THT Mailbag

Last year was a pretty exciting year for the little project that is the Hardball Times, as we put out more articles for more readers than ever before. As we head into 2007, we’re planning some new changes—some visible, some not—as we expand our staff. In the next couple weeks, we’ll be welcoming Richard Barbieri, who will be continuing his work from The Annotated This Day in Baseball, and John Beamer of Beyond the Box Score.

Additionally, a couple weeks ago we asked you for any ideas you might have to make the site more fun for you to read. Several of you responded, and one of the ones we really liked was running more op-ed pieces. There’s just a ton of great baseball writing out in the long tail, so if you have something that you think our readership would be interested, send it along to
. Due to length considerations (just try to make it through this mailbag, I dare you), we won’t be able to publish all the submissions, even the good ones; and of course, we reserve the right to edit any submissions for length, clarity and consistency of style.

Then, every week, we’ll publish our favorite submissions in the Mailbag. So get writing!

Now, onto the questions:

How Clutch Was Jack Morris?

I was reading the Hardball Times Annual and was interested by the WPA info. What I would like to know is, how far back do we have WPA data for? Can we use it for historical comparisons? It seems to me that it could be effectively used to shed light on some longtime debates over starting pitchers, specifically how some pitch their best in tough situations. I’m thinking now of Jack Morris’ reputation for being better than his numbers because he could “bear down” and get people out in close games, while allowing runs more liberally in less-pressurized situations.

– Josh, New York

Dave Studeman: You need play-by-play data to calculate Win Probability Added. Thanks to the tremendous efforts of our friends at Retrosheet, everyone now has access to that data for every year since 1957. I don’t know if anyone has calculated WPA for all those years yet, but I do know of one site that lists WPA stats from 1972 to 2002. It’s called the Baseball Player Value Analysis System.

The period in question does include the Morris years, and it’s interesting to note that Morris doesn’t make the list of top 33 pitchers. In fact, Morris falls far, far below most Hall of Fame pitchers, slightly below such stalwarts as Bob Forsch, Jose DeLeon and Shane Reynolds, all of whom had worse ERA+’s than Morris.

According to this system, Morris had a value of about nine, which means he turned nine losses into wins compared to an average pitcher (or, he was 18 “wins” above .500). His actual won/loss record was 68 “wins” above .500. That is, most of his won/loss success was due to a great offense instead of “pitching to the situation.”

This isn’t a definitive answer to your question, but it strongly suggests that Morris’s reputation isn’t really deserved. That’s just one example of what you can find with WPA data.

Being Foul Doesn’t Pay

Has anyone ever done a study where they can deduce a pitcher’s performance by how many foul balls he gives up in a particular at-bat? For instance, if Derek Jeter works a 12-pitch at-bat against Barry Zito, say, by fouling off pitch after pitch and presumably tiring Zito, how effective will or won’t Zito be in the innings that follow?

– Mike

David Gassko: Keith Woolner comes closest to answering your question here, where he finds that the length of a plate appearance has no effect on the numbers of next batter up. I think this would apply to every other player in the lineup as well. What does it matter if a pitcher throws 15 pitches to one batter or to five? It will result in him coming out of the game in an earlier inning because he’ll hit his limit sooner, but I don’t see why the result of a given pitch (say, the 100th of the game) should be any different if it comes in the eighth inning of an efficiently pitched game or in the seventh because there was a 15 pitch at-bat earlier on.

Additionally, Dan Fox has done some interesting work with foul balls here and here. Keith Woolner has published additional research on foul balls here and here.

Getting Defensive

Are the individual player fielding stats (plus/minus) available in the Hardball Times Annual. If not, where can you get the 2006 numbers?

– Darren J.

A Hardball Times Update
Goodbye for now.

Dave Studeman: Unfortunately, individual plus/minus fielding figures (as computed by John Dewan and Baseball Info Solutions) are not available in the THT Annual or anywhere online. The only source is the Bill James Handbook, which lists the leaders in plus/minus at each position.

At this time, John is still keeping most of the results proprietary. We hope to have some advanced fielding stats of our own on the THT website next year. We’ll let people know more about our developments before the season begins.

