Batter's Box Interactive Magazine Batter's Box Interactive Magazine Batter's Box Interactive Magazine
There has been a lot of debate about high pitch counts for starters and injury. I thought that it might be useful to step back and look at performance.


Baseball Reference contains information about pitcher performance during the 1-25, 26-50, 51-75, 76-100 and 100+ pitch ranges. I was particularly interested in how starters did during the 76-100 range. Usually if a starter has gone 75 pitches in an outing, it means that he didn't come out of the bullpen with nothing and been trashed in one or two innings. This means that the starter's worst starts are not considered at all. On the other hand, 76-100 pitches is generally not thought to be the point of tiredness. So, I looked at the career line with 76-100 pitches of every starter in the American League who had been in the league at least 5 seasons. Here are the results:

The great ones

Pitcher PA Opponent's batting line
Halladay 1328 .271/.320/.388
Maddux 3148 .252/.294/.328
Santana 917 .233/.289/.398
Schilling 2512 .247/.289/.388
Clemens 3525 .224/.289/.341
Mussina 2759 .262/.295/.416

The very good ones

Pitcher PA Opponent's batting line
Burnett 1011 .237/.325/.398
Beckett 851 .243/.315/.410
Sabathia 1104 .248/.309/.375
Pettitte 2140 .291/.338/.427
K. Rogers 2364 .282/.340/.443
Buerhle 1476 .281/.322/.444
Lackey 1138 .301/.365/.445
Escobar 1134 .253/.321/.386
Colon 1750 .264/.330/.425
Millwood 1686 .269/.325/.421
Bonderman
756
.267/.333/.426
Haren
687
.256/.305/.435

The rest

Pitcher PA Opponent's Batting Line
Byrd 1121 .291/.335/.470
Westbrook 850 .257/.310/.364
Cl. Lee
714
.297/.349/.529
Robertson 621 .261/.337/.446
C. Silva 559 .331/.366/.544
Ponson 1338 .289/.354/.453
Garland 1252 .296/.350/.482
Vazquez 1769 .264/.309/.450
Contreras 745 .250/.324/.409
Meche 969 .278/.352/.467
O. Perez 872 .314/.363/.514
Washburn 1418 .264/.315/.444
Batista 1108 .265/.344/.426
JfWeaver 1512 .285/.336/.445
Padilla 852 .282/.344/.446
Ohka
699
.291/.351/.468
VZambrano
535
.236/.354/.383
Towers
466
.298/.338/.472
Maroth
760
.316/.369/.482
Seo
427
.309/.352/.494
RaOrtiz
1144
.267/.339/.491
HoRamirez
492
.281/.357/.514
Kennedy
730
.304/.372/.521
Loaiza
1668
.302/.353/.462

For comparison purposes, the AL averages over the past 5 years are .270/.336/.428. It is pretty clear why managers want to continue sending out the very best pitchers in the game for more than 75 pitches. They are better than the alternatives. But, for all the others, the average reliever will pitch as well or better. Jason Frasor, for example, has been a better pitcher over his career (1-25 or 1-50 pitches) than A. J. Burnett has been after 75 pitches. And in the case of Carlos Silva, for example, almost any pitcher in the Minnesota pen over the last 5 years has been better.

So, what does this data suggest? To me, it suggests that the governing idea of modern starting pitcher usage , the attempt to get 5 starting pitchers to throw between 90-110 pitches each start, is foolish and inefficient. It is a legacy of a different age in baseball, when a complete game by a starting pitcher was a realistic possibility every outing, and consequently "pitcher wins" actually meant a lot more than they do today. Now, managers try hard to give their starting pitchers the opportunity to "win" the game at the expense of the team's likelihood of actually winning.

There are any number of alternatives- 4 man rotations (with each starter limited to 70-80 pitches), 3 day rotations, 5 day hybrids (with the top 1 or 2 starters scheduled to throw as currently and the others paired off and expected to throw roughly 60 pitches each). To implement any of these alternatives, the shape and usage of the bullpen would have to change too. I would not shed a tear about that.

