User Comments, Suggestions, or Complaints | Privacy Policy | Terms of Service | Advertising
Page rendered in 0.9596 seconds
62 querie(s) executed
You are here > Home > Primate Studies > Discussion
Primate Studies
— Where BTF's Members Investigate the Grand Old Game
— Where BTF's Members Investigate the Grand Old Game
Tuesday, February 19, 2008
The Case for the Six-Man Rotation
The New York Mets will be going into the 2008 season with the most expensive pitcher of all time. But in order to finally get to the World Series after two years of late-season futility they will have to set another first: the six-man rotation.By adding Johan Santana and his $150 million salary over seven years (provided that options are picked up) the Mets have arguably built the strongest rotation in the National League. But even with Santana, the Mets management will have to be creative in handling its starters in order to give itself a chance come September and October. Although Santana has logged at least 219 innings in each of the last four years the rotation will have to be treated with care. Because Santana isn’t the problem, they guys that follow him in the rotation are.
In 2006, the Mets’ run to the Series was halted by two ninth-inning runs by the Cardinals in game seven of the National League Championship Series. That they even got to the NLCS was a small miracle considering the injuries to Pedro Martinez and Orlando Hernandez. The Mets’ dearth of pitching in the play-offs, which did bring on the emergence of John Maine and Oliver Perez, got so bad that they strongly considered starting veteran lefty Darren Oliver, who hadn’t made a start all year, in that fateful game seven.
After a season in which the now 36-year old Martinez made only five starts (that’s almost $3 million a start, folks) and Hernandez, whose dubious age might well qualify him for senior citizen fare on public transportation, came undone during the second half, the fragile twosome will have to be treated with more care than ever.
The rotation’s two young studs, John Maine and Oliver Perez fizzled during the second half of last season, as did the rest of the ostensible Santana, Martinez, Maine, Perez, Hernandez rotation.
Pitcher: First Half: Second Half: Johan Santana 11-7, 2.65 ERA 4-6, 4.45 ERA Pedro Martinez DNP 3-1, 2.57 ERA John Maine 11-5, 3.05 ERA 4-5, 5.65 ERA Oliver Perez 10-8, 3.31 ERA 5-2, 4.15 ERA Orlando Hernandez 7-4, 3.12 ERA 2-1, 7.65 ERA
So why not add a sixth man to the rotation? In Jorge Sosa, who did a decent job as the emergency fifth starter last year, winning seven of his first eleven starts and top-prospect Mike Pelfrey (who should be groomed for taking over when Pedro or El Duque leave or retire – both are upcoming free agents) the Mets have two superfluous legitimate starting options. If they used the two to alternate in the sixth slot they could keep the other five starters fresh throughout the season.
And if this sounds like a stretch of the imagination consider this: most of the five starters made a number of starts on five days rest last year, with good results to show for it.
Pitcher: Starts on four days rest: Starts on five days rest: Santana 18 13 Martinez 0 5 Maine 14 14 Perez 10 16 Hernandez 10 10Perez, the only pitcher who pitched on five days rest more often than he did on four, was also the only one of this group who didn’t fizzle out in the second half and it is well documented that Martinez has been a superior pitcher on five days rest for years.
Willie Randolph has shown his willingness to adopt a six-man rotation before. In May 2005, he briefly carried six regular starters after Kazuhisa Ishii returned from injury. Yet he wasn't the first Mets manager to do so. In late July of 1998, Bobby Valentine experimented with the six-man rotation when he had nine starters at his disposal, anchored by Al Leiter and Rick Reed.
What might stand in the way of this innovation is the Mets’ own management. After spending all that money they will want Santana on the bump as often as possible. But having him throw every sixth day instead of every fifth will cost them only 1.75 wins per year. Santana has averaged 33.5 starts in his four years as a full-time starter. Throwing him strictly every sixth day would reduce that number to 27 starts. With a career winning percentage of .679, the Mets could expect Santana to win 4.5 of those 6.5 starts lost. His replacements, Pelfrey and Sosa have a combined career winning percentage of .422 which means they could be expected to win 2.75 wins in those 6.5 starts, leading to a net loss of 1.75 wins at the most per year. Although 1.75 wins would have made the difference between winning and losing the NL East last year, it would be a negligable loss in most years. Especially when you gain a well-rested two-time Cy Young award winner in the play-offs.
