Milwaukee Brewers Blog - Between the Green Pillars : Sabermetrics

Building The Perfect Bullpen

Ryan Topp — Mon, 17 Nov 2008 18:57:00 GMT

One of the longest held beliefs of the Sabermetric community is that the vast majority of major league managers are not using the assets they have in their bullpen to the greatest possible effectiveness. The major argument is that most teams use their best reliever as a "closer" save specialist, meaning they will be used to for 3 outs with a lead of 3 or fewer runs for the majority of their appearances. Some may get in for 4 or 5 outs from time to time and most will be used in "non-save" situations occasionally when they have been idle for a while and need to stay sharp. Most of the time, though, they get three outs in the 9th.

The problem with that usage is, of course, that very often those are not the key outs in a game. How many times has an inferior reliever been brought in to face the oppositions best hitter in the 7th or 7th inning representing the tying or fall behind run? How many times has the "closer" been asked to face the 7, 8 & 9 hitters in a weak lineup with a 3 run lead just to get a save? Shouldn't a teams best reliever be used to get the hardest and most important outs in games and not used in easier situations simply for the sake of adding a save to the total?

The arguments for the closer are not without some merit, especially the ones about guys wanting to know their roles and have a general idea of when they might be used. These are not machines being asked to perform a computation, after all. It is also true that very often the 9th inning does turn out to be the toughest situation in the game and that giving up a run in the 7th inning in a one run game isn't the same as giving up a run in the 9th inning of a one run game. Managers deserve some credit for coming up with this current bullpen useage system, they just haven't been flexible enough in it's employment.

With that in mind, how can the Brewers best build their bullpen for 2009 now that their 9th inning specialist from last year, Salomon Torres has decided to call it a career?

Assuming that Carlos Villanueva and Seth McClung are both headed to the pen (and Melvin is on the record saying that Villanueva will be and McClung might be) the team would be well advised to keep in mind that both of those guys are capable of going more than one inning at a time and that using them as a classic save specialist is probably not the best way to go.

The Brewers are going to need a LHP for the pen and would probably be best served with twol. The question is whether or not it's worth it to bring back Brain Shouse. At this point in his career, he's great against lefties and terrible against righties, which means that his uses are pretty limited in tight situations. Mitch Stetter is waiting in the wings, but his walk happy nature makes him less than a sure thing.

David Riske's ability to come back from his trouble filled 2008 will go a long way towards determining how the middle relief will shake out. The Brewers need him to step up and earn his paycheck this season.

There are veteran closers on the market this winter, like Trevor Hoffman and Jason Isringhausen, but they are also well past their primes and will probably command a premium for their names alone. If one of them could be had at a bargain price, however, it would allow the Brewers to go into spring training with an established situation in the pen, at least until the players prove their worth early on.

Besides Mitch Stetter, the Brewers will head to Arizona with several options from the farm system to take a look at. Tim Dillard has the stuff to be long term big league reliever, though he has yet to show that he is ready to pitch in key situations. Mark DiFelice is a long time minor leaguer who strikes out a lot of batters and walks very few, but his lack of velocity leaves him vulnerable to the long ball. Luis Pena is coming off of a disappointing campaign in AAA, but still has the electric stuff to pitch late in games if he can harness it. Omar Aguilar has top notch stuff and he's been groomed as a closer in the minors. If he can build on his impressive 2008 and cut down on the walks, he could find himself closing in Milwaukee before the end of the year.

Melvin said in a radio interview on 1250 WSSP last week that he was going to wait a while to address the bullpen situation, which means that he's probably not interested in any of the top guys unless their price comes down later on and that he's willing to go with internal options, at least to a large degree. There are some very useful arms already under team control both at the major and minor league levels. If Melvin's track record is any indication, though, he'll probably end up bringing in multiple cheap options for a look see in spring training. It makes sense, because the best way to build a bullpen on a tight budget is to acquire as many cheap options as you can, throw them at the wall and see what sticks.

