CSS Button No Image Css3Menu.com

Baseball Prospectus home
Click here to log in Click here for forgotten password Click here to subscribe
No Previous Article
No Next Article

October 14, 2004

Starters vs. Relievers

The Changing Distribution of Pitching Performance

by James Click

the archives are now free.

All Baseball Prospectus Premium and Fantasy articles more than a year old are now free as a thank you to the entire Internet for making our work possible.

Not a subscriber? Get exclusive content like this delivered hot to your inbox every weekday. Click here for more information on Baseball Prospectus subscriptions or use the buttons to the right to subscribe and get instant access to the best baseball content on the web.

Subscribe for $4.95 per month
Recurring subscription - cancel anytime.

a 33% savings over the monthly price!

Purchase a $39.95 gift subscription
a 33% savings over the monthly price!

Already a subscriber? Click here and use the blue login bar to log in.

Everyone loves plate discipline. Plate discipline allows hitters to control the strike zone, increasing his chance of getting a better pitch to hit. Getting a better pitch in turn increases the chances of putting the ball in play with authority rather than just simply putting it in play. Walks, and consequently the almighty on-base percentage, increase with plate discipline. Pitch counts increase, allowing teams to drive a starter from the game and get into the weaker parts of the bullpen.

This final point is a positive for the offense, but it is based on two assumptions. First, that the opposing pitcher is on a pitch count. While certain pitchers occasionally break through the upper layers of commonly acceptable pitch counts, very few starters will remain in the game past 110-120 pitches. Although the occasional Randy Johnson (162 pitches in 2000), Jason Schmidt (144 this year), or Mark Redman (147 in 2003) will break this mold, it generally holds true.

The second assumption is that the pitchers in the bullpen are inferior to the starter. This belief has largely been based on how teams have traditionally built their bullpens. As teams were transitioning from the idea that starters should pitch the entire game, they placed those pitchers who were not good enough to start in the pen. The idea of a dominant relief pitcher began to take hold in the 1960s and 1970s and, as such, the quality of relief pitchers improved.

However, most relievers were still washed-out starters. As bullpen size and usage has increased in the past 20 years, more and more players are transitioning to relief pitching earlier in their career. Now it's not uncommon for teams to select college relievers near the top of the draft (Bill Bray and Huston Street being two of the most recent examples). Still, very few players grow up wanting to be a LOOGY (Left-Handed One Out guY, thanks John Sickels), so at some point, the better players are likely still to be channeled towards starting.

It's not working. On average, bullpens have now surpassed the average starting pitcher in terms of quality of pitching. Consider the playoffs this year. If you were an opposing manager, would you rather face: Jarrod Washburn or Francisco Rodriguez? John Lackey or Brendan Donnelly? Tom Gordon or Jon Lieber? John Smoltz or Jaret Wright?

Take a look at how relievers and starters have matched up over the past 10 seasons:

           Starters               Bullpen             Difference
Year   AVG    OBP    SLG     AVG    OBP    SLG     AVG    OBP    SLG
1995  .270   .333   .423    .261   .338   .405   -.009   .005  -.018
1996  .274   .336   .435    .261   .339   .409   -.013   .003  -.026
1997  .268   .331   .425    .264   .339   .408   -.004   .008  -.017
1998  .271   .332   .430    .257   .331   .400   -.014  -.001  -.030
1999  .275   .341   .442    .263   .343   .417   -.012   .002  -.025
2000  .274   .341   .446    .264   .344   .418   -.010   .003  -.028
2001  .270   .331   .437    .252   .326   .405   -.018  -.005  -.032
2002  .266   .328   .427    .252   .328   .397   -.014   .000  -.030
2003  .268   .329   .432    .256   .330   .402   -.012   .001  -.030
2004  .270   .332   .438    .258   .331   .408   -.012  -.001  -.030
Bullpens have had a consistent and sizable advantage over starters in both AVG and SLG for ten years, but their Achilles' heel has been walks, a fault that has been dropping somewhat steadily recently. In 2004, relievers were better than starters in all three categories, not quite as dominant as 2001, but still consistently so.

