Tuesday, October 31, 2006

New York Mets 2007 Outlook

2006 results: 97-65, Won NL East, Lost NLCS

Pythagorean record: 91-71

Key free agents: Cliff Floyd, Jose Valentin, Orlando Hernandez, Chad Bradford, Guillermo Mota, Roberto Hernandez, Steve Trachsel, Darren Oliver

Plan for 2007: Plan around a decline from the team’s aging core and the overachievers from 2006

The Mets finished 2006 as the consensus best team in the National League, finishing with the best record in the league by nine games and sporting an everyday lineup with four stars and no huge weaknesses. Their starting pitching was suspect for much of the year, but an excellent bullpen carried them to 97 wins and to the NLCS, where they fell in seven games to the eventual World Champion Cardinals.

How much of that success will carry over to 2007? The good news for the Mets is that they have a great core to build around. Their lineup contains two superstars, Carlos Beltran and David Wright, and two great supporting players in Jose Reyes and Carlos Delgado. The pitching staff returns two future Hall of Fame starters in Tom Glavine and Pedro Martinez, and a potential Hall of Fame closer in Billy Wagner. Though all three, along with Delgado, are well past their prime, they still contribute at a high level when healthy—which, for Glavine, is always.

Even the stars, however, are not without question marks. Are the real Beltran and Reyes the 2006 versions, or 2005? Will Pedro recover from his surgery to become Pedro! in time for the 2007 playoffs? Will Delgado suffer from the sharp mid-30s decline that has killed off many of the best first basemen of the 1990-2005 era? Will Glavine’s stuff hold up in his age 41 season?

The stars will also need some help. Last year’s team had a Pythagorean record of just 91-71, and just 88.7 “third-order wins” (Baseball Prospectus’ stat of choice to determine how well a team has really played after accounting for the game’s natural variance). 91 wins is good, but it’s not a position of extreme strength, especially when the Mets are likely to receive less production in 2007 from several positions, especially the bullpen.

Let’s break down the decline:

Catcher: Paul LoDuca will be 35 in 2007, and catchers typically go through a period of extreme decline in their mid-30s. Projection: -1 win from 2006 value.
First Base: Carlos Delgado’s numbers, age and health point to a consistent decline, and he will be 35 in 2007 as well. Projection: -1 win.
Second Base: Jose Valentin is coming off a career year at age 36. Whether they re-sign him or not, they will experience a decline in production, as no available second baseman can be expected to match his 2006 numbers. Projection if Valentin re-signs: -2 wins.
Shortstop: Jose Reyes will only be 24 next year, but he also jumped substantially in value from 2005 to 2006, so it’s hard to project additional growth. Projection: no change.
Third Base: David Wright has maintained star-level numbers for two years and will be only 24 next year. Aggressive projection: +1 win.
Left Field: Though Cliff Floyd was injured and ineffective, Endy Chavez filled in well in his absence. Lastings Milledge is an exciting prospect, but he lacks on-base skills at this stage in his career. It’s hard to see him matching Chavez’s .306/.348/.431 line from 2006. Projection assuming Milledge is handed the job: -1 win.
Center Field: Carlos Beltran is a stud, but he has never had a season nearly as valuable as his 2006, and he’s one year removed from being labeled a bust. Projection: -2 wins.
Right Field: A full season of Shawn Green might be cause for optimism, but Green has declined significantly, and he isn’t really a better player than Xavier Nady at this stage. Projection: no change.
Bench: If Endy Chavez or Lastings Milledge is available on the bench next year, it should be a significant boost to the Mets’ flexibility. Projection: +1 win.

Add it up, and that’s a net loss of five wins from a team that only had 91-win talent to begin with, and we still haven’t gotten to a pitching staff featuring two 41-year olds, a 35-year old staff ace who will miss half of 2007 recovering from surgery, a non-prospect rookie who posted the best ERA of the rotation, and a bullpen full of guys who had career years in 2006. You may have noticed the high amount of age-related decline, a phenomenon common to teams that build mainly through long-term free-agent contracts.

Put simply, to remain at the 91-win level, the Mets have a lot of work to do. The most important step for the Mets is to realize that this work is necessary. They can’t look at their nine-win advantage over the rest of the NL and assume they can coast to a playoff spot next year. The current Mets incarnation, with all the injuries and decline, probably projects to finish no better than .500 in 2007. To build a real contender, the Mets need to add 10 or more wins. Where can they get them?

Though the bullpen is likely to decline, it makes no sense for the Mets to spend on a free agent when they have live arms available. The rotation will be in need of some help. Glavine, Martinez and John Maine are under contract for next year, but each will likely see his performance decline in 2007, due to age, injury, and regression, respectively. The Mets have the money to break the bank on a free agent ace like Daisuke Matsuzaka or Jason Schmidt, and it would be an ideal fit, one that alone could get them halfway to their 10-win goal. If Trachsel and Hernandez leave, the Mets also need one or two mid-rotation starters. Mike Pelfrey or Aaron Heilman may be the answers, and Oliver Perez could be useful, although the Mets shouldn’t count on him for anything. They may be better served shopping in the bargain bin for Mark Mulder, Ted Lilly, or even Kerry Wood. Each of those players has the potential to be a shutdown pitcher in the playoffs, and each will likely come with a deflated price tag.

Among the position players, the Mets must acquire a good second baseman from the many free-agent options, perhaps Ray Durham or Adam Kennedy, or even re-upping Valentin. They also should shop for a corner outfield bat, despite the public’s false impression of Green and Milledge as above-average players. Carlos Lee or Alfonso Soriano would be good options, as the Mets lineup already leans heavily left-handed. Jose Guillen may be a good fit if the Mets spend their big money on starting pitching. Unfortunately, the market on corner outfielders is thin this winter, so the Mets are in trouble if they can’t land a big fish.

The Mets could also consider trading one of Delgado and LoDuca, two aging players whose perceived values far outweigh their actual contributions, in a deal that either returns a more affordable replacement at the same position or clears a spot for a free agent to offer similar or better output. If a team is desperate for bullpen help, might they be willing to give up a useful bat for Pedro Feliciano or Heilman?

Whatever direction the Mets choose, they must take an aggressive stance and be aware of the likely decline from next year’s team, or their reign atop the NL East will be a short one.