Did they underperform?

badger80 — Wed, 08 Oct 2008 02:18:00 GMT

In a season of ups and downs, the offense caught the most flack for under performing. It was expected that Prince Fielder would again hit 50 home runs and Ryan Braun would be right there with him now that he was getting a full season to show what he could do. Corey Hart was going to be that power/speed combo that would give other teams headaches and Rickie Weeks was going to break out. Everyone expected Rickie Weeks to break out.

It didn't quite work out that way. The 2008 Milwaukee Brewers scored 750 runs this season, 51 fewer runs than 2007. How did this happen when everyone expected this team to take a step forward? Let's go back to the beginning and see if our expectations for the season were realistic.

You can find a number of different publications that do preseason statistical projections. PECOTA is the probably the most popular source and you can find that at Baseball Prospectus.com. Bill James publishes projections, along with John Sickels at Minor League Ball.com. I like to reference Baseball Think Factory's ZiPS Projections, mainly because they're free and easy to find.

For easy reference, here are the links to the Baseball Reference and the ZiPS projections:
2008 Milwaukee Brewers Baseball Reference
Milwaukee Brewers 2008 ZiPS Projections

Prince Fielder
Projection: .282/.380/.571
Actual: .276/.372/.507

Most fans were frustrated by Prince Fielder's power decline this season. He got off to a slow start, needing two weeks and a Joe Dillon bat to hit his first home run of the season. Unfortunately that didn't open the flood gates. His meager power output continued through May before he really started hitting with some authority again in June.

The ZiPS projections were a bit more conservative than fans would have hoped for before the season began. Dreams of a line-up anchor that could OBP .400 and SLG .600 dance through our heads since he almost did that in 2007. Still, a .380 OBP and .571 SLG aren't too shabby and these projections tend to be conservative. Prince came close to his OBP projection, but his power just never came around. He was projected to get 165 hits, 44 home runs and 35 doubles. Fielder almost matched the hit total with 162, but fell 10 home runs and 5 doubles short.

Overall, Prince's power decline was somewhat troubling for a team that relies on quick strikes to score runs, but he was far from the problem that was holding back the team.

Ryan Braun
Projection: .294/.332/.554
Actual: .285/.335/.553

The Brewers golden boy didn't put up the insane numbers he did his rookie season, but is would have been unrealistic to expect a repeat performance since Braun .365 BABIP was healthier than normal. Instead he was brought more into line with a .305 BABIP, and he still put up some pretty impressive numbers.

As you can see, ZiPS was pretty much on the nuts with their projection. His 39 doubles and 37 home runs were higher than projected (33 2B, 33 HR), and he also drew more walks than expected. The scary part is that Braun's intercostal strain clearly hurt his performance in the final month, so a healthy Ryan Braun probably would have surpassed his projection.

J.J. Hardy
Projection: .262/.318/.421
Actual: .283/.343/.478

I think most fans had to be really happy with what they saw out of J.J. Hardy this year. Unlike 2007 when he got off to a torrid start, J.J.'s season was more of a slow burn this year. He expressed that he was uncomfortable batting in front of the pitcher early in the season when the team experimented with batting the pitcher 8th and the punchless Jason Kendall 9th in the order. He found his swing and was eventually moved up to 2nd in the order where he really excelled, hitting for power and getting on base in front of Braun and Fielder.

Hardy is one of the few players who really blew past expectations. He hit for more power, showed more patience and overall showed the maturity we were hoping to see from the entire line-up. Hardy hit 9 more double and home runs than projected, who would have guessed that he would hit 4 triples? J.J. will never be a burner, but we may have to admit that he's faster than he looks.

Corey Hart
Projection: .289/.353/.518
Actual: .268/.300/.459

If computers could have broken hearts, they'd cry in their diary about Corey Hart. ZiPS, like every other blue collar baseball fan in Wisconsin, loved Corey Hart. What the heck happened? Everything started out fine in April and May, but there were some signs in June that Corey needed to make some adjustments. Hart's power spiked in June, but that covered up the fact that his OBP dove 50 points, from .342 to .292. That stuff fluctuates, but unfortunately for Corey and the line-up, he struggled to post a .300 OBP for the next three months before the bottom fell out in September. Hart struggled to post .173/.192/.245 and didn't hit a home run in September.

