More pitching appearances are coming with zero days’ rest
Posted by Andy on February 23, 2011
The title says it all. Among relief pitching appearances, an increasingly higher fraction are coming on zero days' rest for the pitcher in question.
Click through for a graph and discussion.
This plot shows the percentage of relief appearances each season coming on zero days' rest, meaning in back-to-back games on consecutive calendar days.
Off the top of my head, I can think of a few reasons why this stat is trending higher and higher:
- There are fewer scheduled off-days now than there were years ago (thanks in part to 162-game schedule) and therefore a somewhat increased number of back-to-back games each season.
- Use of the 9th-inning specialty closer has been on the rise continually over the last 30 years, and this guy is more likely than ever to appear in back-to-back games.
- Use of one-out specialists has also been on the rise for a long time and given this guy's light work load, he too is more likely to appear in back-to-back games.
I did a few other calculations about this. Save opportunities themselves (regardless of whether they were on zero days' rest or not) represented about 19% of all relief appearances in the 1970s and 1980s. In the 1990s, it went down to about 16%. In the 2000s, it was further down to about 13%. Last season it was 12.5%. This indicates the increased prevalence of more relief appearances, in part as starters go shorter and shorter but also as one-out specialists are used more frequently.
Among just the save opportunity relief appearances, about 25% of them came on zero days' rest in the 1970s and 1980s. In the 1990s it was about 29% and in the 2000s it was about 32%. In 2010 it was a bit off its peak at 31.7%.
So that extra data at least somewhat supports the notion that closers and one-out guys are pitching more on zero days' rest, and that these guys are largely responsible for the overall increase.
February 23rd, 2011 at 8:43 am
Also 8th inning specialists are starting to come into vogue (Rafael Soriano, for instance). If the 8th inning hold is yours, you'll go consecutive days as well.
What I really wonder having read this, is what is the most consecutive days an individual pitcher has thrown? And, of course, who is he?
February 23rd, 2011 at 8:43 am
Very interesting trend. I'm wondering what that graph would look like compared with an inverted graph showing percentage of complete games pitched.
February 23rd, 2011 at 8:43 am
Very interesting stuff. Nothing unexpected, but certainly interesting. Just a few thoughts:
1. What happened in '72? Was there anything specific that led to the drop-off that season?
2. It seems to be down a tick over the past three years from what it was throughout the first part of the decade (but still above the 90's and before). I wonder if this is because of the decrease in runs scored leading to more effective relief appearances, and therefore less reason to use relievers so often? Or if it has to do with the increased influence of statistical analysis leading to slightly more sensible usage patterns and a rejection of the LaRussa/Torre style of seemingly cycling through as many relievers as possible?
February 23rd, 2011 at 9:02 am
A couple of other contributing factors could be that Closers only go one inning - This is sort of a related to the specialist and of course the 8th inning guy.
And Perhaps dilution of talent. There are more teams now therefore there are less quality relievers per team - managers may tend to go back to the same guy more often.
February 23rd, 2011 at 9:14 am
"Save opportunities ... represented about 19% of all relief appearances in the 1970s and 1980s. ... Last season it was 12.5%. This indicates the increased prevalence of more relief appearances...."
Andy -- Taking it a step further, I'd say that downward trend in Save Opportunities as a percentage of total relief appearances (SvOp%) reflects an absolute explosion in taking out the pitcher (SP or RP) when the game isn't close.
Save opportunities are not a fixed commodity; thus, making more pitching changes in close games would tend to increase SvOp%, as one reliever hands over the save opp to the next guy.
In order for the trend of increased relief app's to overcome that "save opp multiplier" effect, there would have to be a huge increase in the number of relief apps that were not save opps.
Also, the decline in SvOp% probably reflects increased scoring; as run average increases, the percentage of games that fall into the save zone will decrease.
February 23rd, 2011 at 9:30 am
I'd have to admit that I was expecting the rate to be higher in the Seventies, when people like Marshall, Granger and Hiller were appearing in nearly {in the case of the former, over} 100 games a season.
February 23rd, 2011 at 9:33 am
Along the lines of other comments, starters are getting pulled earlier and more innings are being pitched by relievers each game. With a limited number of relievers on each team (8 or so?) and around 3-4 innings of ground to cover each game, it's more likely some of these relievers will be used more often. Not sure if this has been looked at in the past, but have the number of extra inning games increased as well the last few years with scoring down?