Looking at it from an economic point of view, the Mets lose $29.1 million by reducing Santana’s starts from 234.5 over seven years to 189, a loss of 45.5 outings or 19.4% of the starts the Mets anticipated getting when they negotiated the contract. That would most likely mean that this plan would get nixed by management who are conscious of the added exposure Santana’s appearances would bring to the ballclub. That would disregard the benefit of minimizing Santana’s risk of injury throughout his prime years – he will be 29 through 35 over the course of his deal with the Mets.
In fact, the net loss of the six-man rotation to the team as a whole would be five wins since the career winning percentage of the first five starters is .595, which means they would probably win 16 games if they split the 27 starts Sosa and Pelfrey would get between them. The latter two’s combined career winning percentage, which is likely to improve considering Pelfrey’s untapped potential, would be expected to yield at least 11 wins next year. The loss of five wins is acceptable when you consider that the five-man rotation projects to win 96 games (if they never got a no-decision) and the six-man rotation should expect to win 91 games. That is however, not taking into account the superior performance of a well-rested pitcher. This should offset the five wins lost (judging by past performance) by adding an extra slot to the rotation. Although the extra starter would rob the team of a reliever, the added rest should allow starters to throw deeper into ballgames, reducing the stress on the bullpen.
Yet some probably believe a six-man rotation would be foolish since a five-man rotation doesn’t even work. In 2002, Rany Jazayerli of Baseball Prospectus wrote: “The five-man rotation is a failure.” He argues that there’s insufficient talent out there to stock even a five-man rotation. Jazayerli also pointed out that the average pitcher in the four-man rotation era was more successful than his five-man rotation counterpart. And he didn’t believe pitchers are kept healthier by resting them more. His research found that “pitching in a four-man rotation is less damaging than pitching in a five-man rotation.” Yet these conclusions were based on a very large sample size with an average age of 29. With a rotation like the Mets’, where four-fifths is either very old or very young, the benefits of adding starters could be much greater.
The six-man rotation seems rather a natural evolution. Once upon a time the five-man rotation was considered odd after all. But now that 30 teams carry at least 11 pitchers each, there are roughly 350 pitchers in the big leagues at all times, more than ever before. As a result, the talent is so thinly spread that it makes sense to optimize its performance by extra rest. Sadly for the baseball purists, a six-man rotation would give starting pitchers an average of only 27 starts per season, compared to 32.4 for a five-man rotation, effectively ridding baseball of the charm of the 20-game winner.
Leander Schaerlaeckens
Posted: February 19, 2008 at 10:40 AM | 52 comment(s)
Related News: Sabermetrics
Related News: Sabermetrics
My Bookmarks
You must be logged in to view your Bookmarks.
Hot Topics
About Baseball Think Factory | Write for Us | Copyright © 1996-2008 Baseball Think Factory
User Comments, Suggestions, or Complaints | Privacy Policy | Terms of Service | Advertising
User Comments, Suggestions, or Complaints | Privacy Policy | Terms of Service | Advertising
Reader Comments and Retorts
Go to end of page
Statements posted here are those of our readers and do not represent the BaseballThinkFactory. Names are provided by the poster and are not verified. We ask that posters follow our submission policy. Please report any inappropriate comments.
Perhaps they would consider using a "modified" six-man rotation, because a strict six-man rotation makes no sense whatsoever for this team. You could make the argument that a rotation which features Santana, Maine and Perez at the top (pitching on "regular" rest) and with Pedro. Duque and "sixth starter" at the bottom (pitching on "extra" rest) would address the problem more effectively. This would allow the three 20-somethings in the rotation to still make nearly 100 starts between them, while leaving the other 60 to 65 to split between the older guys and a sixth guy to use as a placeholder.