What changed? I don't know. None of the Brewers' hitters do a good job of recognizing balls and strikes or identifying pitches to drive, but Hart just fell into the abyss for the last month. Maybe it's in his head? Hart said he didn't like hearing the boos at Miller Park, and it was odd to hear that comment from him since he was well liked by the fans before September. We'll see if the grind just got to him in 2008 and the off season clears his head.

Ultimately, Hart's struggles really ground the offense to a halt. Hardy, Braun and Fielder were all getting on base and Hart was in position much of the season to put up some gaudy numbers. He ended the season with 91 RBIs, but that number is a pretty good example of why RBI totals are overrated.

Rickie Weeks
Projection: .254/.363/.422
Actual: .234/.342/.398

Rickie Weeks was just about every scout and analysts pick as the break out player of 2008. If you check out the comments on the Baseball Think Factory projections page, ESPN and Scouts Inc. writer Keith Law asks for the optimistic projection* for Rickie Weeks which turns out to be .278/.394/.501. Weeks ended 2007 hitting .273/.441/.553 in August and September, which gave hope that his short demotion to Nashville allowed him to figure something out and we'd see him come close to reaching his full potential.

It just never clicked for Weeks. As always, he shows great patience at the plate and probably does a better job with pitch recognition than any of the young core on the team. Rickie just seems to be one of those puzzles with a ton of tiny pieces that takes forever to put together. It'll be awesome when it's finished, but how much time do you want to invest putting it together?

Weeks didn't terribly under perform his projection, though he did benefit from platoon with Ray Durham the last two months of the season. Like most of the young hitters on the Brewers, Weeks had a significant platoon split where he hit .227/.319/.391 against righties and .250/.391/.414 against lefties.

The club needs to give Weeks one more season to right the ship before searching for other options. Unless they make a significant move, he's probably still going to be their best option to lead off, but he needs to be moved down in the order. Bat him lower in the order for the good of his own development even if it means placing a lesser option at the top of the order. He may need a change of scenery, and hopefully it's just in the Brewers batting order and not with another organization that will help him shake his funk.

*Projections are done with a bell curve and broken down into percentages that represents the chance a player will fall into that projection. The Optimistic projection is the top 15% of the bell curve. The projections used for this post were the mean or 50% projection. I do not claim to be an expert and suggest that Google exists if you'd like a better explanation.

Ben Sheets   ERA   WHIP   INN   BB   K
Projection:    3.72   1.172    145    27   128
Optimistic:    2.93   1.034    175   26    166
Actual:          3.09   1.149   198    47   158

Ben Sheets had a rough final two weeks in Milwaukee and may have thrown his last pitch in a Brewers uniform...though up to that point he really performed as well as anyone could have hoped. His strikeouts were down, and his walks were up, but he was by far the most effective pitcher on the staff until Sabathia arrived. Sheets really deserves as much credit as anyone for the Brewers first playoff appearance 26 years.

Yovani Gallardo was lost for the season on May 1, Jeff Suppan was a waste of $42M, Dave Bush and Manny Parra started slowly. Ben Sheets went out and put up his best numbers since 2004, and (if you add in his 2 innings from the All-Star Game) threw 200 innings. Even after he experienced discomfort in his elbow pitching against St. Louis in August, Sheets still managed a sub-2.00 ERA before his final start against the Cubs in the final weekend of the regular season.

It was really sad to see Ben's career in Milwaukee come to a close with a whimper. No matter what anyone says though, Ben Sheets' final season in Milwaukee was a success.

Brewers Pythagorean Record

Ryan Topp — Wed, 11 Jun 2008 02:20:06 GMT

Since we're now over two months into the season, it makes a lot more sense to really start looking at teams Pythagorean records to assess where they are likely to head.

For those who are unfamiliar with the concept, the basic idea is that by looking at the number of runs a team scores and the number they allow you can figure out roughly how many games a team "should" win if you take out some variables like run distribution, which owes a lot to chance. A more thorough explanation can be found here.