Starting Pitcher Fatigue | 13 comments | Create New Account
The following comments are owned by whomever posted them. This site is not responsible for what they say.
Mike Green - Sunday, September 30 2007 @ 10:11 AM EDT (#174898) #
Here are the 2007 team totals for pitches 76-100:

Team Opponent's Line
Bos .271/.325/.462
NYY .301/.361/.453
Tor .240/.297/.406
Bal .254/.329/.418
Tam .269/.333/.467
Cle .280/.340/.443
Det .297/.366/.469
Min .275/.325/.474
CWS .282/.333/.432
KC .293/.348/.472
LAA .294/.353/.445
Sea .303/.372/.514
Oak .300/.355/.472
Tex .329/.398/.521

On average, the starters are clearly significantly below league average by pitches 76-100.
Mike Green - Sunday, September 30 2007 @ 11:39 AM EDT (#174902) #
To put  numbers on it, the league-average starter's opponents hit .285/.345/.463 between pitches 76-100 in 2007.  That is pretty terrible.

The Jay hurlers were the exception in 2007.  They pitched better from pitches 75-100 than earlier in the pitch count.  That was not true in 2006, when they were .288/.349/.460, nor in 2005 when they were .273/.335/.430. 

AWeb - Sunday, September 30 2007 @ 11:47 AM EDT (#174903) #
There are any number of alternatives- 4 man rotations (with each starter limited to 70-80 pitches), 3 day rotations, 5 day hybrids (with the top 1 or 2 starters scheduled to throw as currently and the others paired off and expected to throw roughly 60 pitches each). To implement any of these alternatives, the shape and usage of the bullpen would have to change too. I would not shed a tear about that.

Starters for Toronto will pitch about 1000 innings this year with 993 so far. Relievers picked up the remaining 446.2 innings. Toronto likely leads the league in starter innings pitched. A few basic ideas : relievers do well because they get to concentrate their innings and give max effort for short periods. If relievers have to pitch more often and/or longer each time, they will certainly get less effective in those roles (one example this year : Scott Downs, now fully converted to a lefty specialist facing less than three batters per appearance, has never been close to this effective). The flip side of this is giving the starters a shorter time to pitch would make them more effective, since they no longer need to conserve energy. My basic thought is that if you implemented a system that would give more innings to the relievers and fewer to the starters, you would end up wondering why the hell you wouldn't let the starters (who may end up out-performing the relievers) pitch more innings. The success of the modern bullpen is also driven by the relentless pursuit of the platoon advantage, which is extremely annoying to watch and also quite effective.

If you had 2 ace starters and 6 half-time starters (3 pairs), that's 8 guys in the staff accounted for. The ace starters will still need at least 3 relievers behind them, since they won't be finishing games more than they do now (and throwing away platoon advantage switches late in games would be a step back) .  You'd need another long man to cover for the aces, probably two to cover for the half-starters in case of ineffectiveness (so maybe 8-9 "long men" out of the pen). A blow-out loss could throw the whole pitching staff into flux for weeks (teams would then use AAA as a place to stash guys for 2-3 days at a time to make sure fresh pitchers are available, in essence making pitching staffs much larger and confusing).  It could be interesting though.

To maximize effectiveness for the half-starters, you'd want several lefties to split the time, which could be a major advantage of a different system. Managers would have a much more challenging job, both in game and in managing egos, which I would enjoy from a fan's point of view. Veteran players (and many youngsters) would never accept it and just refuse to pitch for a single team trying this.

I would try a less extreme version of this type of thing...have just one split starter spot (3-5 innings each) in the rotation. Next year, perhaps Litsch/Chacin or Janssen/Tallet (just throwing names out there) could be in this sort of role, which would keep a pitcher "stretched out", in the majors, and ready to fill in once the inevitable Burnett injury strikes, with another reliever pulling into the transitional starter role. Baby steps towards searching for a better way to manage pitching staffs. But I think if you couldn't make one split-starter situation work, there's no point in trying to do it for the whole staff.

Mike Green - Sunday, September 30 2007 @ 02:06 PM EDT (#174904) #
The 4 day rotation works like this.  Let's say that you have 3 lefties in the pen, who we will name Downs, Romero and Ryan, and that you have 4 righties in the pen who we will name League, Janssen, Litsch and Accardo.  Your 4 starters are named conveniently Halladay, Burnett, McGowan and Marcum.