The strict five-man rotation is a big enough mistake - a strict six-man rotation on a staff anchored by Johan Santana would be even worse.
At least he wasn't a bust for his entire contract like Pavano. Pavano isn't expected to pitch this year so his total as a Yankee is 111 innings. For 42 million dollars. That's 378k per inning.
Is this the worst contract of all time? I'm sure some players were total busts but as far as length/salary/failed expectations I would think it must be very near the top.
Since 2002, the only time Pedro posted a better ERA on 4 days rest when compared with 5 days rest is in 2004.
How much of the second half decline was related to fatigue, rather than a regression toward a normalish BABIP?
I mean look at the stats above; without a six-man rotation, the 5 starters still managed to have more starts on 5 days rest then on 4 (58 to 53).
If you are suggesting a spot start every once and a while to back up Martinez and Hernandez one day, that makes sense, but you don't need a six man rotation to achieve your stated goals.
The Mets have two pitchers that are very good against one side of the platoon, plus a closer that's never pitching before the 8th inning and rarely before the 9th. I see guys like Heilman and Sanchez and Wise getting very gassed and instead of having tired starters at the end of the season, you have a spent bullpen.
One of the best things about Santana is that you can count on 7+ innings in about 2/3 of his starts (20 in 2007, 24 in 2006) to rest that bullpen. Oliver Perez, with a good 2007, went 7+ only 11 times. Maine did it only 12 times. El Duque did it 10 times in 24 starts. Pedro is a huge question mark in terms of durability, both in individual games and also potentially missing starts during the season.
The only way I think you can afford a six-man rotation is if you press the starters to throw more innings per start. And I'm not convinced that doing that isn't worse than just having a quicker hook every 5th day.
Maybe for length but the Phillies did pay Danny Tartabull $2M for 7 Abs (6 OPS+) in 1997 dollars.
The money is a bit less than the Pavano contract(42 vs 33 mill) and AZ got slightly more innings out of him(111 vs 138) but that ERA is ghastly(7+) while Pavano at least put up a respectable 4.77 era with NY.
It's close but I guess I'd say the Ortiz contract is worse. Pavano was a huge injury risk and a questionable signing but from the outset the Ortiz contract was just moronic.
Here is how they could start:
Mar 31: Santana
Apr 1: #2
Apr 2: #3
Apr 3: day off
Apr 4: #4
Apr 5: #5
Apr 6: Santana
Apr 7: day off
Apr 8: #2
Apr 9: #3
Apr 10: #4
Apr 11: #5
Apr 12: Santana
Apr 13: #6 or #2 (on four days rest)
Apr 14: day off
Apr 15: #2 (on six days) or #3 (on five days)
Apr 16: #3 or #4
Apr 17: #4 or #5
Apr 18: Santana
etc.
This will create the "odd" situation that sometimes your #6 pitcher will follow your ace, but really, what difference does it make what order they pitch in, as long as you give the most starts to the ace, 2nd most to #2, etc.
I just went through the Mets season and started the best pitcher who had at least 5 days rest, and ended up with 30 starts for #1, 29 starts for #2-4, 25 for #5 and 20 for #6. Through August 5, #1 pitched every sixth day, then had to take an extra day off for an off-day on Aug 11, then another one Sept 4.
#2-4 would start on 5 days rest 22 or 23 times each. Only 6 times total would one of them start on more than 6 days (4 of those occurring right after the A-S break). So, there would be minimal impact to their schedule.
So why not this variation?
But having him throw every sixth day instead of every fifth will cost them only 1.75 wins per year.
I'm thinking 1.75 wins is a LOT, & not just because of how close they came to the divisional title last year.
It's unfortunate that Pedro had his awesome season in 2005 and was hurt in 2006-2007. Swap 2005 with 2006, and the Mets almost certainly go to the WS, given that the NLCS was so closely decided and the Mets were pitching on fumes. Swap 2005 with 2007 and there's no collapse.