Detroit Tigers 2007 Outlook

2006 results: 95-67, Won AL Wild Card, Lost World Series

Pythagorean record: 95-67

Key free agents: Sean Casey

Plan for 2007: Upgrade several positions to avoid regressing to under 90 wins next year

Many recent World Series teams have declined significantly in the year following their trip to the big one. Much of this is due to simple regression to the mean; baseball’s so-called Plexiglass Principle states that all teams tend to gravitate toward a .500 record in future years. However, a considerable part of the problem is the complacence demonstrated by successful teams. Clubs like the 2002 Angels and 2003 Marlins kept their World Series lineups virtually intact the next year, which resulted in 77-85 and 83-79 records, respectively.

Other teams, like the 2005 White Sox or the Yankee teams of recent vintage, have followed up their Series appearances with bold moves to keep the team in contention the next year. Though the Sox failed to make the playoffs in 2006, they won 90 games and were in contention until the last week. Meanwhile, while many criticize George Steinbrenner for overreacting to fill every hole on the New York roster, the Yankees have continued to win the division every year.

The 2006 Tigers face a crossroads. Their Pythagorean record of 95-67 indicates that they really had 95-win talent. They can return virtually the entire 2006 roster—everyone but Sean Casey is under contract for 2007—or they can use the additional revenue from increased ticket sales to target some upgrades in trade or free agency. The latter option is clearly better, but will require good roster management, as the organization will essentially need to remove a player currently under contract for each one they add.

What remains to be seen is whether the Tigers’ 2006 playoff run will cause free agents to warm up to the idea of playing in the Motor City. In recent years, the franchise has pursued free agents like Ivan Rodriguez, Magglio Ordonez and Carl Pavano by offering them far more money and years than any other team, perhaps to compensate for the “stigma” of playing for a team that lost 119 games in 2003.

Where can they upgrade the team? If Mike Maroth returns healthy, Detroit will have five solid starting pitchers, so it would be a waste of resources to sign another, unless they can offer one in trade for a big bat. Joel Zumaya looks poised to be a better closer in 2007 than any free agent reliever. If the Tigers need any pitching, it’s in the back end of the bullpen, which won’t be a big difference-maker over the course of a season.

What about their lineup? Six positions are set for 2007. Rodriguez, Carlos Guillen, Placido Polanco, Brandon Inge and Curtis Granderson are above-average performers at their positions, and all but Rodriguez are signed to cheap, below-market contracts. Ordonez is staying put because of his image and $75 million contract rather than his production, which is around league average for a right fielder, when he is healthy.

That leaves first base, left field, and DH as the easiest positions to improve. These spots are currently occupied by Chris Shelton, Craig Monroe, and Marcus Thames. You can certainly win a championship with these three as your worst everyday players, but they occupy the positions where it’s easiest to find an impact bat, and Shelton is the only one likely to improve next year. Monroe, who can’t take a walk to save his life, is a particularly terrible fit in the number 2 spot. In fact, the Tigers only have one hitter, Guillen, who delivers a really good OBP.

What the Tigers really need is a guy who can get on base in front of their power hitters, preferably a left-handed batter to balance their righty-heavy lineup. That player is Barry Bonds. For all his shortcomings last year, Bonds led the majors with a .454 on-base percentage and should be over .400 again in 2007. Jim Leyland knows how to manage Bonds from their years together in Pittsburgh, and Bonds can DH regularly with an occasional start in left field or at first base. Bonds won’t be cheap, but he is an ideal fit, particularly if they bat him in the leadoff or second spot in the order to get him more plate appearances and more times on base in front of the heart of the order.

Another very good option is a rumored trade with Texas involving Jeremy Bonderman and Mark Teixeira. Though Bonderman has great stats and stuff and will be only 24 next year, a good young starting pitcher simply doesn’t match the value of a top hitter. If this offer is actually on the table, Detroit should quickly accept, even if it means throwing in something of reasonable value, perhaps Shelton. Teixeira would be under the Tigers’ control during his peak years at ages 27 and 28, and brings a great bat and another above-average infield glove for a pitching staff that relies heavily on the defense behind them. He also doesn’t bring the injury or age risk that the big free agents carry with them.

The Tigers’ next choice should be Frank Thomas. Thomas doesn’t quite bring the presence or on-base skills of Bonds, and he is right-handed, but he still fills the need for a big-time power hitter in the middle of the order, when he is healthy. Nomar Garciaparra makes another attractive option. Though he has often been hurt recently, he has never stopped hitting when healthy. Detroit can plug him in at DH to eliminate his defensive shortcomings, and if he gets hurt, they can plug in Thames. Carlos Lee and Alfonso Soriano are better health risks, but they lack the on-base percentages of Thomas or Bonds and will come with expensive, long-term price tags.

The 2007 Tigers, like the 2006 version, are likely to feature a balanced team without any major holes. But without adding an elite-level hitter to the team, they will likely end up with an 85-90 win team, one that needs to overachieve to emerge from the strong AL Central. Their willingness to make sacrifices to get that talent will make all the difference for their chances at returning to the playoffs in 2007.

St. Louis Cardinals 2007 Outlook

2006 results: 83-78, Won NL Central, Won World Series

Pythagorean record: 82-79

Key free agents: Jim Edmonds (team option), Jeff Suppan, Mark Mulder, Jeff Weaver, Jason Marquis, Preston Wilson, Ronnie Belliard

Plan for 2007: Overhaul the roster to prevent it from slipping beneath the level of contention.

The St. Louis Cardinals are World Series champions. Flags fly forever, and the trophy will be theirs for at least a year. Furthermore, in Albert Pujols, Scott Rolen, and Chris Carpenter, they have one of baseball’s strongest cores of superstars to build around.

That’s the good news. The bad news is that the Cardinals are in big trouble next year. Baseball Prospectus publishes a statistic called “third-order wins” which attempts to strip out the variance in baseball and simply tell you how well a team has played in a given season. Usually, a team’s third-order win total will closely approximate its actual win total, although factors like clutch hitting, a great record in one-run games, or an easy schedule can help a team win more games and “outperform” its third-order win total.

The Cardinals won only 83 regular-season games in 2006, and according to third-order wins, they weren’t even that good. They managed just 75.8 third-order wins, the ninth-worst total in baseball. By one measure at least, the Cardinals begin this offseason with one of the ten worst teams in the major leagues, and several key contributors are likely to leave in free agency, leaving them in an even weaker position.

What can the Cards do? They can’t enter a rebuilding period, as their farm system is weak and they need to build a contender around the core of Pujols, Rolen, and Carpenter while they are still playing at a high level. The only real option is a major splurge in the free agent and trade markets this offseason.