Aces Over Fives

In regards to the article about starters, I have to wonder about match-ups in rotations. I know this was a relatively simple analysis, but a #2 starter for a team may pitch in the #4 spot and vice-versa. For instance, Chuck James gave Atlanta 18 starts at an ERA of 3.78. In the analysis, James was given the #2 designation, but in reality, he was pitching as Atlanta’s #5 starter. Therefore, his 3.78 ERA was matched up against ERA of 6.24, and James picks up 11 wins for Atlanta. I am not quite sure what to make of this information, but it might be helpful for teams to consider where the optimal spot in a rotation for their ace (and 2nd, 3rd, 4th, 5th starters).

– Andrew T.

John Walsh: It’s a good question, but it turns out that rotations between different teams don’t actually line up very well over the course of a season. Rotations get jumbled up due to rainouts, off days, etc., and not all teams jumble the same way. Here is a list of the opposing pitcher in each of Chuck James’ starts in 2006, along with the number of runs the Braves scored in each game.

+------------------+-------------+
| Opposing_Pitcher | Braves_runs |
+------------------+-------------+
| Clemens_Roger    |           4 |
| Morillo_Juan B   |           8 |
| Perez_Beltran    |           2 |
| Moyer_Jamie      |           4 |
| Miller_Wade      |           7 |
| Trachsel_Steve   |           5 |
| Hennessey_Brad M |           5 |
| Ortiz_Ramon      |           6 |
| Nolasco_Ricky    |           5 |
| Ortiz_Ramon      |          10 |
| Wolf_Randy       |           3 |
| Harang_Aaron     |           4 |
| Glavine_Tom      |           6 |
| Sanchez_Anibal A |           1 |
| Peavy_Jake       |          10 |
| Suppan_Jeff      |          14 |
| Cabrera_Daniel A |           5 |
| Hendrickson_Mark |           4 |
+------------------+-------------+

You can see with names like Roger Clemens, Jake Peavy and Tom Glavine, James’ opponent was often something better than a 5th starter. The average number of runs scored by the Braves in these starts is 5.7, which is a bit higher than the Braves’ overall average of 5.2 runs per game, but still a long ways from the #5 starter ERA of 6.24. Even more so when you consider that the 5.7 includes unearned runs, while the 6.24 does not.

James was slightly lucky to get 11 wins—a simple Pythagorean analysis suggests he should have won nine games if the Braves had scored their average number of runs in his starts.

Where the Numbers Come From

I was wondering exactly how you calculate your stats such as GB%, LD%, etc. I’m trying to figure out the raw totals of each batted ball type based on the percentages calculated, but my only check that i did it correctly is figuring out home runs allowed by using HR/F and then comparing it with actual home runs allowed. Unfortunately I’m getting the wrong total of home runs. Most of the stats are a function of the number of batted balls allowed by the pitcher, so maybe I’m not calculating that correctly. I took batters faced pitching, and then subtracted walks, strikeouts, and hit batsmen.

– Mike P.

Dave Studeman: Our batted ball stats, such as GB% and LD% are pretty straightforward. Each is calculated as a percent of total batted balls. We don’t list FB% because it’s whatever is leftover from the GB and LD percentages (except for bunts). However, our HR/F stats are adjusted for home parks, so they won’t equal a player’s actual home run totals.

For what you want to do, all of the source statistics are available in a spreadsheet for those who purchase the THT Annual 2007. We highly recommend it!

Vision of the Future

Are we back to the glory days of Kenny Williams? I felt he made a few too many good deals, and got lucky a bit too often for someone with his track record, and he was bound to make up for them. They got a couple nice players, but in a division where 90 wins isn’t good enough, don’t you have to keep Freddy Garcia and Brandon McCarthy? If they couldn’t afford Garcia next year, why did they add Jim Thome? And is there anything more valuable in today’s game than a legit starter who makes the major league minimum?

– Chip

Bryan Tsao: Actually, Williams has really shown me something this offseason. Granted, most White Sox fans probably would have wanted him to address their outfield issues first, but Williams probably increased the overall talent level of his organization while lowering its cost structure. While that might mean slightly fewer wins in the short run, it’s hard to argue with those kinds of moves in the long run.

I think he got Thome to win last season while he still had control over his starting rotation. Going into this season, however, many of his pitchers started getting older, closer to free agency and more expensive. So instead of waiting for them to completely lose their trade value, he moved them for a boatload even younger, pre-arbitration starting pitching talent.

As you note, good pre-arb players are baseball’s ultimate market inefficiency, and stockpiling players with exactly that potential (don’t forget picking up Andrew Sisco for Ross Gload) is a move designed to help them compete in the newly competitive AL Central. Better to move a player a year too early rather than a year too late, the saying goes, and in a sense, Williams is following the relatively recent example of one of his former trading partners: Billy Beane, who moved Tim Hudson and Mark Mulder for even younger talent in return.

