Prospectus Toolbox: Even Further Reading

One of my favorite parts of putting together the Toolbox every week has been assembling the ‘Further Reading’ sections at the end of the columns, which are a starting point for anyone who wants to look for more in-depth examinations of the week’s topic. Over the past couple of months, a number of readers have written in to suggest additions to the further reading lists, and I’ve stumbled on to a few worthy additions of my own. What this means is that from time to time I’m going to update the old columns by adding some additional research options, under the heading of “Even Further Reading,” with a bit of analysis and reader mail mixed in.

Prospectus Toolbox: Value Over Replacement Player

Keith Woolner, “Understanding and Measuring the Replacement Level” in Baseball Prospectus 2002 (Brassey’s, 2002): This one was an outright embarrassing omission, since this 2002 back-of-the-book Fungoes essay sets out a lot of the math and reasoning behind BP’s replacement player calculations.

Prospectus Toolbox: Reliever Tools

Michael Wolverton, “Making the Most of Your Inheritance”: This 2004 column talks about inherited runners, their rate of prevention, and the difference in opportunities between certain kinds of relievers.

Michael Wolverton, “The Tribe’s Quest for Glory-the Worst Bullpens of All Time”: The title pretty much says it all. At the time this article was written in June 2004, Cleveland’s bullpen looked like it could be historically bad, but the Indians relievers actually came around a bit (4.55 RA after the reader’s letter referenced in the article, compared to 7.47 RA before) and actually wound up being a run-of-the-mill lousy bullpen (-12.1 ARP, fifth-worst in the majors that year) as opposed to an all-time awful unit.

So, who actually had the worst bullpens ever? Since 1959, here are the lists by ARP and WXRL:

              Relief        Fair                   Relief         Fair
Year Team      IP     ARP    RA     Year Team       IP     WXRL    RA

1990 Braves   447.1  -78.4  6.01    1999 Royals    485.1  -7.586  6.63
2003 Royals   515.1  -77.4  6.44    1973 Braves    385.2  -6.542  5.79
1993 Rockies  552.2  -74.9  5.96    1990 Braves    447.1  -4.621  6.01
1977 Braves   478.1  -74.4  5.95    1971 Yankees   288.0  -3.859  5.37
1975 Cubs     420.2  -74.3  5.99    1967 Astros    394.0  -3.511  5.23
2005 D'backs  463.0  -72.3  6.16    1993 Mets      373.2  -3.059  5.41
1962 KC A's   413.1  -72.2  6.26    1974 Angels    294.0  -2.940  4.91
1966 Red Sox  495.0  -67.0  5.27    1978 Mets      417.0  -2.929  4.20
1996 Tigers   567.0  -65.5  6.79    1992 Phillies  422.1  -2.608  4.86
1999 Mariners 471.0  -64.6  6.80    1969 Phillies  384.1  -2.597  5.09

You should remember WXRL and ARP from my original article. Fair RA (technically, FAIR_RA_RELIEF) is another reliever metric which incorporates a pitcher’s inherited/bequeathed run performance into a statistic like ERA. Focusing on the main statistics, it’s amazing that only one team overlaps between the two lists, but that shows the big difference between the big two bullpen measures. The 1990 Braves featured the horrifying relief stylings of Rick Luecken, Mark Grant, and Joe Boever, and seems to be the consensus worst ‘pen of the past 48 years, both for their bad performance in big game situations, and their inability to prevent runs overall.

Prospectus Toolbox: Dying Quails and Pitchers BABIP

Erik Allen, Arvin Hsu, and Tom Tango, “Solving DIPS”: This collaboration examines the relative contributions of pitching and defense to hit prevention.

Mitchel Lichtman, “DIPS Revisited”: This 2004 study looks at batted ball types and their respective hit percentages and correlations. Hat tip to reader T.T. for both of the above links.

Clay Davenport, “Minor League Batting Averages on Balls in Play”: This study of BABIP in the minors showed that back when they were down on the farm, major leaguers as a group seemed to have greater control over the rate that balls in play became hits than their fellows who didn’t make it up to The Show.

Tom Tippett, “Can Pitchers Prevent Hits on Balls in Play?”: A 2003 study that looked at BABIP by breaking pitchers down into categories by pitching style-power pitchers, control pitchers, “crafty lefties”-to see the influence that each style had on balls in play.

Prospectus Toolbox: Measuring Team Defense

James Click, “A TAD Here or There-A Closer Look at Team Defense”: A further refinement of the PADE system, this time controlling for pitching factors to isolate defense to produce Team Adjusted Defense.

Reader M.G. had a conceptual problem with my description of PADE a few weeks back:

I’m having a little trouble wrapping my head around PADE for the following (not very thought out) idea: yes, it’s harder to convert hit balls into outs in Fenway Park than in Shea Stadium because, as you mention, Fenway’s got very little foul territory and a big wall getting in the way in left field while Shea has lots of space, and so we might give the Red Sox team defense a boost when we adjust for this. On the other hand, these park features create situations where it doesn’t really matter how good the defenders are, since neither Manny Ramirez nor a Gold Glover is going to convert a ball off the Monster or in the stands into an out, whereas in a big park you need guys who can cover a lot of ground and position themselves well to make outs. So while a pitcher’s park gives the defense more opportunities to make outs, it would also seem to place a higher premium on team defense. I sense there’s something contradictory between this idea and PADE’s corrections for hitter’s parks being harder places to make defensive outs, but can’t quite articulate it.

It’s a good question, albeit one that might be a bit too focused on the specific example that I gave in the column. A ballpark can boost batting average in a number of different ways: small foul territory, depriving the team of foul outs; hard turf that makes it easier for ground balls to skip past infielders; field anomalies, like the Green Monster; altitude, which can increase the speed and distance of batted balls…this list is hardly exhaustive.

Not all of those options indicate a smaller playing field-Coors field, for example, is pretty spacious. Regardless, the basic idea of PADE is to take that portion of the rate at which a team converts balls in play into outs that is attributable to defense, and separate that from the part that’s attributable to the ballpark. You’re right that a ball that’s, say, fifteen feet up the Green Monster is going to be a hit, no matter what. That’s exactly the reason such hits shouldn’t count against the defense.

Prospectus Toolbox: A Secret Affair

Tom Ruane, “RBI Production-A New Look at an Old Stat”: This study looks at RBI in terms of runs expected to produce a stat called RBI production.

Bill James, 1979 Baseball Abstract (Out of Print): As cited in part three of Rich Lederer’s excellent Abstracts from the Abstracts series, this book introduced what James called the Victory-Important RBI. I actually don’t haven’t had the chance to examine the formula for this stat-I’ve never read the ’79 Abstract-but it sounds like James’s idea (motivated by that particular bete noir of the RBI world, the Game-Winning RBI) was to add a leverage factor to RBI. So it was probably a bit like an offensive Win Expectation stat, just ignoring the win-enhancing effect of non-RBI events.

For sabermetric beginners in particular, the Abstracts from the Abstracts series is a resource that keeps on giving. So many of the tools used here and at other websites have their foundations in Bill James’s early work, and the Abstracts series does the invaluable service of summarizing this work, showing how these ideas developed and changed over the twelve years that James did the Baseball Abstracts.