Four relievers, two left and two right, are paired with the starters, say Halladay/Janssen, Burnett/Downs, McGowan/League and Marcum/Romero.    The paired relievers go anywhere from 2-4 innings depending on the game situation when they enter.  Ryan and Accardo deal with the late inning close games.  Litsch appears in blowouts and extra-inning games.  At 70-80 pitches, the starters should average just under 5 innings, allowing for the odd early blowout, or just under 200 innings a year.

The problem with the conventional notion is that the 3rd-7th starters actually are on average hardly better (if, at all) than the average non-closer relief. 

Dave Rutt - Sunday, September 30 2007 @ 04:45 PM EDT (#174906) #
Wouldn't a team with a 4 man rotation have an 8 man bullpen?
Mike Green - Sunday, September 30 2007 @ 05:13 PM EDT (#174907) #
You don't need a 12 man staff.  11 is plenty.  Once you accept the idea that your starter will throw 70-80 pitches and go about 5 innings if all goes well, your assigned "middle man" will go 2-4 innings depending on the score, your two ace relievers will cover key situations in the 8th, 9th and 10th innings, and you have a mop-up man, you have freed up a roster spot for a position player.  It's a significant bonus.

Another bonus results from better flow if role changes are needed.  If one of your starters is injured, it is easier to promote one of the middle men who is regularly working 2-4 innings than the current method of converting a conventional reliever.
AWeb - Sunday, September 30 2007 @ 06:05 PM EDT (#174908) #
Once you accept the idea that your starter will throw 70-80 pitches and go about 5 innings if all goes well, your assigned "middle man" will go 2-4 innings depending on the score

The system falls down when either the middle man or starter struggles badly more than once or twice in a row. You need the flexibility in the bullpen to have consecutive blowouts (or play musical callups). Litsch can't mop up 3-4 innings twice in a row, or more than three times a week.  Which means either the starters need to be able to go longer if necessary (less reasonable to expect as they grow more used to the proposed system), or there needs to be another guy ready to go (rotate the pairs around I suppose, but if you prepare for pitching every four days, don't expect great things if they're forced into a game a few days early).

Also, as I noted above, relievers do better than starters because they don't have to pitch as long each time, and the matchups are optimized for them, often to an infuriatingly ridiculous degree. Sure the starters should improve with only 80 pitches to worry about, but the relievers get worse. I'm not convinced it's worth the tradeoff, which is why I'd like to see something like this run on a small scale (1 rotation pair) before attempting a larger system change. With theoretical pitchers, not already trained and used to the current system, it might make more sense.
Mike Green - Sunday, September 30 2007 @ 08:16 PM EDT (#174912) #
The early collapse by a starter doesn't happen often.  Burnett and McGowan each went out once in the first 2 innings, and Halladay and Marcum not at all.  Marcum went out in the 3rd or 4th inning 4 times, but all you do in that case is extend the middle guy, and Litsch pitches maybe an inning.  If you have the 4/4 system, you have fewer starts by guys like Ohka and Zambrano who can really kill a bullpen. 

The whole idea is that the middle-men pitch longer and less often.  There is less toll on the arm from having to constantly warm up and sit down.

AWeb - Sunday, September 30 2007 @ 10:39 PM EDT (#174913) #
The whole idea is that the middle-men pitch longer and less often.  There is less toll on the arm from having to constantly warm up and sit down.

I wouldn't be so worried about the starter collapsing, which as you point out is fairly rare. I'd be more worried about the second guy (the 5th-8th best long men in available) getting knocked around fairly often. A 4-man rotation could expect about 40 starts each. This puts the starters near the same number of innings each (200 or so, although with one less starter, so in all the total would only be 800 rather than 1000 in a year). The secondary guys have to throw about 120 innings each. Actually, the Jays would be in a good position to give something like this a try, as they have 8 pitchers that you might want to pitch that much, unlike some teams. It would be the type of move that many of us hoped JP would be willing to try : taking a strength, and rather than leaning on it and standing pat, trying to maximize it.  I wouldn't be terribly upset if the Jays decided to adopt something like you propose, although I would be shocked.