The Brewers recent winning streak at home, which included a few blowout wins, helped bring the team more into line with their Pythagorean record. You can find the current standings on baseballreference.com's league expanded standings.

Heading into play on June 10th, the Brewers Pythagorean record is 31-32, which is two games worse than their actual record of 33-30. A two game split at this point either way doesn't point to much of a difference, so the team is probably a win or two above where they "should be" but that can change quickly with a big win or a big loss.

Going forward, what this tells us is something pretty in line with common sense: The Brewers need to start outscoring their opponents more often. They are currently 11th in the league in runs per game and 10th in the league in runs allowed per game. If those numbers stay where they are, they can expect to end up right about where they are: around .500.

The key is going to be scoring more runs, because it seems like the pitching is probably doing about what you can ask from it with guys like Gallardo and Riske out and Gagne barely useful. They can do that, but will need to get on base more for the big flies to drive the team forward.

More support for batting Kendall 9th

Ryan Topp — Fri, 21 Mar 2008 20:38:22 GMT

As was reported earlier on this blog, there is some support out there for the general concept of batting your pitcher 8th. In that case, they looked at Tony LaRussa's decision to do so last year.

Now, Dan Fox over at Baseball Prospectus, has some more specific info on the Brewers situation, and it centers around Kendall's rather unique (ahem) abilities:

What makes it work in this case is that Kendall's combination of an adequate OBP and a woeful slugging percentage is not that different from the typical ninth spot hitter in the order. This means that when batting eighth he won't be driving in many of the four-through-seventh place hitters who reach base, but when hitting ninth he gets on base enough to allow Fielder and Braun opportunities to drive him in, resulting in a net offensive gain. It appears the break-even for this strategy as far as the Brewers go is somewhere around Kendall delivering a .360 SLG. That is, if Kendall's slugging percentage were .360 or higher (while delivering the same OBP (of .324) ), it would essentially be a wash as to whether to bat him eighth or ninth.

So when Yost talks about the "numbers" on this issue, he's not just blowing smoke. What remains to be seen is if he's going to stick to this or not after the first time the pitcher bats with two outs and runners in scoring position and the team goes on to lose. One thing going against Yost here is that the benefits of this move (Kendall being on more in front of the top of the order hitters) are probably going to be less readily apparent than the negative aspects of having the pitcher come up a spot earlier. There is going to be a lot of gnashing of teeth by the fans and possibly the media. Does Yost have the stones to stick with his plan, even in the midst of public criticism?

Finally, do you remember part of the original plan here was that Braun and Fielder (or Fielder and Braun) would move up a spot each and bat second and third? What ever happened to that?

Thanks to Badger80 for digging this up over at BP.

Expect the unexpected?

Ryan Topp — Fri, 21 Mar 2008 20:28:40 GMT

There has been a lot already made of Manager Ned Yost's decision to bat catcher Jason Kendall 9th this year, moving the pitcher up a spot to hitting 8th. The move is backed by analysis done by some in the Sabermetric community as having the potential to add a few runs over the course of a season. However, outside of some flirting done by Tony LaRussa when his teams were fringe contenders at best, no manager has had the juevos to actually try this in the midst of a pennant race.

Another topic along these lines that readily pops to mind is the way in which "closers" are used in the modern game. The model established by Tony LaRussa in the late 1980's with the Oakland A's is that a manager should use a series of relievers after a starter is pulled from a game, often taking advantage of match-ups, to be able to hand the lead to the teams best reliever or "closer" who almost exclusively pitches in the 9th, or maybe comes in for 4 or 5 outs on occasion. Many people, Sabermetircally inclined or otherwise, have pointed out that there are times when saving ones best reliever for the 9th inning while allowing a lesser reliever to blow a lead in the 7th isn't the best use of resources. But managers stick to this LaRussian model because it is the accepted way of managing.

This all begs several related questions: Should a manager use information gleaned from statistical analysis or otherwise like this, even if it falls well outside of the accepted orthodoxy of how you run a baseball team? Perhaps even more relevant to a fan: Should a fan be able to expect that a manager would use this information to try to get an edge over opponents, even if it means breaking with the accepted norms?

What do you think?