Fortunately for the Cards, they have two big things going for them: a creative general manager in Walt Jocketty and a devoted fan base that will generate revenue and perhaps encourage free agents to choose St. Louis. But where can they look for the additional production they’ll need to contend in 2007?

In addition to the many holes created by the departing free agents, the Cardinals feature two everyday players whose bats are subpar, postseason heroes David Eckstein and Yadier Molina. At age 24 and with excellent defense, Molina is worth bringing back as a starter. Eckstein is 32 and coming off three poor seasons out of four, making him a prime candidate for an upgrade. Unfortunately, the only better shortstops in the free agent market are Julio Lugo and Craig Counsell. If the Cardinals can sign one of them and trade Eckstein for a useful part, it’s a move worth making.

One move St. Louis should definitely make is to pick up Jim Edmonds’ option. At $7 million—the difference between the $10 million option and $3 million buyout—this move looked like a no-brainer before Edmonds suffered through injuries and slumps during most of 2006. But in a poor free agent market, $7 million for one year is a bargain for a player who should be worth five wins if he can play 130 games next year.

If St. Louis is stuck with Eckstein and Molina in the lineup, they can’t afford to start the no-hit Aaron Miles at second. Ray Durham, Adam Kennedy, Jose Valentin, and Todd Walker are all getting old, but each would be a significant upgrade. Kennedy would be the best fit on a team with a soft-tossing groundball staff, while Durham is the best hitter of the group. It remains to be seen which will be the better fit for the Cards in 2007.

In the corner outfield spots, the Cardinals have two very similar players in John Rodriguez and Chris Duncan, left-handed hitters who can mash against righties but can’t hit a lick against lefties. They also still owe $10 million to Juan Encarnacion. If there’s a trade market for any of these three, and particularly if they can receive a starting pitcher in return, Jocketty should pull the trigger.

Fixing the pitching staff will be a challenge, with lots of holes to fill and the league-wide demand for starters at an all-time high. Signing Mark Mulder to an affordable, incentive-laden deal may be a good move, although Mulder was in steep decline long before his injury. Suppan is likely to sign a contract in the neighborhood of four years and $40 million, way too much for a pitcher who struggles to strike out five men per nine innings. Jeff Weaver will probably come cheaper, and should be roughly as good going forward. If the Cardinals sign Weaver to a one-year, $5 million deal to re-establish himself for next year’s free agent market, it would be a good move for both parties. Marquis is not likely to come back, nor should he be. His peripheral stats in 2005 foreshadowed his awful 2006, and he is unlikely to be a league-average starter again.

With Jason Isringhausen returning from surgery, the Cardinals should give Adam Wainwright a shot to start in the middle of the rotation. He can always move back into relief if the experiment fails. This will probably leave two open starting pitching slots to fill via trades or free agency.

The best free agent starter is Daisuke Matsuzaka. Matsuzaka will be the most expensive starter after including his posting fee, but his NPB stats and stuff are both very good, and many Japanese pitchers with unorthodox deliveries have made a big impact in their first season, like Hideo Nomo, Akinori Otsuka, and Shingo Takatsu. Other top-level pitchers include Jason Schmidt, whose strikeout rate has held up even though his other numbers have not, and Mike Mussina, whose option is expected to be declined by the Yankees. The Cards can target one of these elite starters and one lower-tier, perhaps Adam Eaton.

All in all, the Cardinals have a lot of work to do to even get back to the playoffs, let alone repeat as World Champions. They’ll need to be aggressive and willing to overpay to keep a contender on the field, or risk spending September out of the playoff hunt for the first time in seven years.

Wednesday, October 25, 2006

Zumaya's Real Folly

I said no analysis, and I lied. Today's issue-du-jour is Joel Zumaya's decision to throw to third base on a grounder back to the mound with runners on first and second and nobody out.

Because the play resulted in a throwing error that scored two runs, everyone was all over him for throwing to third base instead of second base. Of course, had the Tigers turned a 1-5-3 double play followed by a sharp single or wild pitch that only advanced the runner to third, everyone would have discussed what a crucial play it was to get the lead runner.

Let's take a non-results-oriented viewpoint. The question is, if Zumaya can pause and consider, what is his best option?

We'll analyze this using run expectancy. I normally use win expectancy, but a team that is down 2 runs in the seventh will rarely win regardless of the outcome. Using data from 2004-06, the number of runs the Cardinals expect to score in the inning is:

.350 if the 1-5-3 double play is completed, leaving a man on second and two outs

.369 if he completes a 1-4-3 or 1-6-3 double play, leaving a man on third and two outs

.943 if he gets only the lead runner at third, leaving runners on first and second and one out

Remember, we're not even considering the possibility of an error here. All the talk that the error occurred because Inge was not ready to receive the throw is totally overblown; he is one of the best defensive third basemen in the majors and has great reflexes and agility.

It is still harder for him to turn the double play, however. And what do we find? Assuming the Tigers turn a double play, they gain .019 runs by nailing the lead runner. However, if they don't turn the double play, entirely likely when throwing to a third baseman who is not used to making this play, it costs them .574 runs, over 30 times as much.

In other words, assuming they would have turned the double play by throwing to second (and it should have been easy with the gimpy Pujols running), they need to turn this double play 97% of the time for throwing to third to be the better play. A player like Inge is going to turn this the majority of the time, but 97% is more than one can expect out of him.

Should Zumaya be able to do this math in his head in a split second? Of course not. But it costs a team virtually nothing to hire someone to do this calculation in the offseason and teach the players that the notion of getting the lead runner is totally overrated.

Baseball managers, like many of us, have a bias towards expecting the best possible outcome from a situation. With runners on second and third and none or one out, they will intentionally walk the batter, not because it is the best play or even close, but because they are obsessed with allowing no runs, and remember past instances of getting out of a big jam via a crucial double-play ball. Tony La Russa remembered Yadier Molina hitting a big home run and (apparently) believed he would do so again in Game 2. Joel Zumaya may have remembered turning a triple play on a slowfooted batter at some point in his life. Perhaps he just wanted to keep the lead runner off third. It wasn't worth it.
No analysis today, although I am somewhat proud of Jim Leyland for using Joel Zumaya in a game that was relatively close even though the Tigers were behind.

I do want to discuss a moment from Game 3 that i thoroughly enjoyed. FOX Shill Chris Myers is in the left field stands, going on and on about how many free tacos Taco Bell is poised to give away if a home run is hit to left field, and how generous it was of Taco Bell to give him lots of money so that baseball fans could spend the game hearing about tacos instead of watching baseball. Meanwhile, a bunch of chumps around him are waving big placards reading "Free Taco Here".