I'm not sure about the warrming up causing a great deal of wear and tear on pitchers; is there any hard information on that sort of thing? Since some pitching coaches recommend daily throwing anyway, and warming up is not done at 100%, I might be persuaded that warming up to pitch often could actually be helpful, at least in moderation.
halejon - Monday, October 01 2007 @ 02:58 AM EDT (#174914) #
Assuming he ever starts for us again, I would love to see this sort of experiment done with Chacin/Tallet. Over his career, Gus has been great for the first 25 pitches, hung on for the second 25, and then totally lost it for the third (.296/.357/.467). It's not a statistical fluke- you can see him start to lose it and yet there's a lot of pressure to do whatever it takes to get him through 5 and have a chance for the win (or 6 and give the bullpen a break) no matter the situation. He's going to tax the bullpen anyway, may as well make it official and have someone ready to go long ahead of time.
Newton - Monday, October 01 2007 @ 02:40 PM EDT (#174922) #

I'd prefer a much simpler alternative where  the 5th starter only pitches when necessary to keep the others on regular rest .  ie Doc, AJ, McGowan, and Marcum go every fifth day with the 5th guy being skipped wherever off days and rainouts allow.  The 5th starter would likely make 15 or so starts under this scenario. 

If you go with 4 guys on pitch counts they'll end up with fewer wins (these guys and their agents all like them to get credited with the W) and who knows how their arms would handle being revved up every 4 days.   

Don't forget the Marcum conundrum; what to do with a guy on a pitch count chucking a gem.

If your 5th starter never gets into a groove under this system, too bad, he's the fifth starter and probably have had middling numbers or worse in any event. 

R Billie - Monday, October 01 2007 @ 06:14 PM EDT (#174927) #

I think the idea that a starter has to be given an opportunity to win the game for himself just to pad his personal stats is one of the most destructive philosophies the Jays have followed under Gibbons.

I can remember one game in particular which clearly illustrates the problem with this...Gustavo Chacin was facing Minnesotta and his command was clearly all over the place.  He was allowed to stay in the game to face Torii Hunter with two men on and two outs in the 4th because the manager wanted him to have a chance at the win.  The fact that there was a good chance the win would not be decided in the 5th inning of a close game was seemingly beside the point.

Hunter went on to hit a three run homerun and the Jays never recovered in the game.  This is a definate example of a manager negatively influencing the ability of his team to win games when calling a fresh armed righty out of the pen would have likely gotten them out of the inning just fine.

At the end of the year when everyone is counting up the Blue Jays 83 wins, is anyone besides Gustavo Chacin looking back at a game in May or whenever it was and remembering the HE was rewarded the win in a game rather than a teammate out of the pen?  This practice is (as a opposing team's announcer said at the time) clearly putting the welfare of one guy over the welfare of the team and the organization.  And not in a meaningful way but in a very superficial way.

I think the insistence of most teams to continue with 5 man rotations and an insistence on getting as deep as you can into the game with the starter is the reason so many poor quality innings are thrown.  For instance, I'm not sure why AJ Burnett stayed in the game yesterday after it became clear the Rays were sitting on his fastball and hitting it hard.  It took 5 runs scored and a change from a 2 run lead to a 3 run deficit to finally get him pulled.  I wonder if this was May rather than the last game of the year whether Gibbons would have been much quicker with the hook.

The philosophy that the starter HAS to be left in as long as possible makes no sense.  Baseball should always be a meritocracy.  A pitcher stays in if his performance and the situation warrants it.  The first 5-6 innings have nothing to do with the next 2 or 3. 

CeeBee - Monday, October 01 2007 @ 08:09 PM EDT (#174929) #

"I think the idea that a starter has to be given an opportunity to win the game for himself just to pad his personal stats is one of the most destructive philosophies the Jays have followed under Gibbons."

But it's not only the Jays or Gibbons. Every team in baseball follows this philosophy and until the method of handing out wins is changed this philosophy is unlikely to change as well.

Starting Pitcher Fatigue | 13 comments | Create New Account
The following comments are owned by whomever posted them. This site is not responsible for what they say.