Baseball for the Thinking Fan

Login | Register | Feedback

btf_logo
You are here > Home > Primate Studies > Discussion
Primate Studies
— Where BTF's Members Investigate the Grand Old Game

Monday, March 31, 2003

Beware of Falling Prices Strikeout Rates

Is Bartolo Colon’s declining strikeout rate a bad omen?

Bartolo Colon?s Drop in Strikeout Rate

 

Last year, Bartolo Colon became the pitcher that his fans thought he should be. He won 20 games, posted an ERA below three, and would have been prominent in the Cy Young voting had he not spent half the season in each league. Of course, with the circus in Montreal, Colon immediately became the center of trade talks this winter, and was eventually dealt to the White Sox. Through it all, one thing disturbed me (and some others) about Colon?s wonderful season ? an alarming drop in his strikeout rate.

From the beginning of his career, Colon has been a power pitcher. In his first two years as a regular starter (1998 and 1999), Colon averaged seven strikeouts per nine innings ? above average, but not an exceptional rate in modern baseball. His walk rate in those years was also a little better than average, at 3.4 per 9 IP.

In 2000, Bartolo became a strikeout machine. In 188 innings, he struck out 212 batters ?10.15 per nine innings. His walk rate also shot up, all the way to 4.7 per nine innings. In 2001, both of those numbers fell back down to earth. Colon recorded his second 200-strikeout season that year, but his K rate fell to a still-impressive 8.14 per nine. His walk rate went to 3.64, which was a little worse than average. Bartolo was still clearly a strikeout pitcher, and still seemed on the verge of that breakout season.

Enter 2002. Before the trade from Cleveland to Montreal, Colon went 10-4 with a 2.55 ERA and an impressive walk rate of 2.4 per nine innings ? but only 5.8 strikeouts per nine. After going to the Expos, the walk rate and ERA increased some, but remained good ? but his strikeout rate went down even more, to 5.69. Overall in 2002, he struck out 149 batters in 233 innings, or 5.75 per nine innings. This was well below the MLB average of about 6.5.

Bartolo?s K rate was essentially the same in both halves of the year, and it doesn?t appear that he had some kind of injury that resulted in the strikeout drop. It seems that Colon just decided to stop trying to blow the ball past opposing batters and let his fielders do some of the work. As a result, his walk rate fell, and his baserunners per 9 innings was a career-best. The tactic worked; at age 29, Colon was one of the best pitchers in baseball last year.

Colon?s drop in strikeout rate was dramatic. Did I say dramatic? Actually, it was historic. The chart below is a list of pitchers since 1950 who, from the ages of 28 to 29, experienced a drop in strikeout rate of at least 1.50 strikeouts per 9 innings pitched (minmum 100 IP each year). Colon in 2002 had the fourth-largest decrease in strikeout rate for a 29-year-old pitcher.

Pitcher

Year 1

SO/ 9 IP

RSAA

WS

Year 2

SO/ 9 IP

RSAA

WS

K rate change

Sonny Siebert

1965

9.10

-3

10

1966

6.09

16

15

-3.01

Jim Maloney

1968

7.87

11

14

1969

5.13

37

22

-2.74

Jim Bibby

1973

7.66

-14

4

1974

5.08

5

12

-2.58

Bartolo Colon

2001

8.14

2

12

2002

5.75

28

20

-2.39

Mark Gardner

1990

7.96

-16

6

1991

5.72

3

11

-2.24

Jack Sanford

1957

7.14

20

17

1958

5.13

18

20

-2.01

Bob Turley

1959

6.49

24

20

1960

4.53

13

19

-1.96

Tom Glavine

1994

7.62

19

15

1995

5.75

2

14

-1.87

Mario Soto

1985

7.50

0

10

1986

5.74

-38

8

-1.76

Floyd Bannister

1983

7.99

3

8

1984

6.28

-11

6

-1.71

Gaylord Perry

1967

7.06

11

17

1968

5.35

5

13

-1.71

John Montefusco

1978

6.67

-15

9

1979

4.99

-7

5

-1.68

John Tudor

1982

6.72

17

20

1983

5.06

5

14

-1.66

Steve Trachsel

1999

6.53

19

17

2000

4.93

-8

9

-1.60

Jim Kaat

1967

7.21

5

14

1968

5.62

-9

2

-1.59

Jack Morris

1983

7.11

18

20

1984

5.54

-10

7

-1.57

 