Myers then turns to the 11-year-old he's interviewing and asks if this promotion is why he's in the left field stands. His response:

"Yeah, I'm out here to watch the game, but a taco wouldn't be bad." He instantly turns his head back and watches the next pitch, while Myers drones on about free tacos for everyone.

Let me tell you, when I was 11, I would have killed for a Taco Bell taco, almost as much as I would have killed for tickets to nonexistent Cubs playoff games. Bravo, Bobby Kelly. It's people like you who kept Spider-Man webs off our bases. Now if you could only reach out to the jackasses in your section who are holding up the Taco Bell signs.

Monday, October 23, 2006

Predictive and Non-Predictive Stats

Several years ago, Voros McCracken stirred up a hornet's nest with this article on Baseball Prospectus' web site, which drew a line between the numbers a pitcher does and does not have control over. While most of these were clear-cut (strikeouts, W-L record), McCracken asserted that earned runs and hits allowed were largely in the hands of the defense behind the pitcher.

With some reservations, everyone in the baseball analysis community has accepted McCracken's conclusion that hits allowed and ERA are a poor predictor of future success as a pitcher when compared to defense-independent pitching statistics (DIPS).

Of course, this doesn't mean that major-league baseball teams have listened. Russ Ortiz, who for years had put up terrible DIPS numbers despite above-average ERAs, was given a $34 million contract by the Diamondbacks before the 2005 season. Ortiz was released, with full payment of his remaining contract, less than halfway through the deal by a team that valued his empty roster spot more than a pitcher they had recently valued at $8.5 million per annum. He continued his failures with the Orioles, who signed him for the major-league minimum.

As a whole, baseball teams do a poor job of recognizing which statistics will yield future success and which will not. Fifty years from now, baseball historians will look back and laugh at the idea that teams and fans of this era used save totals as their primary evaluator of relievers, laugh at how the Mariners and Yankees were fooled by the career years of Adrian Beltre and Carl Pavano, laugh at how teams were still overpaying for pitchers whose lack of strikeouts made it clear they wouldn't repeat their success.

(Warning: NSFW Earl Weaver link)

Thirty years ago, Earl Weaver understood the value of players who would put up numbers that will always win you ballgames: walks and homers. Whatever team signs Juan Pierre for way too much money this offseason is going to learn an expensive lesson about what his skill set is really worth. Meanwhile, Russ Branyan can't stick with a team despite being an above-league average player for his career.

Smart baseball teams profit from the knowledge that plate discipline, making contact, and power are the paramount skills for hitters, and commanding the strike zone and keeping the ball down are the keys for pitchers, and that almost everything else is just fluctuation. Most of the current terrible contracts in the league would never have been given out if teams kept this basic concept in mind.
I'm getting a decent amount of flak from people who defend every decision Jim Leyland and Tony La Russa make, because they managed less-than-elite teams into the World Series.

I want to make a few things clear:

- Jim Leyland and Tony La Russa are good managers.
- I don't believe I could manage a team better than either of these men.
- Nevertheless, they make poor choices with regard to many obvious decisions.

I play a lot of poker. In the poker world, when a winning player makes obvious errors that compound in the long run, they are called leaks. Though this player will still win plenty in the long run, perhaps even enough to earn a living, he will still fail to reach his potential win rate unless he stops making these basic errors.

Though there are a countless number of decisions a winning poker player must get right in the long run, two things are of paramount importance:

- Get the big decisions right. Big decisions usually mean close decisions in big pots. If you are able to win one big pot each session that a mediocre player wouldn't, you are way ahead of the game.

- Get the frequent decisions right. These are decisions that aren't a big deal in the short run, but come up quite often. For example, if you decide to just once play a terrible hand, like 8-4 offsuit in hold 'em, it doesn't cost you that much. If it costs you $10 to call, you may lose $4 or $5 per call in the long run. But if you start playing every hand, you will get killed in the long run as these losses add up.

In baseball, there are few truly big decisions that come up in big situations, and in most of these it is obvious what the best play is. Because sports "analysis" is based primarily on results, the praise usually goes to the manager who made the decision that turned out best, even if it made no sense at the time. If Yadier Molina had delivered a game-tying or game-winning hit in the ninth inning yesterday, La Russa would still be fighting off reporters asking him how he knew Molina was going to deliver. Jim Leyland could have made a (correct) big decision by bringing in Joel Zumaya after Todd Jones allowed the tying run to reach second base, but instead he looks brilliant for sticking with Jones, who got Molina to ground out to end the game.

We know why Leyland stuck with Jones: Jones is a "closer", and that label cannot be removed when it's still a save situation. I don't understand why Molina was allowed to bat, but presumably La Russa considered his options and decided the weak-hitting catcher was the best hitter he had available at the time, perhaps because of Molina's NLCS-winning homer last week. Either way, two managers each made poor decisions in a big spot, and the fact that one of them won the game should not affect the way they are viewed.

More common are the small errors that come up all the time. Writing the wrong player into a lineup card is a small error, but one that affects your team every game until you stop doing it. Every time Jim Leyland starts Ramon Santiago when Chris Shelton could be playing, he is costing himself perhaps .04 wins. This isn't a big deal if he does it only once per year. When he does it every day, that's over six wins a season lost, or in the case of the World Series, potentially .28 wins lost. In a series where a single game is often the difference between a championship ring and a long off-season of regret, it is absolutely unacceptable to make this kind of error.

Continuing to use Jones over Zumaya as the team's closer this season is a similar mistake. This one ended up costing the Tigers only a couple of wins this year, as Jones converted many more saves than he should have based on his other numbers, because he gave up most of his runs in non-save situations and constantly put the tying and winning runs on base before getting out of the inning, collecting the same "save" despite doing a much poorer job. Managers treat saves as a predictive stat, when they are simply a record of what has happened in the past.

If the managers in the major leagues cannot get these decisions right, it's time for them to hire someone who can. The NFL long ago hired specialized coaches for every position, and baseball teams have hitting coaches, base coaches, pitching coaches, bench coaches, and bullpen coaches. Baseball Prospectus estimates the value of an additional win at over $2 million. If teams can find the room in their budgets to give $40 million to a below-average, injury-prone starter, they can afford $20,000 for someone to make decisions like this correctly. It would take only .01 additional wins per year to earn this money back.
Cardinals Don't Want to Win

The Cardinals did not want to win tonight.

That's about all that can be said about the decisions to let Preston Wilson and Yadier Molina bat tonight with the tying run at second base and two outs in the ninth.

