Jeff Keppinger          4      16     0.3
CIN     Brandon Phillips        4     662   -10.0
CIN     Jeff Keppinger          5      34     1.8
CIN     Ryan Freel              5      52     0.5
CIN     Juan Castro             5      21    -1.5
CIN     Edwin Encarnacion       5     402   -10.5
CIN     Alex Gonzalez           6     420     2.0
CIN     Juan Castro             6      44    -0.5
CIN     Jeff Keppinger          6     184    -1.6
CIN     Pedro Lopez             6      41    -1.7
-------------------------------------------------
HOU     Chris Sampson           1      25     1.9
HOU     Roy Oswalt              1      38     1.2
HOU     Wandy Rodriguez         1      22     1.0
HOU     Woody Williams          1      40     0.2
HOU     Dave Borkowski          1      21     0.0
HOU     Brad Ausmus             2      35     0.1
HOU     Eric Munson             2      18    -0.2
HOU     Mark Loretta            3      51     2.9
HOU     Mike Lamb               3      67    -1.1
HOU     Lance Berkman           3     277    -3.5
HOU     Chris Burke             4     150     2.5
HOU     Mark Loretta            4      77    -0.3
HOU     Craig Biggio            4     398    -9.7
HOU     Ty Wigginton            5     143     1.7
HOU     Mark Loretta            5      48    -0.5
HOU     Mike Lamb               5     144    -3.5
HOU     Morgan Ensberg          5     186    -9.0
HOU     Adam Everett            6     281     9.4
HOU     Eric Bruntlett          6     180     0.3
HOU     Cody Ransom             6      35    -0.1
HOU     Mark Loretta            6     248    -2.2
-------------------------------------------------
MIL     Jeff Suppan             1      35     2.1
MIL     Chris Capuano           1      30     1.1
MIL     Dave Bush               1      26     0.4
MIL     Claudio Vargas          1      15     0.2
MIL     Johnny Estrada          2      38     0.8
MIL     Tony Graffanino         3      25     0.7
MIL     Prince Fielder          3     325    -9.7
MIL     Craig Counsell          4      84     3.0
MIL     Tony Graffanino         4      92     0.5
MIL     Rickie Weeks            4     439   -21.4
MIL     Craig Counsell          5     112     3.9
MIL     Tony Graffanino         5      74     2.2
MIL     Ryan Braun              5     309   -24.7
MIL     Craig Counsell          6      79    -0.1
MIL     J.J. Hardy              6     619    -1.7
-------------------------------------------------
PIT     Matt Morris             1      18     1.6
PIT     Ian Snell               1      20     1.6
PIT     Paul Maholm             1      35     0.8
PIT     Shawn Chacon            1      20     0.4
PIT     Tom Gorzelanny          1      28    -0.4
PIT     Zach Duke               1      23    -0.6
PIT     Ronny Paulino           2      42     0.7
PIT     Adam LaRoche            3     327     3.0
PIT     Josh Phelps             3      25     1.2
PIT     Matt Kata               4      18    -1.9
PIT     Jose Castillo           4      56    -5.7
PIT     Freddy Sanchez          4     526   -13.0
PIT     Jose Castillo           5     130     2.6
PIT     Matt Kata               5      23     0.1
PIT     Jose Bautista           5     410    -4.8
PIT     Cesar Izturis           6     107    -0.5
PIT     Jack Wilson             6     665    -0.6
PIT     Jose Castillo           6      32    -0.6
-------------------------------------------------
SLN     Adam Wainwright         1      29     1.8
SLN     Braden Looper           1      21     0.4
SLN     Joel Pineiro            1      15     0.2
SLN     Kip Wells               1      23    -0.5
SLN     Yadier Molina           2      34     0.9
SLN     Gary Bennett            2      17     0.1
SLN     Albert Pujols           3     396    10.6
SLN     Brendan Ryan            4      64     1.3
SLN     Miguel Cairo            4      19     0.6
SLN     Aaron Miles             4     242    -0.8
SLN     Adam Kennedy            4     333    -2.8
SLN     Scott Rolen             5     351    13.7
SLN     Scott Spiezio           5      68     2.4
SLN     Russell Branyan         5      28     1.6
SLN     Miguel Cairo            5      31     1.5
SLN     Brendan Ryan            5      72     0.3
SLN     Brendan Ryan            6      91     3.7
SLN     Brian Barden            6      17     1.0
SLN     David Eckstein          6     493    -7.2
SLN     Aaron Miles             6     155    -8.6

A Pair of GM Interviews

One of the interviewers I always make a priority to read, and I'm not just saying this because he's a colleague, is David Laurila at Baseball Prospectus. Whether it's catching up with a former player, talking to a scouting director, or even a general manager as he did last week with the Pirates new GM Neal Huntington (pictured right), Laurila is a must read.

What makes Laurila's interviews so good, and the skill most important for any interviewer I would guess, is that I find he asks the questions I would ask only in a way that never seems to offend the interviewee and cause them to shut down. And of course he has a great way of transitioning from one topic to another making the interviews pleasant to read.