Ranking the GMs

I know how you like challenges so I thought: Would it be possible to rank the current general managers?

Some folks over at Motown Sports have talked about it and the problems with how to compare budgets, farm systems and owner influence over day-to-day operations.

While I have no idea where to start and what is even possible to measure statwise, do you think it would be possible?

– Mike W., United Kingdom

David Gassko: Rating general managers is incredibly difficult and ultimately, something that cannot be perfectly quantified. Part of it is a lack of information; for example, we know some Yankees moves are initiated by George Steinbrenner, but we don’t know which ones. So how can we judge Brian Cashman as a general manager when we don’t even know what deals we can ascribe to him? I happen to think that Cashman is a very, very good general manager, but his track record is conspicuously mediocre—unless you believe that some of the particularly egregious Yankees mistakes were not his ideas (as I do).

But that’s a unique situation. Worse still is that ideally we’d like to judge general managers not just on what they do, but what they want to do. For example, the Red Sox were very much in the running to sign Carl Pavano a couple years ago, and gave him a pretty big contract offer. Pavano went to the Yankees, and as it turns out, the Sox dodged a pretty big bullet. But shouldn’t we debit Theo Epstein for being willing to offer Pavano so much money? What if the Yankees hadn’t signed him? The Red Sox would have made a pretty hefty mistake. Of course, we mostly don’t know what is on various general managers’ minds so we can’t really know what bad moves they have avoided.

The Pavano story brings up another issue: What is the proper way to judge the moves a general manger does make? Let’s say that Pavano projected pretty well. Is it fair to say the Yankees made a bad move in signing him if it looked like a good move at the time? Isn’t it unfair to look at things in retrospective? Shouldn’t we rather be judging moves based on all the information available at the time (which is after all, the only information the general manager could possibly take into account when making a decision)? Or do only results matter? What if there is a general manager who is really good at guessing which players will do better or worse than what the numbers tell us? Shouldn’t he get credit for that? But how can we know that he’s good and not just lucky? That’s eight questions in one paragraph, and I can’t answer any of them.

As you can see, judging general managers is an extremely difficult task. However, I do have a few basic rules of thumb to go by:

– How much does the general manager get out of his budget? Does he build a contender with little money, or does he have trouble putting together a winning team with lots of cash at-hand? Does his team win about as many games as you would expect given its payroll?

– How many egregiously bad moves has he made? A good general manager can make mistakes, or have deals unexpectedly blowup in his face, but only a bad general manager will make a move that seems like a disaster from the moment it’s announced.

– Does he properly recognize the value of his farm system? Most general managers are fine here, but some consistently neglect their minor league teams or trade great prospects for mediocre veterans. No farm system should be completely barren, and if it is, that’s the sign of a poor general manager.

– Does he understand the value of a dollar? Again, most general managers pay players at least roughly how much they’re worth, but a few will give millions to replacement-level players or something ridiculous like that. Actually, most general managers who fail this point are those who would rather pay a veteran presence millions than let a similarly-skilled and younger player play. (Note: This is distinctly different from the first point. The Marlins, for example, are an 80-win team on a $15 million payroll. If they were to sign four guys worth 2.5 wins extra each and give each $15 million a year, they would be a 90-win team on a $75 million payroll, which is still remarkably good. But, those would also be terrible signings, and I would say that their GM should be fired on the spot if he did that.)

– How well does he adapt to his circumstances? Does he know whether to be a buyer or seller at the trade deadline, or does just stand pat when moves need to be made? Is he a small-market GM who allocates a lot of his payroll towards one player, especially one who isn’t necessarily great?

– Finally, does he understand context? This is the most sabermetric of my questions. Does he understand park and age effects? Does he sign players who mesh well with his team and park (i.e. the Dodgers play in an environment particularly conducive to groundball pitchers while the Reds’ terrible outfield defense makes it tough for guys who allow a lot of fly balls)? Can he identify bargains who are underrated either because other teams have not properly adjusted their statistics or because they haven’t been given a chance?

If you ask all these questions about a general manager, I think you can get a very good feel for both his strengths and weaknesses as well as his performance overall. But it’s impossible to make a definitive list that tells us, here’s how good this guy is, and here is how good this other general manager is.


Comments are closed.