February 23rd, 2011 at 9:57 am
DD @7 -- There is a broad (inverse) correlation between scoring average and percentage of games that go extra innings, and the rate of extra-inning games was higher last year than in most of the years since 1992.
However, there is also a lot of random variation. For example, 2005 saw the lowest scoring average in the 10 years from 2000-09 (4.59 R/G), but it also had the fewest extra-inning games. And while 2010 had the lowest scoring average (4.38 R/G) since 1992, the number of extra-inning games (425 last year) was slightly less than in 2007 (4.80 R/G, 429 EIG), 2004 (4.81 R/G, 428 EIG) and 1996 (5.04 R/G, 428 EIG; note also that there were 6.7% fewer games in '96 due to 2 fewer teams in MLB).
February 23rd, 2011 at 10:02 am
Jim Dunne @3, re: the possibility of "statistical analysis leading to slightly more sensible usage patterns and a rejection of the LaRussa/Torre style...."
I'd love to believe that was true. But if you've managed to detect such a trend, your powers of observation are keener than mine!
February 23rd, 2011 at 10:06 am
Every starting pitcher is on a pitch-count.
February 23rd, 2011 at 10:08 am
Great thread! Another contributing factor albeit small, is a lot less rainouts due to domes and retractable roofs. So no accidental days off.
February 23rd, 2011 at 10:15 am
I'd guess that the slight decline since 2007 in the percentage of relief app's that come on 0 days' rest mainly reflects the fact that total relief app's declined by 3.5% between 2007 and 2010 (from 14,432 in '07 to 13,926 last year), while the average number of relievers carried by a team has probably grown slightly.
February 23rd, 2011 at 10:21 am
BTW, for anyone who hasn't recently read Bill James's study/essay on the optimal usage pattern for relief aces from his Historical Baseball Abstract, you can read it here:
http://www.amazon.com/Bill-James-Historical-Baseball-Abstract/dp/0684806975
(Click the "Look inside" link, then do a "Search inside this book" for the phrase "usage pattern".)
February 23rd, 2011 at 10:32 am
@ 10 Chris
2001 - 2010 = 3 pitchers went > 149 pitches. Edwin Jackson's no-no was the last.
1991 - 2000 = 41 pitchers went > 149 pitches.
1988 - 1990 (3 years) = 48 pitchers had games of > 149 pitches.
February 23rd, 2011 at 11:16 am
Here are the percentages of actual saves (as opposed to save opportunities) coming on 0 days' rest over the past 5 years:
2006 -- 33.8% (405 of 1200 total saves)
2007 -- 33.5% (401 of 1198)
2008 -- 33.1% (392 of 1184)
2009 -- 33.8% (409 of 1201)
2010 -- 33.4% (402 of 1204)
(Why saves rather than save opportunities? For pitchers as a group, used under the modern pattern, the term "save opportunity" is misleading. Many appearances that are technically classified as "save opps" have no real chance of becoming actual saves; e.g., a lefty specialist entering in the 7th inning is never going to finish the game.)
Observations:
-- Whereas relief appearances on 0 days' rest spiked from 2006 to 2007, then fell back in 2008-10, the same 5-year period saw virtually no change in the percentage of saves that came on 0 days' rest.
-- For both 2006-10 and 1970-75, the percentage of saves that came on 0 days' rest was higher than the percentage of relief appearances on 0 days' rest -- and I would guess the same holds true for all the years in between. Here are the percentages of saves on 0 days' rest for 1970-75:
1970 -- 23.8% (209 of 877 total saves)
1971 -- 24.2% (167 of 689)
1972 -- 19.8% (145 of 732)
1973 -- 21.4% (175 of 819)
1974 -- 27.3% (141 of 517)
1975 -- 26.0% (174 of 669)
Two things look interesting in that last data set:
-- The difference in 0-days-Sv% between 1972/73 and 1974/75.
-- The sharp drop in total saves from 1973 (819) to 1974 (517).
February 23rd, 2011 at 11:24 am
I'd call this the "mike Meyers" effect.
In 96-97 the Tigers just burned this guy up- 80 plus appearances two years in a row. Then, when he's almost out of gas and about to hit free agency you just unload him. Payroll wise, it makes sense.
February 23rd, 2011 at 11:48 am
i've always found it weird how relievers seem to flame out quicker than starters these days. you'd think they'd last longer. i don't know if it's because pitching on consecutive days hurts more than more pitches on less frequent days, or if it's because of pressure pitches, or if it's because they're just worse and they can't be starters, or if it's financial, or what.