In John Rodriguez, Chris Duncan and Gary Bennett, Tony La Russa's bench had two lefties who have peformed quite well against right-handed pitching and a backup catcher to replace the pinch hitter for Molina. The Tigers had a soft-tossing righty on the mound and no one warming in the bullpen. This is the easiest pinch-hitting decision ever.

And La Russa blew it.

Yadier Molina was the worst-hitting regular catcher in the majors this year, which basically makes him the worst-hitting regular overall.

It's not unreasonable to suggest that letting Molina bat over Duncan or Rodriguez halved the Cardinals' chances of winning this game, taking them perhaps from 16% to 8%. (The historical percentage is 13%, but Jones is a below-average closer and Duncan and Rodriguez represent above-average hitters against a righty.) Anything that costs you 8% of a win is a huge boneheaded move, perhaps equivalent to starting your number 4 over your ace in a game 7 when both are on full rest.

Letting Wilson bat was equally dumb. Wilson was released outright this year from a contending team due to ineffectiveness, and it's not like the Astros had a great outfield. Replacing him with Rodriguez represents no real defensive downgrade, Rodriguez is clearly the better hitter against a righty, and bench depth was not an issue at this stage in the game.

Why these players were put on the roster to not be used is beyond me, though they'll likely see more playing time in the series than Neifi Perez, who was put in over Chris Shelton AGAIN, perhaps because Leyland needs Neifi's bat to pinch hit in the games at Busch.

Thursday, October 19, 2006

Tony La Russa cannot manage a bullpen

The situation: bottom of the sixth, tie game, runners on second and third, one out. The Mets are about to send up a bunch of lefties who can't hit lefties. Your starting pitcher doesn't look tired, but he is mediocre to begin with. Your move.

As you might have guessed, La Russa opted to improve the Mets' position by intentionally walking the batter, Shawn Green. Surprisingly, he then declined to go to a left-hander in the bullpen, a clear mistake. The Mets have very little bench flexibility, and Willie Randolph has shown a willingness to stick with impotent right-handed batters instead of pinch hitting against the Cardinal southpaws.

As with every decision I tend to criticize, it worked out for the Cardinals, as Suppan retired the next two batters without a run scoring.

In the top of the seventh, Suppan came to bat, to the surprise of many at home and in the booth. The problem with this move isn't that it was terribly costly to let him bat--doing so only costs the Cardinals about .05 runs on average--but that Suppan, after throwing 90 pitches, is hardly a better pitcher than what the bullpen had to offer. In an elimination game with the next day off, the Cards could afford to throw everyone out there.

They stuck with Suppan, he gave them (at least) one more good inning, and now the bullpen is better set up for a potential long extra-innings affair. But that's an unlikely outcome.

Wednesday, October 18, 2006

How were the Tigers really built?

You're going to hear a lot of stories, either now or after the World Series, about how the Tigers came together and rebounded from the 119-loss season of 2003, and how the 2006 version was built through a series of prudent acquisitions.

I leafed through a copy of Baseball Prospectus 2004 for their comments on how the 2003 team would need to rebuild. They looked at previous futile teams and showed that they tend to turn over their entire roster before becoming contenders again. The Tigers do have a few players remaining from the 2003 team, but not many; they underwent a successful rebuilding program by importing players from outside the organization.

Since we're not results-oriented here, let's look at how the team really came together, as the decisions looked at the time. First, a look at the hitters.

C: Ivan Rodriguez (Free Agent, 2004)

Rodriguez was given an outrageous 4-year, $40 million contract after the 2003 season in an attempt to restore credibility to the franchise after they went 43-119 the season before. His contract had a theme common to the free agents on this list: he only signed with the Tigers because they gave him a much more lucrative deal than he could have gotten elsewhere.

Rodriguez's contract runs through ages 33-36, a period where many catchers deteriorate both physically and on the field. At the time it was signed, he had missed more than 50 games in three of the past four years. To their credit, Detroit did hedge their bets by allowing themselves to void his contract if he missed extensive time with back injuries. Though it has worked out well, this contract was correctly denounced at the time.

1B: Sean Casey (Trade, 2006)

Casey was acquired in trade with Pittsburgh at the deadline because the Tigers "needed" a left-handed bat in the lineup (they did) and a first baseman (they didn't). Since then, Casey has failed to outhit Chris Shelton, yet still gets the starts, even against lefties. Shelton was inexplicably left off the postseason roster even though he is better than Casey by almost any measure. If the Tigers re-sign Casey, they will regret it, both for the wasted money and blocking their best first baseman from the lineup.

2B: Placido Polanco (Trade, 2005)

Polanco was acquired in a trade with the Phillies, reportedly because Ugueth Urbina got into a fight on the team plane and needed to be dealt quickly. The Tigers did well to pick up Polanco, who was unable to get an everyday job after he surprised the Phils by accepting arbitration, knowing he would come back in a part-time role behind Chase Utley and David Bell.

Polanco finished second in the majors in batting average last year. Though his stats have dipped heavily this year, he still has plus defense and a .300 average, a package that made him one of the league's best second basemen until 2006. Looking back, it's a shame he never found full-time work until now, as he was a legitimate 7-8 win player at his peak.

Though his hand was forced, Dombrowski made a great pickup of an undervalued commodity, and signed him to a very reasonable contract (4 years, $18.4 million).

3B: Brandon Inge (Homegrown)

Inge has had one of the strangest career paths in history. Entering 2004, he was 27 years old, had established himself as the worst-hitting regular catcher in the league, and the Tigers had just made a long-term commitment to Ivan Rodriguez.

Any normal team would have brought back Inge as a backup catcher if at all. Instead, the Tigers converted him to a full-time third baseman. Then a funny thing happened: Inge became a league-average hitter at third base, where the standards for batters are much higher, and became one of MLB's premier defenders at the hot corner.

It's possible the Tigers scouts believed that Inge's bat and glove could flourish after shedding the tools of ignorance, but I doubt it. No reasonable analyst could have foreseen Inge becoming this valuable as a third baseman.

SS: Carlos Guillen (Trade, 2004)

There must have been something in the water in Detroit in the winter of 2003-04, because like Inge, Guillen began flourishing at that point. Guillen arrived in Detroit a .270/.330/.400 type with below-average defense, more a placeholder than a player who drives a team into contention. Since then, he has developed decent power for a shortstop, batted at least .318 every year, and posted two star-level seasons, while remaining on pace for one in half a season in 2005. He has become the first player in major league history to increase his batting average seven years in a row.