The column “K rate change” in the above table means “change in strikeout rate from year 1 to year 2.” To measure value, I used two metrics ? Lee Sinins? Runs Saved Above Average and Bill James? Win Shares. Colon?s value skyrocketed in 2002, but eleven of the 16 pitchers above saw their value decrease, by one of the two value measures, in the second year of the study. This leaves five pitchers who, like Colon, experienced a significant increase in value along with a large decrease in strikeout rate:

Pitcher

Year 1

SO/ 9 IP

RSAA

WS

Year 2

SO/ 9 IP

RSAA

WS

K rate change

Jim Maloney

1968

7.87

-3

10

1969

5.13

16

15

-2.74

Bartolo Colon

2001

8.14

11

14

2002

5.75

37

22

-2.39

Bob Turley

1959

6.49

-14

4

1960

4.53

5

12

-1.96

Tom Glavine

1994

7.62

2

12

1995

5.75

28

20

-1.87

Steve Trachsel

1999

6.53

-16

6

2000

4.93

3

11

-1.60

 

Each of these guys improved from age 28 to age 29, despite experiencing a large dropoff in strikeout rate. Let?s take a look at each, starting from the bottom:

Steve Trachsel can hardly be compared to Bartolo Colon; Trachsel has never been a strikeout pitcher and his “improvement” was from 8-18, 5.57 to 8-15, 4.80.

Tom Glavine?s strikeout rate in 1994 is an aberration when compared to the rest of his career, and the decline from 7.62 to 5.75 K?s per nine innings is more of a return to normalcy than a dramatic change. His change in value in both Win Shares and RSAA, however, is identical to Colon?s, and Glavine also experienced a significant decrease in his walk rate. Of course, Glavine has remained a very good pitcher since then.

Bob Turley?s ages 26-27 seasons would look better in this study. Like Colon, Turley was a young fireballer who came into his own with a 20-win campaign (in Turley?s case, it was 21 wins) that was accompanied by a large drop in strikeout rate. In 1959, the 28-year-old Turley struggled, but he improved to respectability in 1960. It would be Turley?s last decent year, and he was finished by age 32.

Hard-throwing Jim Maloney had four straight 200-strikeout seasons in the mid-60s, and had a massive dropoff in strikeout rate from ages 28-29. One contributing factor to this decrease was surely the lowering of the pitcher?s mound in 1969. Despite all this, Maloney?s ERA improved from 3.61 in 1968 to 2.77 in 1969. Relative to the league, Maloney?s ERA was 17% worse than the league average in ?68, but 30% better than average in ?69. Maloney severed his Achilles tendon early in the 1970 season, and would not earn a single Win Share for the rest of his career.

The pitcher who may be the best match isn?t even in the study ? Andy Benes. Benes doesn?t make the study because his Year 1 came at age 27 (he?d make the 27-28 list, but this is a 28-29 list). Still, he matches pretty well with Colon. Both were power pitchers. Benes led the league in strikeouts at in 1994, at a rate of almost 10 per 9 IP, and his ERA was 3.86. Colon in 2000 had a strikeout rate of just over 10, and posted a 3.88 ERA. Both players had good strikeout rates the next season (Benes was at 8.47), and both saw their ERAs climb to above four. The next season, Benes? strikeout rate fell to 6.25, a drop of 2.22 strikeouts per 9 innings. Colon?s drop was to 5.75. Both pitchers also had improved control, and Colon had 20 wins and Benes a career-high 18. Benes? strikeout rate was back near 9 in 1997, but it fell off for good after that. He never again won more than 14 games, and only once reached 200 innings pitched. Benes was 34 last year, and appears finished.