The way this deal is viewed historically is going to depend very seriously on the 2008 season. I still think there's a small chance that Pedro can put up a real monster season in 2008. There is some hope that the surgery actually fixed something.
I think the next big change you'll see in pitching is the return of using starting pitchers as relief pitchers between starts. Particularly with a 6-man rotation, if you've got 5 days off between starts, couldn't you simply always make the pitcher who's on his 3rd day of rest available for bullpen work that day? Even with a 5-man rotation, simply having starting pitchers available to go an inning or two in relief on the days when they'd otherwise be throwing on the side gets you an extra bullpen arm without having to use up an extra roster spot.
The Dreifort contract is worse than both. ~ 205 IP, for a five year, $55 million dollar contract, signed in 2000, making the actual value of the contract much greater than either Pavano's or Ortiz'.
This is a baffling statement to me. 5 wins is HUGE.
If the Mets go boom like they did last year, everyone involved in this project gets fired, its not really reasonable to ask a team that thinks they are going to win a pennant to make a radical change. Now a team like Pittsburgh what do they have to lose? And maybe they somehow get to .500.
So far, Pedro has given the Mets 377.2 IP of 124 ERA+ pitching. A 180 IP, 120 era+ season from Pedro in 2008 and some playoff starts and I'd call that a good free agent contract for a pitcher, all things considered.
I imagine in a generation the new Mets fan statheads will look at the nearly $50 million the Mets paid Piazza to be an average catcher in 2003-2005 and decide that the entire contract was not worth it. But none of the Mets fans that were watching the team rise during that period would agree. There are deals where you can look back at the dollars per win and decide that they were financial mistakes, but if you asked the owner he wouldn't take them back in a second. I suspect that even if Pedro flops in 2008 his deal may be viewed in this light.
I played around with the schedule and assigned starters to each game, while taking into account off-days and necessary rest periods. The results look like this:
Santana: 32 starts (12 on full rest, 19 on five days or more rest)
Maine: 31 starts (10 on full rest, 20 on five days or more rest)
Perez: 30 starts (10 on full rest, 19 on five days or more rest)
Pedro: 29 starts (4 on full rest, 24 on five days or more rest)
El Duque: 29 starts (6 on full rest, 22 on five days or more rest)
Sixth Starter: 10 starts (never on less than five days of rest)
Only Maine and Perez make a start on four days rest in April. Santana makes his first such start in mid-May and both Pedro and Duque get all the way until mid-June before making their first four-day rest start. Both Pedro and Duque make only two starts on four days' rest before the All Star Break, then each gets rewarded with 10 days between starts. As far as I can tell, the only time the schedule calls for back-to-back starts on four days' rest for any pitcher is in the last two weeks of September - and that's only if the Mets are in a tight pennant race and can no longer afford to spread guys out.
The "sixth starter" is an interesting case. His first apearance would be on April 21 and he makes two starts before being sent down. He comes back May 15, makes three starts and gets sent down on June 2. In between, he can pitch relief in the May 20-22 series with Atlanta. The next call-up is on July 6, where he makes 2 starts and then gets sent down over the All Star Break. There's the need for a spot start on August 1 - this guy can either stay with team and pitch long relief from August 6 to August 14 or get sent down after the spot start. Finally, he makes three starts from August 19 to September 7 and is done for the season, never pitching on only four days' rest at the major league level. Sounds like the perfect role for Pelfrey, if you ask me.
Of course, this is contingent on having no injuries and no rainouts, so of course you would have to modify it throughout the course of the season. You'd also have to count on Willie Randolph being innovative enough to make it work, so I have no hope that this will do anything other than waste my lunch hour at work.
I am very surprised at the people predicting stuff like 20 starts or 150 innings as a good over-under for Pedro. This is a 36-year old pitcher who has had 2 years where he didn't carry an ERA+ of 125 or higher in his career. 28 innings isn't very useful in terms of predicting that he's back to being 1.5-2 runs better than average in ERA, but it is useful in showing that his arm isn't going to fall off.