I don't really know how to explain this one, either. He has been able to sustain the growth, so it doesn't look like a fluke, but why the sudden turnaround? Did the Detroit coaches know something no one else did? Something in the clubhouse coffee? It's hard to pin down.

LF: Craig Monroe (Claimed off waivers, 2002)

Monroe is exactly the kind of player who gets overrated by the masses because he has decent HR totals, has collected several clutch hits this year, and can't get on base to save his life. He has batted second for much of the postseason, even though he is very poorly suited for the role.

Despite good power, Monroe remains a below-average corner outfielder due to his lack of walks, poor defense and declining batting average. He resembles Jacque Jones or Juan Encarnacion, a corner guy with no on-base skills and moderate power. Like Jones or Encarnacion, Monroe will be overpaid when he hits the free agent market. He makes a great candidate for a trade or non-tender after this year.

CF: Curtis Granderson (Homegrown)

Granderson is an example of the player development that teams need to do to win, and that the Tigers have failed at for years. Even when the Yankees "bought" their way to four World Series wins in the late 90's, they had a core of homegrown players like Derek Jeter, Mariano Rivera and Jorge Posada who were cheap due to lack of service time. By getting a very productive player for the cost of the major-league minimum, you save the team $10 million that can be used to patch a variety of holes (though not enough to pay the Tigers' two closers this year.)

Granderson was knocked for playing at a non-baseball school (He is the only University of Illinois-Chicago grad to ever play in the majors), but the Tigers scouts and minor league coaches deserve credit for discovering and developing him.

RF: Magglio Ordonez (Free Agent, 2005)

Ordonez was given a contract even worse than Pudge's: 5 years, $75 million after he underwent a mysterious European surgery on his knee. It seems likely that Maggs would not have gotten more than perhaps 4 years, $50 million with any non-Tigers team, but they felt the need to add a big-name free agent to turn around the franchise from years of losing.

Since arriving, Ordonez has continued a decline in power that makes him a longshot to reach 30 homers again. He's settled in as a .300/.360/.480 corner outfielder in a neutral park with poor defense. Plenty of players with that skill set can't even get everyday jobs.

It's hard to criticize any move that helps a team to a World Series run that has revitalized the franchise, but this is a contract that already looks bad and will only appear worse two years from now.

DH: Marcus Thames (Free Agent, 2004)

Thames has spent the past several years as a 4A player: one who destroys AAA pitching but couldn't stick in the majors. Though many of these guys fail to amount to anything in the majors, many do (Brady Clark and Ken Phelps come to mind), and the Tigers deserve credit for giving him another chance, although they still refuse to give him the full-time job.

Sunday, October 15, 2006

Green Costing Mets

The Mets' addition of Shawn Green in August was lauded by few and criticized by many. The analyst community was particularly pessimistic, realizing Green was done as a productive player and was now a significant liability in the field.

Green has now made three costly misplays in the NLCS. In Game 2, he turned a potential game-turning catch into a two-run Scott Spiezio triple. In Game 3, he turned what would have been a routine running catch for a right fielder with normal range into another two-run triple for Spiezio. Tonight, he took a very poor route to a ball in the gap, giving Juan Encarnacion an extra base and ensuring Spiezio would score from first. (Think Spiezio is going to remember to send Green a Hanukkah card this year?)

Green's bat is not worth putting up with this abysmal defense. If and when Cliff Floyd comes back, Endy Chavez should be starting in right field.
Fire Tom Coughlin

I'm writing an NFL post for a change.

Today, the Giants played the Falcons. With 3:45 left and the Giants leading 20-14, they scored a touchdown to pad the lead to 12. The Giants then sent in their kicking unit and converted the extra point to make the score 27-14.

This is just an indefensible move, even though the Giants went on to win the game by the same score. With under four minutes to play, there is no way the Falcons are going to engineer three different scoring drives to get a touchdown and two field goals. Barring an unlikely missed extra point, there is no real difference between a 12-point and 13-point lead in that spot.

However, there is a big difference between a 13-point and 14-point lead. In the unlikely event the Falcons score two touchdowns, a 14-point lead gives the Giants another chance to win the game in overtime, since the Falcons are unwilling to think outside the box and go for two when it is correct to do so.

Even if the chance of a successful two-point conversion was as low as 10%, it was the right move to go for it. Whoever is in charge of these decisions in the Giants coaching staff needs to be relieved of those duties immediately.
Unlikely Advances

When the playoffs began, two teams were left for dead by the oddsmakers and the public. They were both coming off late-season collapses, one resulting in blowing a division lead of 10 games with 50 to play and having to settle for the wild card, the other narrowly avoiding blowing a 7.5 game lead with 11 to play.

One of these teams, the Cardinals, was really a weak team; they won just 83 games and should have won even fewer. Though they weren't technically the fourth seed, they were considered the runt of the litter, a team that would continue its late-season collapse and make a first-round exit.

The Tigers, on the other hand, were a strong contender disguised as a pushover, cast by the media as a flash in the pan that had come back to earth and would be quickly disposed of by the most intimidating lineup in MLB history. Admittedly, a package of average hitting, good pitching and great defense isn't as sexy as a lineup of nine former all-stars.

A funny thing happened, though: The Tigers have won seven straight games to clinch the AL title, including three against the juggernaut Yankees, and the Cards are up 2-1 and favored to meet them in the World Series. What does this tell us?

1. Momentum is meaningless
2. Anything can happen in baseball
3. The Tigers are pretty good at baseball

Number 1 is something that should have become clear with last year's White Sox, who narrowly avoided one of the biggest late-season collapses of all time, but rebounded to go 11-1 in the playoffs. Of course, by the time they were in the World Series, everyone was talking about how they now had all the momentum, what with their playoff run. Apparently it shifted without anyone noticing.

This brings up an interesting point. At the beginning of tonight's Fox broadcast of the Cardinals-Mets game, they pointed out that the momentum in this series could shift very easily. If momentum can shift that easily, does it really even exist? Apparently baseball talking heads are not familiar with the definition of momentum. Momentum in real life not only actually exists, but it is not easily reversed. Swing a bat at a car barreling down on you, and you may dent the car as it runs you over. Swing a bat at a terrible 0-2 delivery from Brad Lidge, and you completely reverse the momentum of the series.