I guess the key question I?m trying to raise is this: A pitcher with a low strikeout rate has a much shorter career expectation than a pitcher with a high strikeout rate ? but a good pitcher also has a much longer career expectation than a mediocre pitcher.  In the case of Bartolo Colon, a rare trick has been turned ? simultaneously, a big decrease in strikeout rate and a big increase in value. How do you balance these two factors? Is he better off, in the long run, or worse?  How much of one offsets the other?  Is there a legitimate tradeoff?

I don’t know how to answer those questions, and I doubt anybody does.  And so, like a basketball player unable to grab the rebound himself, I am flipping it up in the air, with the hope that maybe a teammate can pull it in.   How can you even find a record of pitchers who have had simultaneous increases in effectiveness and decreases in K rates? How can you shake the data to make those guys fall out?  And when you get the data, what do you do with it?  

Colon will be a free agent after the 2003 season. If I?m a GM, do I want him on my team? Of course; the guy is a 20-game winner. But how much should I worry about that sudden drop in strikeout rate? And how the heck do you quantify worry?

Matthew Namee is a research assistant to baseball author Bill James and writes a regular column for StealHome.com, which launches on Monday, March 31

 

Matthew Namee Posted: March 31, 2003 at 06:00 AM | 14 comment(s) Login to Bookmark
  Related News:

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.

   1. User unknown in local recipient table (Craig B) Posted: March 31, 2003 at 02:52 AM (#610112)
Thanks Matthew. Awesome analysis. One thing that struck me is that Colon, like these other pitchers, was able to see his K rate drop so much because he had a lot of strikeouts to lose. This is a terrific group of pitchers: every single pitcher on this list was or is a good pitcher.
   2. tangotiger Posted: March 31, 2003 at 02:53 AM (#610119)
Yes, excellent look.
   3. Dave Studenmund Posted: March 31, 2003 at 02:53 AM (#610123)
Very good analysis, Matthew. As a Chicago resident, I've wondered the same thing. Maybe even the infamous Phil Rogers has brought it up. BTW, today's Chicago Tribune focuses almost exclusively on the wonderful young pitching of both the Cubs and White Sox. Interesting take for a local newspaper, and a legitimate attempt to educate local readers.

Sorry for the hijack. Seems to me that sacrificing velocity for control is not necessarily a bad thing. A potential line of inquiry: how many pitchers have made that transition in their careers? Other question: where can the Average Joe find Win Shares for 2002?
   4. Walt Davis Posted: March 31, 2003 at 02:53 AM (#610129)
Nice job. However, rather than looking at the biggest drops from the previous year, it might have been better to look at biggest drops relative to career rate. That probably keeps Glavine out. It also might have been better to look at the percent drop in rate rather than the raw drop.

Anyway, if the QDIPS estimates are accurate, I think we'd have to say it was a good tradeoff. His K-rate dropped, but the drop in his walk rate and a fairly substantial drop in his HR rate as well compensated, leaving him with essentially the same QDIPS ERA. So while maybe he's not a great pitcher, if these changes are permanent he's as good a pitcher while throwing fewer pitches and presumably able to go deeper into games and/or having a healthier career.

One thing to remember is that balls in play are a good thing. The OPS against for a BIP is under 700. A pitcher who could magically keep every ball in play (this would mean no walks or HRs) would be an awesome pitcher.

And the interaction among Ks, BBs, and HRs is not particularly straightforward. If we define OBN as the OBP allowed on balls not in play, SLN as the SLG on balls not in play and OPN as the OPS against on balls not in play, here are Colon's OBN/SLN/OPN for 2001 and 2002:

2001: 370/458/828 (33.7% out-of-play)
2002: 382/473/855 (24.9% out-of-play)

So the 2002 combo of Ks, BBs and HRs produces something close to the same OPN as in 2001. The slightly higher OPN in 2002 is compensated for by having MORE BIP (which produce a lower OPS against).