I expect that he'll be better than 180 IP and 120 ERA+. If I had to guess, I would put him at around 200 innings of about a 3.00-3.25 ERA.
You have to be completely ignorant of historical perspective to bash the Piazza deal. The Mets were a laughingstock of a franchise, so bad that books were written about how poorly the team was run. Certainly Piazza was only a part of the solution, but he was definitely the face of the new Mets that demanded your respect.
Pedro Martinez could not be Piazza because the Mets really weren't bad for as long. But I definitely believe that the Pedro signing sent a critical message to fans, to the other teams, and to other free agents (like Beltran): the Mets were going to go hard after the very best talent, and pay what it took to get them signed. Without Pedro, I don't know that Beltran signs, and I think Santana looks very differently at waiving his no-trade at the 2008 Mets without those two guys.
I already look at Pedro's 4th year as a necessary evil. Pretty much every competitive "money" franchise would have been happy to get the first three years of Pedro's contract before hindsight told them he'd miss practically a whole season. The only reason six to eight teams didn't get into a bidding war was because the Mets were willing to go for that fourth year.
But I think if Pedro has anything close to an average Pedro year in 2008, people will consider this contract to be worthwhile ignoring those other issues considering how unreliable starting pitching is. Pedro's 2005-2008 contract will be compared to the ones Zito and Randy Johnson got (Yankees and 2nd D'Back contract); a good 2008 and it probably compares pretty favorably to both of them.
I think if Pedro has a strong 2008, he will sign another contract with the Mets as well. That contract might not be so well-regarded.
That's a Cy Young candidate.
Honestly, I expect Pedro to be really good when healthy because he has always been really good when he was healthy. That ERA wouldn't surprise me. But 200 innings is an optimistic projection to say the least.
What's crazy about that is a 3.25 ERA in 200 innings would be, compared to most years of his career, a "bad" year.
Honestly, I expect Pedro to be really good when healthy because he has always been really good when he was healthy. That ERA wouldn't surprise me. But 200 innings is an optimistic projection to say the least.
I think it's optimistic but not outrageous. In the past 4 years, Pedro had 2 years where he pitched 217 innings, one where he was hurt, and one where he was rehabbing from surgery. The last time anyone saw healthy Pedro was 2005 and he was well over 200 innings.
What really made the difference for me was not the quality of innings he threw on his return, but his ability to throw 90-100 pitches without showing serious fatigue. He was throwing a little less than a full game and still striking out batters in his final inning. I think he's healthy.
If we see this happen, it will last until the first big name goes down with an injury to his pitching arm. It won't matter whether it was related to pitching in relief--what will matter is somebody's 50 million dollar contract winding up in the toilet.
2007(1): 10-6, 2.75 2007(2): 5-7, 4.04
2006(1): 9-5, 2.95 2006(2): 10-1, 2.54
2005(1): 7-5, 3.98 2005(2): 9-2, 1.59
2004(1): 7-6, 3.78 2004(2): 13-0, 1.21
Looks to me like the 2007 season is the outlier in that crop; at a minimum, you have a real uphill battle convincing me that this is a guy who needs to be saved from his late-season fades.
Minn 2007 RS/ML Rank: 718/25th
NYM 2007 RS/ML Rank: 804/10th
It's going to be interesting to see if the Mets consider using Sosa as a spot starter in 2008. In his last 47 starts, he has a 107 ERA+ (4.03 ERA) and a 4.94 fips. I think Sosa can be valuable in a longman/swingman role for the Mets.
People and especially projection systems are skittish because he's a 36 yo who has not pitched a full season in three years. Coming off a shoulder injury. There's a significant chance he gets hurt again and misses a large portion of the season. I wouldn't say that's likely to happen, necessarily, but its very optimistic to just assume he'll be healthy all year.
Without Mike Piazza, the Mets would have no post-season appearances between 1988 and 2006.
Even worse, they were selling Tim Spehr jersys at the Shea Stadium concession stands.
Chan Ho Park says "Hi."
380 innings at below replacement level. And all for only $48 million over 4 years. I'm not sure if it's worse than Pavano, but it's in the running.