Number 2 is often stated as "anything can happen in a short series". I think even this underestimates the variance involved in baseball. Entering the season, the AL Central featured four very similarly talented teams (sorry, Royals fans). The Tigers shocked the world--who, except for stat geeks, had all vastly underrated them at the beginning of the year--by emerging from the pack to make the playoffs. If you played the 2006 season over, it wouldn't surprise me at all to see the Indians win the division and the Tigers finish fourth. That's what happens in baseball. The world champs improve their roster and finish in third place. You want a dynasty, go watch the NBA.

Of course, short playoff series just amplify this effect. The ultimate example is the World Baseball Classic, which "proved" that neither the USA, nor Venezuela, nor the Dominican Republic, are among the two best countries in international baseball, and only the DR is in the top four.

As for number 3, I can understand, up to a point, why people gave up all hope on the Tigers. The Yankees were totally overblown in the media as an unstoppable force. No one respected the Tigers going into this year, so no one had serious reservations about writing them off when they tanked late in the year. (By the way, if you think the '05 White Sox and '06 Tigers are similar now, wait until you see how much differently the media treats the Tigers entering the 2007 season. Remember how the White Sox were the favorites to repeat this year, after being rated the 5th or 6th most likely team to emerge from the '05 playoffs?)

Really, in the end, it came down to momentum. The Twins didn't play any better than the Tigers this year, but they ended the season hot, so they were rated as 5-1 dogs to win the World Series rather than the ice-cold, 20-1 Tigers. (Obviously some of this was also because the Tigers faced the heavily favored Yankees in round 1.)

But the smart people realized the Tigers had more than a puncher's chance. This is baseball, after all, and no team is realistically going to be much more than a 2-1 dog in a 5-game series unless it is an utter mismatch like last year's Cardinals-Padres NLDS.

Congrats to the Tigers, and I hope Jim Leyland doesn't screw up their march to the title by batting Neifi Perez second again or anything.

Saturday, October 14, 2006

Tigers Wasting At Bats

After starting Marcus Thames at DH for the entire series against the Yankees, Jim Leyland now appears committed to Alexis Gomez, starting him two out of three games against right-handed pitchers, with Omar Infante inexplicably taking the third start.

Thames may have acquired the label of being a lefty-masher, but his career platoon split is a difference of only 28 points of OPS, a fairly low number. The Fox talking heads quoted Jim Leyland saying something about how Gomez is going to be good, but needs seasoning. Setting aside that the playoffs are not the place to get seasoning (with one exception), Gomez is now 28 years old. He's not getting any better, and his current self isn't good enough to play every day, with a career OPS of .649.

In four games of replacing Sean Casey, Leyland has now used Ramon Santiago three times and Neifi Perez once. It should be obvious to anyone who can read a stat line that Infante is a better hitter than either, and he's a fine fielder to boot. Perhaps this, and not the omission of Chris Shelton from the postseason roster, is the most dangerous consequence of carrying three middle infielders: that Leyland will actually give regular playing time to the two unnecessary ones.

Tiger fans can hope that Shelton is playing first base and Thames DHing in the World Series, and both of these problems will correct themselves.

Friday, October 13, 2006

Tony La Russa waffled and finally decided to start Chris Carpenter tonight on full rest, allowing him to pitch game 6, also on full rest. Without the rainout Wednesday, Carpenter would have been forced to start games 3 and 7 unless he went on three days' rest. The talking heads on Fox and ESPN are discussing how the rainout helped the Cardinals by allowing them to move Carpenter up in the rotation.

The rainout was a Greek gift. Carpenter should have remained on schedule to pitch in games 3 and 7. By using him in games 2 and 6, they've all but eliminated the chance he can start three games in the World Series, a change which would have improved their chances of winning the WS by about 4% if they had won the NLCS in 5 or 6 games. Moving Carpenter up, in contrast, increases the chance of winning the NLCS by a paltry 0.4%. The result is a net loss of 0.4% to the Cardinals' WS title hopes, a significant hit when their total chances are under 12%.

You'll probably hear, or have heard, a lot of talk of momentum and how teams rarely come back from 2-0 deficits, so the Cards need to put Carpenter in now to avoid the 2-0 hole. Obviously, teams in 2-0 holes will rarely come back to win. If the teams are evenly matched, the team down 2-0 will need to win four of the next five games. There are 32 possible outcomes of a five-game series, assuming they continue playing even after one team has clinched. Out of these, the team in the 0-2 hole will win four or five only 6 times, so they are more than a 4-1 dog. Historically, teams in an 0-2 hole have performed in line with these numbers. As Fox displayed in a graphic during today's ALCS game, plenty of teams have come back from 0-2 deficits. It is not the end of the world.

Hypothetically, let's say the Cardinals automatically win Carpenter's start and lose Suppan's. After game 3, is there any additional value in having won game 2 instead of game 3 if you're down 2-1 either way? If anything, wouldn't you rather have the "momentum" of having won your last game? Furthermore, an extra day of rest is more likely to help Carpenter, who has tired down the stretch, than Suppan, who has pitched much better in the second half.

Moving Carpenter up wasn't extremely costly, but the Cardinals can't afford any slip-ups at this point.

Thursday, October 12, 2006

Many people are still defending Jim Leyland's use of Joel Zumaya, even in light of his injury. I'll attempt to clarify my position.

- Leyland did not injure Zumaya. He simply increased the change Zumaya would be injured, perhaps by 1-2%.
- Despite this, this added injury risk, combined with the possibility of Zumaya being unavailable for future games, made it a poor decision to bring him in with a 6-run lead and a 5-run lead.
- It does matter that he was injured while being used in a pointless appearance.

Let's look at an analagous situation from the world of poker. Say you're playing a hold 'em tournament. You have $10,000 in chips in front of you, and you think your opponent is very unlikely to have a hand that can call a $10,000 all-in bluff bet. If there is already $20,000 in the pot, bluffing is a very attractive proposition; the vast majority of the time, you will win the $20,000. There's a small chance you will lose your $10,000, but overall the bet is a winning proposition.

This situation is similar to using Zumaya in a close game. You're risking the very small chance of an injury and a small chance that he will be unavailable in a critical spot tomorrow, but the potential reward is high.

But what if the pot contains only $100? Now betting all $10,000 is a foolish idea. You will usually win the $100, but $100 isn't that much compared to the amount as risk, and the times he calls and takes your $10,000 will make it a losing play in the long run.

This is more like using Zumaya in, oh, let's say a 6-0 game with two outs in the eighth.

Yes, it was unlucky that Zumaya was hurt, just like you would be unlucky to get caught bluffing. That doesn't mean it wasn't a mistake to put him out there.
Rob Neyer wrote today that the Tigers' chances of winning the World Series are "nearly 50 percent." As usual, his math is pretty good, even if he was just throwing a guess out there.