February 23rd, 2011 at 11:52 am
also, the link reminded me of something. is there a dummy version of bill james' book on the shelves or something? i looked for it at borders and found the one that looked like the one in the link. no real math in it whatsoever. just your typical almanac-style history of baseball book. where can i find the nitty-gritty of what bill james was doing? the models, proofs, and whatnot?
February 23rd, 2011 at 12:07 pm
I see the overall trend upward (given all the reasons Andy listed). But as a whole from 1974 through 1991 (from the dawn of the "fireman" to the beginnings of the "1-inning closer", the number jumped back and forth across the 20% line.
Obviously the "V" created by the 1970 through 1974 data is interesting, but it is equally interesting that 1969 and 1991 have virtually the same % (not too much of an upward trend. The real upward trends are the 1960's and 2000's. I don't really count before 1960 as it was as likely for a starter to be involved in a 0 days rest scenario as it was a reliever (only).
February 23rd, 2011 at 1:14 pm
@ Barkie (still prefer your old handle)
What about Pedro Feliciano for the Mets the last 4 years?
344 games.
78 in 2007
86 in 2008
88 in 2009
92...92!!! in '10.
and wouldn't you know it, free agent.
February 23rd, 2011 at 1:16 pm
and btw...
Pedro had great #s, Mike Meyers was mediocre. Not the comedian or the fictional serial killer, the lefty reliever.
February 23rd, 2011 at 1:23 pm
ohh...
and as far as i can tell, Pedro's last 4 years, pitching in 344 games seems to be a record. I think paul Quantril comes closest, then Tekulve, then Mike Marshal.
I'm not sure how to accurately check that record, i just checked for guys with multiple seasons where they appeared in 75 or more games, and went to their player pages.
But I'm pretty sure 344 is the record over 4 years. That's over two seasons of ball. Impressive.
February 23rd, 2011 at 1:58 pm
Paul Quantrill: 341
Steve Kline: 332
that's the highest I could find. Each faced a lot more batters per season than Pedro did.
February 23rd, 2011 at 2:24 pm
@23 Doug,
Kent Tekulve had a 335 stretch.
Quantril had a 5 year stretch at 409, which pedro just missed in his last 5 years, pitching in 408 games. 409 for 5 years looks like the record. Ray King had a 5 year stretch of 401, and Tekulve would be the master if it were not for the '81 strike.
He had a ten year period of 738, '76-85. In the strike shortened '81 he had only 45, but the average of the 2 years before and after would be 83, giving him nearly 800 games in ten years, with some pretty high inning totals to boot.
February 23rd, 2011 at 4:29 pm
@16, 20, 22, 23, 24
You can find the 4-year record for G by doing a PI search for career years 1-4, then for years 2-5, etc. A little bit of work, but it is complete. It also lets you see how the record changes with the "age" of the pitcher:
1-4: 308 Mitch Williams
2-5: 332 Steve Kline
3-6: 324 Ray King
4-7: 335 Kent Tekulve
5-8: 344 Pedro Feliciano
6-9: 313 Scott Eyre
7-10: 317 Eyre
8-11: 313 Bob Howry
9-12: 323 Paul Quantrill
10-13: 341 Quantrill
11-14: 311 Quantrill
12-15: 301 Mike Timlin
13-16: 297 Timlin
14-17: 292 Russ Springer
15-18: 284 Tom Gordon
16-19: 293 Mike Stanton
17-20: 271 Jesse Orosco
18-21: 230 Dennis Eckersley
19-22: 224 Eckersley
20-23: 217 Eckersley
21-24: 222 Eckersley
So Feliciano does hold the record, with Quantrill close behind. Marshall topped out at 329 in his 4th-7th years. The only others with 320 or more were John Rauch (322, 4th-7th), Tekulve (321, 3rd-6th) and Marshall (320, 5th-8th).
As might be expected the higher numbers were recorded somewhat earlier in their careers, except for Quantrill.
