Tuesday, December 27, 2005

The Hardball Times

Hello again. I said before that this would be a "review" of the 2006 Hardball Times Baseball Annual, but that's no exactly true. Actually, I'm going to look at some of the articles in it and the ideas expressed and see what there is to learn.
The annual begins with a look at each division and what went on as the season unfolded. They use statistics to tell the story of the season, and while I don't always agree with their conclusions, they are approached with an admirable degree of reason and good sense. The writers for The Hardball Times are a bizarre assortment; what you have are essentially big-time baseball fans and stat geeks from across the country who contribute to this website as writers and analysts, mostly in their spare time. Some succeed moreso than others (such as Steve Treder's entertaining look at the NL West), but you really get the sense of a work in progress here. The Times doesn't offer pat truths or intractable absolutes; they are working to solve a problem and are simply seeking to make a genuine effort toward a solution. There is a sense of working through the problems in the book, which comes off not as amateur, but rather as honest and straightforward. No one here pretends to have the Holy Grail of statistical analysis, but their frank admission of such makes them much more useful than 100 self-righteous vapid-heads.
Rob Neyer gives a preview of his next book with an article on big mistakes made in the past season. It's not a matter of pure second-guessing or a generous helping of hindsight, but rather something about which people should have known better, even at the time the decision was made. A frank look at past mistakes is a brilliant way to acquire insight and to avoid repeating these mistakes; but insight is a difficult problem for baseball executives, sabermetric or no.
Neyer takes the Yankees and Red Sox to task (as so many have), but he makes a good point that no one else had considered: why in the world did the Phillies trade Placido Polanco? Yes, the Phillies had a young rookie in Chase Utley who was ready to play second. But the problem predates that issue. Neyer wonders why in the blazes the Phillies felt the need to sign David F'N Bell to a big contract when he's a pretty dreadful hitter. They could have simply moved Polanco to third, saved the money on Bell, and had both Polanco and Utley in the infield. It's a good point, and no one else has made it. Often we see a problem as it existed when it reached the crisis point, rather than looking back on how it could have been avoided long before.
Matt Welch, in the article Getting with the Program, defends the Los Angeles Angels, who are often the target of a good deal of scorn from sabermetricians. Welch does temper the negativity surrounding the Angels' philosophies but doesn't do a very good job of countering the negativity itself. He does point out the fact that the Angels' system has produced a good record of winning (compared to the sabermetrically-inclined A's), but doesn't successfully dispute the criticisms surrounding the team. You get a good sense that Welch is a fan of the Angels looking to excuse his team, rather than someone looking to make a straightforward argument, and this is reflected in his commentary.
On the other side of LA, Jon Weisman writes a fine article about Paul DePodesta's star-crossed time in Los Angeles. It's been commonly accepted since his firing that Paul was a failure, but Jon does a good job of pointing out that this isn't true. Without cheerleading for DePo, Weisman points out his level of success which, while not absolute, was enough to at least ensure his job. DePo was simply another scapegoat, one which the media were all the more ready to accept given his background.
Maury Brown previews the battle over the upcoming Collective Bargaining Agreement. The current CBA expires after the 2006 season, so get ready for some more titanic squabbles among millionaires. But Brown does a good job of examining the issue, coming to the basic conclusion that it won't be too bad this time around. Craig Burley and Thomas Ayers preview the World Baseball Classic. This is for another blog entry, but the Classic has taken some turns even since this article was written, and not looking like as golden an opportunity as people thought.
Bill James looks at the amount of young pitching used by the Royals. He wonders if the Royals used a historic amount of young pitching? While it wasn't historic, it was large, and James' hypothesis about bad teams using more young pitching was generally borne out. The most interesting bit from this article was James' look at the best pitching staffs by decade, in terms of RSAA (Runs Saved Above Average). The #1 pitching decade, so to speak, goes to the 1990s Atlanta Braves. Next were the 1940s Cardinals, whom James describes as "the first great farm system churning out arms like IHOP frying pancakes." Next are the 1940s Tigers (Hal Newhouser, Dizzy Trout), the 1900s Cubs (Mordecai Brown, Orval Overall, etc.) and the 1960s Cardinals (Bob Gibson and friends). But the 9th team on the list is the 1950's Boston Red Sox.
This is a bit of a surprise. Or, as James exclaims, "What? Ike Delock is a Hall of Famer?"
While it may be true that RSAA is missing something, James looks back at the team and discovers that this forgettable era in Boston history did have a very underrated pitching staff, led by Frank Sullivan, who had a couple damn good seasons (since forgotten). It's always fascinating when statistical analysis brings back an answer contrary to what you believe, and how you seek to reconcile the two.
James does another study of Bert Blyleven's career. Blyleven is, in my opinion and several others (James included) the best pitcher not in Cooperstown by far. The knock on Blyleven was that he didn't "match the effort" of the team. While Blyleven's pitching stats are very good(3.31 career ERA, 117 ERA+) , his won-lost record is just fair (287-250). Perhaps, as his critics have claimed, Blyleven didn't "match the effort." That is to say, did he win games 9-0 only to lose another one 2-1? Was he saving runs in games where it didn't mattered but giving them up in close games when they did?
James studies the issue and finds that this is somewhat true. He concludes that Blyleven did fail to "match the effort," although I don't really see that. It's really impossible to prove whether his record is the result of effort-matching or simply bad luck. Do pitchers really have the ability to say, "Okay, we're ahead by 9, so I can give up 6 or 7 runs?" Can they really control when they give up lots of runs? To a certain extent, yes, a pitcher can control whether or not to bear down on a hitter depending on the situation. But is this factor really significant enough to account for Blyleven's record? And if it is, can we prove that it's anything but bad luck on Blyleven's part that he won the easy games and lost the tough ones? I can't see that in James' research, although he still doesn't see it as enough to keep Bert out of the Hall.
John DeWan, former CEO of STATS, Inc. writes an article where he argues that there's nothing really magical about the 100-pitch limit. In fact, he finds more evidence that pitchers throwing more pitches tend to do better in the rest of the season, rather than tiring out. His article is counter-intuitive to what sabermetrics think, that throwing a lot of pitches in one start is generally bad on the pitcher. There's not enough in DeWan's article to truly isolate the factors he wants to, and still several explanations for his study other than the conclusion he comes to, but it is an interesting idea meriting another look.
Pitch counts are an example of stats taken way out of control. Just as Saves are now far over-valued beyond what they intend to measure, pitch counts have become used far beyond their usefulness. All pitchers are held to low pitch counts, which may or not be the best idea for them. Pitch counts are virtually the only significant measure taken to determine pitcher abuse, whereas other factors must be taken into account. Many old-timers grumble about all the pitch counts in baseball, and while they all come from a genuinely good idea, they are used far beyond their usefulness and practicality.
Perhaps the most interesting (and complex) data in the book are the statistics on batted balls. Thanks to Baseball Info Solutions, THT is able to determine many heretofore unknown facts, such as how often batters hit line drives, and what percentage of a pitcher's outfield flies become home runs. John Fox breaks down a team's basic offensive and defensive statistics and determines from these how much runs a team should have score and prevented. This helps eliminate a good deal of luck; hitters will no longer be punished for a line drive that was hit right at the shortstop, nor will pitchers be punished for the broken bat blooper that falls for a single. It determines the heart of what the team did based on the event itself, not based on the result of the event, which may or may not have been positive. When you break down an offense (or pitching staff, etc.) to these basic bits of information, you can determine just how good they should have been, based on these components. Fox then determines which teams were the luckiest, according to how much they outperformed their components. Some may argue that there's more than luck involved, and I suppose that's true to a certain extent, but this helps eliminate a lot of the luck and chance that we previously thought was uneliminatable (if that's a word). This tells us a lot about who got lucky, and helps us predict who might be in for a "luck adjustment" next year.
There are several other very fascinating articles: Dave Studenmund breaks down the park effects into much more useful detail; J.C. Bradbury and David Gassko ask how much control a hitter has over a batted ball and subsequently make us wonder just how much control a pitcher has over home runs allowed; Bradbury also shows us "PrOPS," a method of removing luck on the individual level which gives us a pretty good idea of who got lucky (and unlucky) in 2005 (wouldn't it be great to know that free agent Alex Gonzalez, a bad hitter in 2005, was actually lucky and might have been even worse than his numbers); Dan Fox introduces us to Incremental Runs, a way of measuring who the good baserunners are beyond steals; and Dave Studenmund finishes it off with Net Win Share Value, which basically shows how much teams over-(or under-) paid their players, based on their actual Win Share value and their circumstances. Obviously, the best value is a rookie making the league-minimum who plays like an MVP; but that player can't make $15 million, and NWSV adjusts for that. Even a big-name free agent can be a great value if he produces far and above how he's getting paid, based on his status as a free agent (or arbitration-eligible player). The best and worst values in 2005: (WSAB refers to Win Shares Above Bench, or Win Shares above what a replacement-level AAA player would produce)
Jason Bay, PIT (Salary: $355,000) WS: 34 WSAB: 20.4
D. Willis, FLA (Salary: $378,500) WS: 26 WSAB: 19.2
T. Hafner, CLE (Salary: $500,000) WS: 27 WSAB: 18.9
Derrek Lee, CHC (Salary: $7.6 million) WS: 37 WSAB: 24.1
D. Ortiz, BOS (Salary: $5.25 mil.) WS: 31 WSAB: 22