Of course that doesn't mean that the falling K rate isn't of concern. If it's a sign of losing stuff or an injury, then Colon is in trouble. But it could be that what we need to be concerned with are drops in K/BB rates (largely determinant of OBN) or drops in K rates that aren't accompanied by drops in HR rates (determinant of SLN).
   5. Tracy Posted: March 31, 2003 at 02:53 AM (#610134)
Three comments:

1) As Walt Davis pointed out, using the percentage change in strikeout rate would be better. If Pitcher A goes from 10.0 K/9 to 7.5 K/9, while Pitcher B goes from 8 K/9 to 6K/9, A looks worse in this instance, despite their equal change.

2) Showing how the pitchers performed the season after their big decline might help us in trying to determine how Colon will perform this season.

3) The sixteen pitchers lost an average of one (1) Win Share.
   6. tangotiger Posted: March 31, 2003 at 02:53 AM (#610138)
You should really use PA and not IP in the denominator. As for whether to use differential or division, I'd stick with ratios. So, K/(PA-K) in year 1 divided by K/(PA-K) in year 2.

I wrote a lengthy email to Matt suggesting a possible study that I was thinkning of doing (but I don't have the time/patience at the moment). If someone like Walt, or some other sabermetrician wants to try this as well (and you've got the Lahman DB), send me an e-mail.
   7. Walt Davis Posted: March 31, 2003 at 02:53 AM (#610143)
David,

Are the two mutually exclusive? I know that high-K pitchers have longer careers, but is that because of the high K or because, as time goes on, they gain control, thereby increasing the K/BB ratio. Similarly, drops in K-rates are often a sign of trouble, but is that true when you also see an "equal" drop in walk and HR rates?

Anyway, Colon's future is an open question. The point is, with those rates, he was as effective last year as he has been earlier in his career, and those results are entirely consistent with DIPS. If he can maintain those rates is an open question, but he will be as effective a pitcher as he was before if he does.

As to health, there are a million mysteries and presumably a safe workload for one pitcher is unsafe for another. But I can't think of any good reason to think that saving yourself 5-10 pitches per start would possibly hurt anyone.

Also I know that high-K pitchers have longer careers, but are there studies showing they're in fact healthier careers? I ask because I don't know. But since high-K pitchers are usually more effective pitchers, it stands to reason that they'd have longer careers.
   8. Jay Jaffe Posted: April 01, 2003 at 02:53 AM (#610150)
Walt,

Two answers for you:

First, in general a longer career IS a healthier career. Injuries don't generally draw out players' careers and prevent them from retiring, they do quite the opposite.

Second, and perhaps more enlighteningly... last summer I read a brief snippet in ESPN Mag about Tom House's studies of pitcher mechanics:

"House's company, Bio-Kinetics, has studied computerized analyses of [Mark] Prior since he was a sophomore at University High in San Diego... House has more than 300 big league pitchers cataloged in his digital files, and already Prior, 21, ranks fifth in the efficiency of his mechanics, behind Schilling, Johnson, Clemens and Ryan."

This got me to thinking about some possible explanations for why strikeout pitchers are more successful in the long run, and one of them is simply that efficient mechanics promote longevity. Not all of your high-K guys necessarily have good mechanics, and not all of your good mechanics guys are high-K ones. But the pitchers that maintain their high-K status for an extended period of time do so because they by and large are able to throw the ball hard(er) repeatedly without taxing their bodies as much as the next guy. They keep showing up long after the guys with bad mechanics have gone rehab, or worse. That's worth a lot.
   9. tangotiger Posted: April 01, 2003 at 02:53 AM (#610152)
If i can continue Jay's comment:
mechanics + quality -> longevity
mechanics + quality -> (usually) high-K
mechanics + quality -> (sometimes) low-K