Translating these to runs is slightly difficult because of the overlap between starters and relievers in nearly every game. Many relievers are employed to strand baserunners, a task at which they are consistently successful at a rate of around 63% (starters strand about 61% of baserunners). However, because of baseball's arcane rules about run allocation, all of the runners bequeathed by one pitcher to another are charged to the former. Not accounting for those differences yields the following ERA and RA numbers:

          Starters        Bullpen        Difference
         RA     ERA      RA     ERA      RA     ERA
1995    4.98    4.53    4.69    4.29   -0.29   -0.24
1996    5.19    4.73    4.81    4.34   -0.38   -0.39
1997    4.87    4.45    4.68    4.24   -0.19   -0.21
1998    4.95    4.55    4.57    4.17   -0.38   -0.37
1999    5.26    4.82    4.90    4.46   -0.36   -0.37
2000    5.28    4.87    5.02    4.56   -0.26   -0.31
2001    4.98    4.57    4.51    4.10   -0.47   -0.47
2002    4.79    4.41    4.40    4.01   -0.39   -0.40
2003    4.90    4.52    4.52    4.14   -0.38   -0.38
2004    5.00    4.62    4.56    4.15   -0.45   -0.47
Looking at these numbers would give the bullpen a decisive and decided advantage over the starters. While it may seem unfair to allocate all the bequeathed runs allowed by the bullpen to the starters, considering the fairly equal strand rates (63% to 61%), the bullpen seems to be doing at least as well as the starters would have had they stayed in the game. Rather than fight about who's responsible for whom, let's look at the numbers if all bequeathed runners disappeared:

          Starters        Bullpen        Difference
         RA     ERA      RA     ERA      RA     ERA
1995    4.61    4.15    4.69    4.29    0.08    0.14
1996    4.78    4.31    4.81    4.34    0.03    0.03
1997    4.52    4.11    4.68    4.24    0.16    0.13
1998    4.59    4.19    4.57    4.17   -0.02   -0.02
1999    4.88    4.44    4.90    4.46    0.02    0.02
2000    4.90    4.49    5.02    4.56    0.12    0.07
2001    4.64    4.23    4.51    4.10   -0.13   -0.13
2002    4.47    4.08    4.40    4.01   -0.07   -0.07
2003    4.56    4.18    4.52    4.14   -0.04   -0.04
2004    4.67    4.28    4.56    4.15   -0.11   -0.13
Over the past four seasons, the tide has turned in the bullpens' favor. Combine these numbers with relievers' continuing superiority in suppressing batting average and slugging percentage and suddenly the idea of working the count to get into the bullpen doesn't make quite so much sense any more.

Finally, look at how relievers and starters have fared in the major rate statistics for pitching:

           Starters               Bullpen             Difference
YEAR    HR/9   BB/9   SO/9    HR/9   BB/9   SO/9     HR/9   BB/9   SO/9
1995    1.03   3.31   6.01    0.99   4.04   7.02    -0.04   0.73   1.01
1996    1.14   3.28   6.13    1.03   4.14   7.25    -0.10   0.86   1.12
1997    1.06   3.27   6.35    0.98   3.93   7.29    -0.08   0.66   0.94
1998    1.09   3.20   6.36    0.96   3.85   7.14    -0.14   0.65   0.78
1999    1.18   3.49   6.14    1.09   4.20   7.16    -0.09   0.71   1.03
2000    1.23   3.57   6.22    1.09   4.25   7.14    -0.15   0.68   0.92
2001    1.18   3.10   6.37    1.05   3.65   7.46    -0.12   0.55   1.09
2002    1.10   3.17   6.19    0.96   3.80   7.20    -0.14   0.64   1.00
2003    1.12   3.09   6.05    1.00   3.71   7.07    -0.13   0.61   1.01
2004    1.18   3.18   6.23    1.04   3.72   7.30    -0.14   0.54   1.07
Bullpens have had an advantage over starters in both home-run rate and strikeout rate in every season for the past 10 years, and it hasn't been particularly close. These advantages have been muted by the dramatic difference in walk rates.

Lately, relievers have reduced their walk rates from an average of 4.09 from 1995-2000 to 3.72 from 2001-2004 while maintaining a healthy strikeout rate (7.17 to 7.26) and home-run rate (1.02 to 1.01). Starters also reduced their walks, dropping from 3.35 BB/9 to 3.13, while holding steady in the other two categories. The improvement in relief pitching, however, has been much more dramatic and has been enough to push an already close race in the bullpens' favor.