That said, here's a just a brief excerpt from the Huntington interview that illustrates both skills as Laurila both asks about what can be a hot button topic in statistical analysis and does so by piggy-backing off of a series of questions on the Pirates new manager John Russell.

DL: In a recent interview you talked about how the front office will be utilizing statistical analysis to evaluate talent. To what extent do you expect John Russell and his coaching staff to utilize that type of information?

NH: From an advance scouting standpoint, to try to gain an advantage, it will be vital. John and I have talked about that extensively, and we expect him, along with his hitting and pitching coaches, to use that information objectively and creatively. During the course of the season I expect that John and I will on the phone frequently, if not daily. We’ll talk about a multitude of issues, with the use of statistics being one of them.

DL: In the same interview, you said, “we’ll rely on the more traditional objective evaluations like OPS, WHIP, Runs Created, Component ERA…” Have we reached the point where those metrics can now be considered traditional?

NH: I think so. Cleveland--led by Chris Antonetti and Mike Chernoff--does an excellent job of objectively evaluating players, and we hope to build upon that model in Pittsburgh. We look at “traditional” metrics as well as more advanced metrics any time we are breaking players down. As a matter of fact, as an organization I want us to go to the next level of metrics. I hope that we can get to the point where we’ve built an environment to create and utilize them in-house. We have a lot of work to do, but we have an outstanding statistical consultant on board, and we’re creating a computer system to help us not only store, track and access data but to also analyze the data that has been collected.

DL: Can you touch briefly on some of the more advanced metrics you alluded to, and the role you expect them to play in your evaluation and decision-making process?

NH: In terms of advanced metrics, we are trying to utilize equivalences and adjusted performance numbers as well as percentile ranks to deepen the objective evaluation of the player. An interesting development has been the objective evaluation of defensive performance and the overall impact on a player’s value. The overall role will be as a supplement to our subjective evaluations to provide us with the best information to increase our probability for success.

So here we learn that Huntington is being aggressive in pursuing analysis to the extent of commissioning the construction of a new system for the organization, that it incorporates advanced metrics, and that at least in part it deals with advanced scouting data, an area that you don't always hear about in terms of analysis where the focus is often more on player projections and translating performance across levels of competition. And so it would appear that Pittsburgh is entering the realm of the more sophisticated teams in terms of quantitative analysis. The rest of the interview is enlightening as well with thoughts on player development and a couple specific prospects.

In a second GM interview, Baseball Digest Daily's Joe Hamrahi hits the high points in an interview with the Dodger's Ned Colletti. As you might imagine the focus was on the cadre of young players the Dodgers have and what their opportunities might be in 2007 given the unfortunate decision in the 2006 off-season to load up on veterans at their expense. When asked what he might have done differently in 2007 given the chance Colletti said...

That’s a great question. I’m sure there are (some things). One thing I know I wouldn’t have done differently. I’m glad I didn’t trade 3 or 4 kids at the deadline. I think that decision will prove to be valuable for a very long time. People ask about the young players all the time. Teams need young players.

Well, yes, one would think that was obvious and of course by the trading deadline there were few people encouraging the Dodgers to give up Andre Ethier, James Loney, Matt Kemp, Andy LaRoche, or Chad Billingsley for a rent-a-player. Earlier in the interview Colletti seemed very excited about the finish of Loney in particular but there too it's not as if there weren't solid reasons to believe Loney would be a contributor long before that as indicated by his PECOTA projection and his .385/.442/.603 line before the break in limited time.

But now it appears as if either Kemp or Ethier will lose time in the wake of the signing of Andruw Jones since one doubts that Juan Pierre and his large contract will be the victim. Sometimes injuries have a way of making these kinds of problems moot as the season approaches but it'll be interesting to see what Colletti's next move is.

Finally, in relation to the signing of Jones some have wondered what the difference between Jones and Pierre might be in terms of opponent's baserunning given that Jones has among the best and Pierre the worst reputations for their throwing arms.

In an essay titled "Expanding the Cannon: Quantifying the Impact of Outfield Throwing Arms" for this year's Baseball Prospectus annual, I take a look at exactly that question and the results for Jones and Pierre confirm the common wisdom. In 2007 Jones was worth +4 runs over average in terms of throwing runners out and stopping them from advancing. The combination of his positioning and arm meant that runners advanced just 31% of the time, tops among centerfielders with 100 or more adjusted games played at the position. On the contrary Pierre found himself at the other end of the spectrum and rated a -7 for his arm as opponents advanced 47% of the time - easily the worst at his position. Suffice it to say that 2007 was not an aberration as the essay also includes 2005 and 2006 leaderboards and totals for the same time period. By simply replacing Pierre with Jones (and not considering how much Pierre might play in left where his arm will hurt the team far less), the Dodgers stand to gain around a win only considering outfield throwing. When you add the larger offensive impact, you're talking about a significant difference...