February 23rd, 2011 at 5:06 pm
I suppose you could also do it in PI using age ranges as well. That gives a slightly different perspective:
16-19: 107 Willie McGill, Monte Ward
17-20: 177 Ward
18-21: 226 Amos Rusie
19-22: 249 Rusie
20-23: 236 Rusie, B.-H. Kim, M. Kilroy, B. Mathews
21-24: 308 Williams
22-25: 294 Broxton
23-26: 290 Broxton
24-27: 297 Marmol
25-28: 332 Kline
26-29: 320 Kline
27-30: 324 King
28-31: 329 Marshall
29-32: 321 Tekulve
30-33: 344 Feliciano
31-34: 323 Quantrill
32-35: 341 Quantrill
33-36: 311 Quantrill
34-37: 291 Remlinger
35-38: 294 Tekulve
36-39: 301 Timlin
37-40: 297 Timlin
38-41: 294 Tekulve
39-42: 271 Orosco
40-43: 268 Doug Jones
41-44: 234 Wilhelm
42-45: 233 Wilhelm
So the peak is in the early 30s. Lately there has been a trend towards using younger (early to mid 20s) relievers more, at least at the top end. The teenagers are dominated by pre-1900 starting pitchers. The other thing that stands out to me is that Tekulve shows up three times in a time span longer than a decade.
February 23rd, 2011 at 5:35 pm
#1, I don't know the record for consecutive days pitched, but Wikipedia says the record for consecutive team games with a relief appearance is 13, held by Mike Marshall (no surprise there), Los Angeles Dodgers, 18 June through 3 July 1974, and Dale Mohorcic, Texas Rangers 6 through 20 August 1986.
It's possible that the record for consecutive days pitched is held by some 19th century guy like Will White. In 1879, the Reds played 81 games, and White started - and finished - 75 of them. He also had one relief appearance. Of course, with an 81-game schedule, a team gets a lot of days off. The Reds never played more than 6 days in a row, so White 1879 is not the answer. But someone else from that era might be.
February 23rd, 2011 at 8:04 pm
I always wondered how much the old APBA rule replicated the day to day handling of a bullpen in MLB: pitch three days in a row and your pitcher has to sit a game, and pitch over 2 1/3 in any game and you have to sit the next game. The latter is unlikely to occur as rarely do relievers pitch over 2 innings now. APBA may be considering changes in order to "roll" with the times.
February 23rd, 2011 at 8:53 pm
Pedro Feliciano's 344 games over the past 4 years is nice. (And as a Mets fan, I am grateful, and wish him well in his next stop.)
But in those 344 games, Feliciano pitched a grand total of 239.1 IP, about 0.70 IP per game, with a 124 ERA+ and 1.36 WHIP. And that's while having the platoon advantage more often than an average reliever. (Though he was not strictly a "LOOGY".)
In Mike Marshall's best 4-year stretch (1972-75), he pitched "just" 320 games, but logged 612.2 IP -- 1.91 IP per game -- with a 141 ERA+ and 1.23 WHIP, with no special platoon treatment. (In over 1/3 of those games, Marshall cycled through the entire lineup at least once.)
If you subtract Feliciano's IP and ER from Marshall's, you get 373.1 IP -- more than any reliever has worked over the past 4 years -- and a 2.00 ERA.
I'm just sayin'....
February 24th, 2011 at 12:27 am
[...] More pitching appearances are coming with zero days’ rest » Baseball-Reference Blog » Blog Archi.... GA_googleFillSlot("jmb_468x60"); var addthis_pub="realjohnnybravo"; [...]
February 24th, 2011 at 12:27 am
Mike Marshall also pitched in 13 consecutive games in 1974. Was the record then. I'm not sure if anyone has exceeded it since.
He's an interesting guy, who studied (and got a doctorate) in kinesthetics. He did this to help teach himself how to pitch effectively to both left and right-handed batters.
February 24th, 2011 at 12:37 am
Forgot about off days. During that streak of 13 consecutive games Mike Marshall pitched in, he threw on 8 consecutive days in May 17-24th.
February 24th, 2011 at 10:21 am
[...] More pitching appearances are coming with zero days’ rest (Baseball-Reference). The number was around 10% in 1920, 15% in the ’50s, and 20% or more since the early-’90s. [...]
February 24th, 2011 at 2:05 pm
Makes me think of something I read when Bob Feller died:
He was 15-4 on ONE day's rest.
Of course, the earlier games could have been early exits, and some of the later games could have been in relief...but still...
Makes you wonder if those 4-man rotations were better. Feller, Spahn, etc....lots of guys pitched a lot of years that way.
March 4th, 2011 at 1:46 pm
so what is the record for most consecutive zero-days-rest appearances by a pitcher?