There are many possible reasons for a transition like this. A great deal of attention has been focused lately on the long, winding histories of many of baseball's more dominant relievers. In 2002 it was Chris Hammond, out of baseball for several seasons, suddenly emerging with a 0.95 ERA for the Braves. Virtually the entire Anaheim Angels bullpen was constructed from independent-league players and washed-up prospects. Chad Bradford got his own chapter in Moneyball after wallowing in the obscurity in the White Sox system before becoming one of baseball's best pitchers at stranding inherited runners.

Perhaps teams are willing to entertain more options in the bullpen these days rather than recycling the same small set of relievers bouncing around the league as they drifted in and out of style and demand. If teams are truly under pressure to save money while improving, the bullpen would be the single easiest aspect of the team to overhaul on the cheap. As teams have sought out market inefficiencies in the pen, performance has improved.

Of course, this theory assumes that it is easier to find relievers than starters. Since we're using league averages for starters and relievers, though, it's possible that the discrepancy is more a case of the lower-tier starters dragging down their averages more than the lower-tier relievers. The idea that starters are a more scarce commodity around baseball than relievers is prevalent. As long as it remains so, teams will be more likely to stick with a struggling starter than with a struggling reliever because of the idea that the starter is not as easily replaced.

Another possibility is that starters who washed out because of injuries can now return as top-flight firemen rather than being removed from baseball entirely. Pitchers like Jason Isringhausen, who now have access to modern medical techniques but are still hoping to reduce stress on their arms, can return to the bullpen to pitch instead of players who were weeded out simply because of talent.

Finally, there are issues with using simply averages for all pitchers over the course of a season. Primarily, talent-based playing time distribution may be different with starters than with the pen. Despite the occasional attempts at reversion to the four-man rotation, starters for the most part pitch every fifth or sixth day at a maximum whereas top relievers approach 80-90 innings per season, usually in one-inning stints. Depending on how many innings a starter pitches and how little creativity a manager has, it's much easier to distribute a disproportionate amount of the bullpen innings to a smaller group of top relievers than it is to do so with starters. In short, starters are limited by their spot in the rotation as well as their performance whereas the bullpen restrictions are based more on performance and less on frequency of use. The extra restriction can skew the results in the pens' favor.

Whatever the reason, many teams across baseball now feature stronger relief men than starters, and the number of teams that do so is increasing every year. Depending on the quality of the starting pitcher and relievers who may not be available because of recent overuse, working the pitch count can still yield the added benefit of forcing the opposition to use a pitcher of lower quality. However, the frequency of that situation is dwindling with the increase in effective relievers. While it's unlikely that this recent trend signals a significant shift in roster construction or usage patterns for pitchers, it is clear that one of the theoretical advantages of plate discipline isn't so beneficial as it may appear.

Related Content:  Relievers,  Starters,  The Who,  Bullpen,  3-0 Count

0 comments have been left for this article.

No Previous Article
No Next Article

Prospectus Feature: Arbitration Clash
Fantasy Article Tale of the Tape, Dynasty Edition: Jackie Br...
Fantasy Article Fantasy Players to Avoid: Long-Term Outfield...
Fantasy Article Dynasty League Positional Rankings: The Top ...
Premium Article Baseball Therapy: Is Win Probability Broken?
Cold Takes: Smoak on the Roster
Banjo Hitter: PECOTA's Breakout Bets: Hitter...

Premium Article Prospectus Today: Where Have You Gone, Spark...
Under The Knife: The World's Most Famous Ank...
Premium Article Lies, Damned Lies: Using the Golden Run Rati...

2005-01-12 - Premium Article The Lowe-Down
2004-11-09 - Premium Article Time to Get PADE Again
2004-11-03 - Premium Article Is Patience a Virtue?
2004-10-14 - Premium Article Starters vs. Relievers
2004-10-07 - Trending Downward
2004-09-02 - Premium Article The Disappearing Southpaw
2004-08-16 - Premium Article How Parks Affect Baserunning