-- MWE
I think teams do that to some extent, but more in the form of avoiding starting certain pitchers against certain teams that give them trouble. I also think it's harder to do now than it was 50 years ago. Especially in the AL, there aren't many cakewalk lineups or heavily lopsided lineups one way or the other.
I think 2008 is going to be a very big shock to the system for Pedro believers. I think he's Jeff Suppan with major injury issues.
If I were thinking about a pitcher that is pretty much the opposite of Pedro Martinez, I'd pick Jeff Suppan.
Suppan is about a good a bet as anyone to make 30 starts in 2008. Pedro's health is in question.
Suppan has never averaged 5.5 k/9 in any season in which he pitched more than 100 innings. Pedro's one of the premier strikeout of his generation and was able to get strikeouts last season coming off surgery.
Suppan's a groundball pitcher. Pedro's a flyball pitcher.
I could go on but I am puzzled by this. What do you mean that Pedro is Jeff Suppan with health issues?
Since the reason for most of these strategies is to keep the best pitchers from getting injured by overwork, he probably just wished he hadn't had to worry about injuries.
I suspect that explanation will make more sense than the one he will give.
Your problem is that you are thinking of this as a comparison. In fact, Pedro Martinez is, quite literally, Jeff Suppan. There was a lot more to Pedro's "shoulder" surgery than most of the public is aware. The procedure was actually a complete DNA transplant. Of course, the external manifestations of this operation will take years to become apparent. But in almost all ways that truly matter, Pedro Martinez is now Jeff Suppan. With a bum shoulder though - there isn't anything anyone can do about that.
A key point, implicit in the article but not fully articulated, is that baseball players wear down psychologically. The Mets collapse last year was chiefly psychological, but the fact is that the bullpen was overused, and its physical weakness contributed to its panicky state, and this reverberated throughout the pitching ranks. The same phenomenon is observed in all walks of life. Soldiers are much more likely to panic and be routed when they are exhausted; a six-day rotation would help insulate the pitchers from the psychological stresses of a long, grinding season.
An ingenious idea, again.
Keep 'em coming.
More to the point, is there any evidence that modern pitchers suffer more psychological duress than the pitchers of most of history that pitched in four man rotations?
I'm an advocate of the return of the four man rotation, or rather, a modification of it; more to the point, I'm a big believer in tailoring your pitcher usage to the particular pitchers on your staff. I'm extremely dubious that a strict six-man rotation is the best way to do it, not unless you actually have six starting pitchers of roughly the same quality. You simply need to get more work out of your ace than the 29 or 30 starts a six-man rotation allows for, and that goes double when your ace is the best pitcher in the world.
In practice, a lot of teams already use some vague approximation of a six man rotation, especially in midsummer when off days get fewer and farther between, generally when somebody needs bumped back a day and/or because there's not much quality difference between the #5 guy and three other guys they have in AAA. It's not really a 'rotation', but teams have been in the habit for awhile now of working a non-rotation guy (from the bullpen or AAA) in now and again when everybody else needs an extra day's rest.
That's a totally different scenario. I prefer the latter solution, particularly after roster expansion, where you can have a much deeper pen. I want to rest guys like Heilman and Sanchez as well if the Mets are well ahead.
...
Perez, the only pitcher who pitched on five days rest more often than he did on four, was also the only one of this group who didn’t fizzle out in the second half
I'm confused...did Perez fizzle out in the second half? He was certainly worse, but as Danny pointed out, how much of that was due to defense (including his own)?
I'm always amazed at how poor the analysis is by these stat geeks. Leander wants to focus on one season, and announce he's figured it out. Well, Johan Santana is far better in the 2nd half throughout his career. 2007 was the exception. Pedro Martinez has not had huge 2nd half drop-offs either. John Maine's 2007 workload was far greater than in his previous seasons. I just think it was more like a rookie wall in his first full MLB season. Maine's endurance should be better this year.
You must be Registered and Logged In to post comments.
<< Back to main