The causative agent to longevity is not high-K, but proper mechanics and quality of pitcher. high-K is a by-product, as is longevity. Since they are both byproducts of the similar causative agents, you expect high correlation between these byproducts.
   10. Dave Studenmund Posted: April 01, 2003 at 02:53 AM (#610153)
So, back to the point, what does this mean for Colon? He had a high K rate, and it declined. Is this a reflection of his underlying pitching mechanics? Have they changed? Or were his mechanics already below par, and his K rate is beginning to reflect that? I will admit that I'm no judge of pitching mechanics.
   11. tangotiger Posted: April 01, 2003 at 02:53 AM (#610155)
For Colon specifically, you need a scout to tell you. Did he consciously change his pitching approach? Or is there something mechanical/physical that changed about him? Is the speed of his fastball, when he wants it the same? Lots of questions to ask.

In studies such as these, you'll get some general answers that might apply to Colon. But, for specific cases, you need much more information than the performance numbers will give you.
   12. Tracy Posted: April 02, 2003 at 02:53 AM (#610171)
At the risk of introducing actual statistical analysis into the discussion...

Performing an ANOVA test on Win Shares shows no evidence that the mean Win Shares are significantly different from year 1 to year 2 (F value of 0.26, p value of 0.61). Similarly, the same test on RSAA comes up negative (F value of 0.87, p value of 0.36). So it appears that the change in strikeout rate has no predictive value for the change in Win Shares or RSAA.

As a result, it seems to me that this article really doesn?t shed any new light or answer the question posed in the subtitle, and that as noted before, seeing how the pitchers performed the season following their decline would be more illuminating.
   13. Matthew Namee Posted: April 02, 2003 at 02:53 AM (#610174)
I want to clarify something... The point of my article was not to guess what Colon will do in 2003, but rather what (if any) effect the decline in K rate will have on his longevity. I don't expect a collapse from Colon soon, but if he really does settle in as a 5.75 K/9 guy, I've got to wonder about his expected career length.

I did look at each comparable pitcher case-by-case, and discussed their future performance, but I didn't focus on "next year" because I really wasn't trying to find out anything about Colon's "next year."

Those DIPS figures that David Smyth posted earlier are interesting, and something I hadn't yet seen. Tango has suggested alternate ways of doing this study, so I can't say we're at a dead end, but the data I found didn't lead me to a conclusion in which I was confident. I do think it's a worthwhile topic, and very relevant (as Colon will be a free agent after this year), but his situation is extremely rare. If anyone else has a better way of looking at this issue, I'd love to see it (and I say that with sincerity and not sarcasm). Thanks for all the support.
   14. Walt Davis Posted: April 07, 2003 at 01:56 AM (#610294)
Defensive? Moi??

But I'd still be interested in an answer to my question. The studies which look at career longevity -- did they look at the impact of K-rates while controlling for K/BB rates? do K/BB rates have an impact independent of K-rates? does the general warning about dropping K-rates hold in cases where the pitcher also drops their walk and HR-rates?

I could well be wrong, but I'd bet that there are relatively few cases like Colon -- guys who saw big drops in K-rates but "equal" drops in BB and HR.

My question about career length and injuries is whether low-K control pitchers get injured or do they lose effectiveness?

And regardless of findings about high K-rates, I can't imagine how throwing 5-10 fewer pitches per game could possibly increase the injury risk. I can easily imagine that a drop in K-rate is a sign that there's an already existing injury that's getting worse, but that's a different kettle of fish.

Anyway, I'll stand by my original statement. He was as effective a pitcher in 2002 as he was in 2001. If he maintains his 2002 rates, he'll be as effective a pitcher as he was before while throwing fewer pitches, which should allow him either to go deeper into games or have a longer/healthier career. How that got turned into a statement about the probability of him maintaining those rates is beyond me.

You must be Registered and Logged In to post comments.

 

 

<< Back to main

BBTF Partner

Support BBTF

donate

Thanks to
BarrysLazyBoy
for his generous support.

Bookmarks

You must be logged in to view your Bookmarks.

Syndicate

Page rendered in 0.4697 seconds
47 querie(s) executed