The WS race as I see it:

Tigers 48%
Mets 26%
Cardinals 15%
A's 11%

The Cardinals' best hope is to win the NLCS without having Carpenter go twice, thus freeing him up to pitch games 1, 4 and 7 of the WS.
Dusty Baker is not a good manager by any stretch, but as much as he had to do with the injury-plagued careers of Kerry Wood and Mark Prior, he had no choice but to get a lot of innings out of them in 2003, especially in the playoffs. When your team is in a tight playoff race, you need to get your best arms out there when it counts.

Jim Leyland, on the other hand, contributed to Joel Zumaya's sore wrist by making him throw when the team didn't really need him. 6-run and 5-run leads do not call for your ace reliever with two rounds of playoffs to go.
The Tigers showed off the fruits of Jim Leyland's earlier decisions in Wednesday's game.

- With Chris Shelton unavailable and Dmitri Young released for "performance-related reasons", Sean Casey was replaced in the lineup by Neifi Perez.

Perez, besides having a legitimate claim to the title of worst hitter in MLB history, is clearly not even the best middle infielder on the team. Omar Infante started16 games at DH this year. Setting aside the idiocy of handing the DH spot to your backup middle infielder, it is clear Leyland knows the guy can hit a lot better than Neifi.

- With Joel Zumaya having been used to "protect" a 5-0 lead in the eighth inning yesterday, Leyland turned to inferior relievers Wil Ledezma (7-4 Tigers lead, sixth inning) and Jason Grilli (7-5, seventh inning).

Neither of these was an extremely critical spot, but consider the difference between the following situations:

- 5-0 lead, bottom eighth. A's score a run. Their win probability increases from 1.6% to 4.7%. If they go scoreless, it is reduced to 0.5%.
- 7-5 lead, bottom seventh, two outs. A's score a run. Their win probability increases from 14.3% to 27.9%. If they fail to score, it goes down to 12.7%.

If you're counting at home, this means preventing a run is worth nearly four times as much in situation 2, since you have cost yourself 15.2% of a win instead of 4.2%. You can talk all you want about establishing the tone of the series or picking up momentum by icing down the first game; the bottom line is that intangibles do not increase the value of your appearance by 260%.

Also of note is that Fernando Rodney and Todd Jones each pitched an inning despite having thrown yesterday. If Leyland feels this game was close enough to get two of his three best relievers in there, why not the best one? Is he saving him for later with a two-run lead, but not with a five-run lead?

It doesn't make any more sense than using the first-base spot in your lineup to bat the worst active hitter in MLB.

Tuesday, October 10, 2006

Jim Leyland may win manager of the year, but he isn't getting my vote. Despite being credited with the renaissance of a franchise long ago left for dead, he's made some basic mistakes that may take center stage in games 2-7 of the ALCS:

1) Leaving Chris Shelton off the playoff roster to keep three utility infielders
2) Refusing to trust the back of his bullpen to hold any lead

Now, with Sean Casey hurt and Shelton unavailable, the Tigers' contingency plan is to move Carlos Guillen to first and give the shortstop job to Omar Infante or Ramon Santiago. Essentially, they're using the first-base slot in their lineup for Infante or Santiago, both of whom hit, well, like backup middle infielders. Meanwhile, Neifi Perez, who took Shelton's postseason roster spot, will stay on the bench without anything to do.

Shelton was left off the roster, in essence, because he hasn't hit well since April. His stats for this year, while not as good as his breakout 2005, are still above-average at first base once you adjust for park factors.

Ah, but Shelton hasn't hit well lately. I've got news for you, Jim: Sean Casey has been terrible since donning a Tiger uniform. In fact, with Detroit, he's hit worse than post-April Shelton. So we can't chalk this up to being fooled by a small sample size. Clearly, Leyland is only willing to trust grizzled veterans with clutch situations.

There's a problem with that conclusion, however: it directly contradicts his philosophy of bullpen usage, which is that no lead is ever too big to bring in his 21-year old relief ace. Consider how he has "protected" his big leads in the playoffs thusfar:

ALDS Game 3: 6-0 lead through 7.2 innings. Brings in Joel Zumaya, who threw 2.1 innings yesterday, to get out of the eighth, then closer Todd Jones, who also worked the day before, to pitch the ninth. Never mind that this is the second of four consecutive days with games scheduled.

ALDS Game 4: 8-1 lead through 8.1 innings. Brings in Jamie Walker, his best left-handed reliever, to face the lefty-heavy Yankees. This one isn't really a big deal, because it was an elimination game and they had two days off afterward.

ALCS Game 1: 5-0 lead through 5 innings. Brings in his second-best reliever, Fernando Rodney, to pitch the sixth and seventh, Zumaya for the eighth, and Jones for the ninth.

To give you an idea of how this type of decision-making should be made, we'll use win probabilities from walkoffbalk.com. The Tigers' chances of winning entering the bottom of the eighth inning of Tuesday's game were 98.4%. By bringing Zumaya into the game, they perhaps increased their chances of winning to 99%, while they would have been lowered to 97.5% with Jason Grilli on the mound. So the decision added 1.5% to Detroit's chances of winning. (In reality, it didn't even do that, as Zumaya gave up a run.)

It seems like smart managing, until you realize that this inning restricted either Zumaya's availability or his effectiveness in future games, because there is a diminishing return on each additional inning you ask your relievers to pitch. If Zumaya tires later in the playoffs during a close game, it will largely be blamed on his youth and heavy workload during the season, but it may actually happen because Leyland won't save his arm for when it is really needed.

The whole idea behind optimal reliever usage is to save your best pitchers for when you really need them. In other words, you want them in there when one run will often change the result of the game. These situations occur only when the game is close, and more often later in the game. So far, in five games, only one of which was anything resembling close beyond the sixth inning, Leyland has used the back end of his bullpen to throw exactly two pitches, both by Jason Grilli in ALDS Game 1 to retire one Yankee batter.

If the game is 4-3 Tigers in the seventh Wednesday, will Zumaya have enough left in the tank to go for two innings? If he does, will he be good for another two innings Friday? Regardless, his effectiveness in those games will be affected by his workload up to this point. Leyland should take a lesson from Ken Macha, who used everyone but his two best relievers (Justin Duchscherer and Huston Street) to get out of Game 1. If you're never going to use your worst relievers, then take them off the roster and use that spot to keep Chris Shelton on the roster. Doing anything else is just a waste of roster spots, which are a valuable resource.