Lee shows up ahead of lower-cost players because NWSV accounts for the real world; getting a player like Lee for $7 2/3 million on the free agent market is just as valuable as developing a rookie like Bay who makes the league minimum.

S. Sosa, BAL (Salary: $17 mil.) WS: 4 WSAB: -3.3
B. Bonds, SF (Salary: $22 mil.) WS: 2 WSAB: 1.5
K. Brown, NYY (Salary: $15.7 mil.) WS: 0 WSAB: -2.2
J. Bagwell, HOU (Salary: $18 mil.) WS: 3 WSAB: 1.3
C. Park, TEX-SD (Salary: $15 mil.) WS: 5 WSAB: 0.5
This is why you don't sign old players to big, back-loaded, long-term contracts. Sure, Bonds and Sosa were worth that much money when they first signed the deal, but as you approach 40, nobody's worth that much money. Teams regret adding that one extra year to the contract, and so will many teams that signed long-term contracts this year, thinking: "Oh, what's one more year, anyway?"

THT also publishes many useful (and hard-to-find) stats in the back of the book.
Hopefully, this book will become an annual event, much-anticipated in the sabermetric community. THT is a little-known website not affiliated with anyone at all, but they've built up quite a bit of influence and respect in the sabermetric community. Getting people like Rob Neyer and Bill James to contribute to your annual is a pretty impressive sign, as THT is becoming one of the must-read analysis sites on the internet. The fact that it's essentially a labor of love and a hobby makes it all the more charming. You won't agree with all of their conclusions, but the very fact that they're out there making the arguments and stirring up the intellectual soup is more than worthwhile.

No comments: