Baseball Beat January 23, 2009
2008 Leaders and Laggards: K/100 Pitches

Three years ago, I introduced the concept of evaluating pitchers based on strikeouts per pitch, which has a higher correlation to runs allowed than strikeouts per batter faced or strikeouts per inning.

In a follow-up article in November 2007, I summarized the value of this new measurement:

All of us like pitchers who can rack up strikeouts. There is no argument between statheads and the scouting community over the value of missing bats. In a nutshell, Ks are the out of choice. The more, the merrier.

We also know that pitch counts are important. The fewer, the better. As such, it seems logical that combining high strikeout and low pitch totals is a recipe for success . . . The best way to measure such effectiveness is via K/100 pitches. The formula is (strikeouts divided by total pitches) x 100.

Interestingly, the average starter's workload has been roughly 100 pitches per start for the past several years. As such, K/100P gives us additional insight as to the approximate number of strikeouts per start.

The argument against K/100P is that it has vestiges of BB and BABIP mixed into the formula and, therefore, is not a pure stat. If that is the case, then the same holds true for K/BF, as compared to K/IP or K/9. The latter is nothing more than strikeouts per out. A pitcher who gives up a lot of walks and hits is going to face more batters and increase the likelihood of striking out more hitters per inning pitched. Therefore, strikeouts per batter faced tells us more than strikeouts per inning.

Similarly, a pitcher who throws a lot of pitches is going to increase his chances of striking out more hitters per batter faced. Accordingly, strikeouts per pitch improves upon strikeouts per batter faced.

Not surprisingly, K/P has the highest correlation to ERA and RA. K/BF has the second-highest correlation and K/IP has the lowest correlation. In any other words, K/P > K/BF > K/IP.

Last season, there were 142 pitchers who threw 100 or more innings. The correlation between K/100P and ERA among these pitchers was meaningful at -.576.

The distribution of K/100P was as follows:

```Best                    7.37
Top 10%                 5.80
Top Quartile            5.24
Average                 4.54
Median                  4.51
Bottom Quartile         3.58
Bottom 10%              3.13
Worst                   2.39
```

Rich Harden was No. 1 with 7.37 strikeouts per 100 pitches. Livan Hernandez ranked dead last at 2.39 K/100P.

Let's take a look at the top and bottom 30 pitchers in terms of K/100P:

TOP 30

```Rich Harden             7.37
Tim Lincecum            7.20
Joba Chamberlain        6.90
CC Sabathia             6.58
Josh Beckett            6.40
A.J. Burnett            6.33
Ervin Santana           6.24
Dan Haren               6.17
Edinson Volquez         6.08
Scott Kazmir            6.04
Randy Johnson           5.97
Javier Vazquez          5.92
Jake Peavy              5.80
Wandy Rodriguez         5.76
Ricky Nolasco           5.74
Johan Santana           5.73
Cole Hamels             5.72
Ted Lilly               5.68
Jorge de la Rosa        5.67
Zack Greinke            5.67
Ryan Dempster           5.60
Jonathan Sanchez        5.55
Felix Hernandez         5.47
Brandon Webb            5.45
Brett Myers             5.40
Clayton Kershaw         5.38
John Lackey             5.36
Oliver Perez            5.35
```

When healthy, Harden ranks among the best pitchers in the game. The 27-year-old righthander dominated American League hitters when he pitched for Oakland and National League batters after he was traded to the Chicago Cubs in early July. On a combined basis, Harden was 10-2 with a 2.07 ERA. He struck out more than 30 percent of all hitters and 11 per nine innings. The good news for Cubs fans is that Harden has decided not to pitch in the World Baseball Classic.

Tim Lincecum led the majors in K/9 (10.51) among those with 162 or more innings and strikeouts (265). The National League Cy Young Award winner fanned 28.6 percent of the batters faced en route to an 18-5 record and a 2.62 ERA.

Joba Chamberlain only started 12 games while making 30 appearances out of the bullpen. He threw a total of 100.1 innings, striking out 118 (10.58 K/9). The hard-throwing righthander whiffed 27.0 percent of batters as a starter and 30.8 percent as a reliever. He began and ended the season in the latter role and is expected to serve as a set-up man starter for the Yankees in 2009.

Edinson Volquez is a good example of K/100P as a metric of effectiveness. He ranked second in the majors in K/9 (9.46) but only seventh among qualifiers in K/100P. The other side of the Josh Hamilton trade struck out 206 batters in 196 innings but walked 93. He placed eighth in most pitches per plate appearance (4.04) in the majors, which negatively affected his K/100P ranking.

BOTTOM 30

```Kevin Correia           3.49
Luke Hochevar           3.47
Miguel Batista          3.46
Joe Saunders            3.43
Jason Marquis           3.42
Darrell Rasner          3.42
Joe Blanton             3.42
Nick Blackburn          3.33
Mike Pelfrey            3.31
Jarrod Washburn         3.30
Paul Byrd               3.24
Jeff Suppan             3.18
Jeremy Sowers           3.16
Daniel Cabrera          3.15
Glen Perkins            3.13
Aaron Cook              3.13
Zach Miner              3.06
R.A. Dickey             3.03
Scott Feldman           2.98
Zach Duke               2.97
Brian Burres            2.90
Kenny Rogers            2.88
Fausto Carmona          2.86
Jon Garland             2.82
Carlos Silva            2.81
Sidney Ponson           2.68
Kyle Kendrick           2.60
Livan Hernandez         2.39
```

Joe Saunders was the most successful of those pitchers listed in the bottom 30. The Angels lefthander ranked sixth in wins (17) and W-L % (.708) and seventh in ERA (3.41) and ERA+ (130) in the AL. Rather than being named to the All-Star team for a second consecutive season in 2009, look for Saunders to regress based on his low K/100P rate and BABIP (.269). I would agree with the projection systems that call for him to win 11-13 games with an ERA of 3.90-4.25.

Fausto Carmona, on the other hand, was the most disappointing pitcher last season. In 2007, the Indians righthander was second in wins (19), ERA (3.06), and ERA+ (151) in the AL while finishing fourth in the Cy Young Award voting. However, he ranked below the average and median in K/100P (4.37) that season, perhaps portending a difficult repeat performance in 2008 when he went 8-7 with a 5.44 ERA.

Although Luke Hochevar is only 25, it doesn't appear as if he is going to make good on being the No. 1 selection in the 2006 draft. Kudos to the Dodgers in holding the line on his bonus demands when taken with the 40th pick in the 2005 draft.

Pay attention to K/100P. This metric will add more value than K/BF and K/9.

Update: 2008 K/100P Rankings (100 or more IP).

Very interesting Rich. Am I missing a link? Can we see the entire list of eligible pitchers and not just the top and bottom 30?

"[Joba Chamberlain] began and ended the season in the latter role and is expected to serve as a set-up man for the Yankees in 2009."

No. The Yankees plan to use Joba as their fifth starter. If he stays healthy, he is expected to throw somewhere around 140 innings.

I stand corrected. Thanks.

Thanks for this, Rich. As a Mets fan, I liked seeing Oliver Perez sneak into the top 30; I really hope they re-sign the guy. And I also can't be surprised that Mike Pelfrey is in the bottom 30. I would seriously investigate trading him for an outfield bat or credible 2d baseman while his value is high, rather than wait for his results to regress.

When Carmona is right he is an extreme ground ball pitcher. Noting the high number of Indian's pitchers in the low group, you should note the system wide philosophy of pitching to contact with sinkers. This keeps pitch counts low and gives Peralta a much higher SS range factor than is in any way warranted by his play.

Interesting article. I can't help but notice the top 30 in K/100IP also looks to be the top 30 "great when healthy, which is not often" list.

This looks to be a great way to look at pitchers from year to year. Looking at the top 60 in K/100 pitchers for 2007 and 2008 (and filtering for a minimum of 100ip) and you see a lot of repeat performers.

37 of the top 60 appear in both years. You can disqualify 30 of the 120 performers because they didn't reach the 100ip threshold for various reasons in both years. (Rookies, injuries, etc.)

I'm not as well informed on statistics as many of you on this site, so please correct me if this is not a helpful statement.

If you take the 30 pitchers that didn't meet the 100ip threshold both years from the 120 top performers in both years, that means that 74/90 available performers repeated. That is a remarkable rate.

I now see why guys like Shaun Marcum and John Danks came out of nowhere (at least in my mind) to have a great 2008. You can also see how Ervin Santana really underperformed in 2007 and was primed for a return to usefulness in 2008.

It also makes me think that a guy like Carlos Villanueva is a huge sleeper for 2009. I haven't heard anything about him all off season, but he is on both lists, and I would anticipate his HR/9 to drop from the inordinately high 1.5/9.

Anyway, great work on looking at this metric. I think it will prove to be very helpful in watching for breakout performances.

Interesting stuff, Rich. Is there a place that this stat is found throughout the season?

http://statisticianmagician.mlblogs.com/

Wow, great stuff!

Looking over the list, I couldn't help but see that the Giants have three starters on the Top 30 list: Lincecum, Johnson, and Sanchez. In addition, Matt Cain just missed the list with a 5.16 K/100, according to my calculation.

However, now the Yankees have three of the top 6 in Joba, CC, and AJ. Cubs also have three of them too. Phillies, Dodgers, D-backs, and Angels got 2 as well. And I think the Royals too (Greinke for sure; De La Rosa?)

How does the correlation of K/100P to ERA & RA compare to the correlation of the latter 2 to WHIP?

Gues I should have just researched this myself - took all of a minute.

The correlation of WHIP and ERA is outstanding - .9. But the repeatability of WHIP is far lower that that of K rates. So using it as a predictive tool is much less reliable.

But many of you probably knew that.

And that .9 correlation with WHIP is from a case, not a large data set. It's likely lower than that overall.

Rich, with all due respect to your efforts on behalf of Bert Blyleven, your work on K/100P and your categorization of pitchers by K and groundall rates is what I find most valuable. I've been looking forward to this article since Christmas. Thanks for continuing to provide this outstanding information!

as always I love the Lederer k/100 column.

It came just in time, too, as I wasnt sure who to take next in a fantasy draft and this little post was enough to give Nolasco the job.

Some year soon people might start to notice Chad Billingsley in the top 10. No other Dodger from last year made either the top 30 or bottom 30.

Sorry - Clayton Kershaw just makes it into the top 30 as well. In the meantime Lowe and Penny have both moved on at very high salaries without making it in here.

luke hochevar should record a few more K's to be effective, but he has a developing sinking fastball that may make his lack of K's be a damning quality.

don't be surprised to see luke become a very steady middle of the rotation pitcher for many years.

Excellent article. I'm also wondering what the correlation is of batters faced per inning pitched and batters faced per runs allowed in terms of ERA and RA/9 (even though, given the fact pitchers rarely last past 6 innings RA/6 may be better). Batters faced per run allowed may be a more telling stat than WHIP since WHIP doesn't tell one the sequencing of hits. For instance, a walk, followed by a single, followed by a HR is far more damaging than a HR followed by a single followed bay a walk. I did a study of it in 2006 and found a strong correlation between Batters Faced per Run Allowed and Batters Faced per Inning Pitched as well as Base Runners per Inning Pitched and Runs Allowed per 9.

I did something after the 2007 season (I haven't plugged in 2008 stats yet) and found a very strong correlation between baserunners allowed per game and the number of runs scored per game. I found the correlatio to be the nearly identical for both the AL and the NL. I know this seems like a "duh" stat but I think the bottom line is look at pitchers who are consistently successful at keeping the other team from having men on base. This greatly reduces the chances of even an error being costly, as the 2006 World Series best showcases. The Tigers' errors helped cost them Games 4 and 5 because the Cardinals kept having men on base. I'm wondering how this corresponds to K/100P?

I was just reminded that Joe Saunders pitched with kidney stones for at least part of 2008, so maybe that KCP* goes up in '09.

(*C... using the Roman numeral for 100)

I'm not down on Saunders as much as I think it will be nearly impossible for him to repeat his excellent season in 2008. Even if his strikeout rate improves, I would still expect his ERA to go up due to an unsustainably low BABIP last year. I see Saunders as a solid pitcher with a slightly above-average fastball and change who throws strikes and does a pretty good job at keeping the ball on the ground. One can eat a lot of innings and make tens of millions of dollars doing just that for a career.

Great Article, although I do have one question. From the article:

"A pitcher who gives up a lot of walks and hits is going to face more batters and increase the likelihood of striking out more hitters per inning pitched."

Could you (or someone else) explain why this would be? It seems to me that if you have 3 outs per inning, someone getting on base wouldn't really give you more opportunity to accumulate strikeouts. You still only get 3 outs. Why is it that by facing more batters in an inning you are increasing the likelyhood that those outs will be obtained through strikeouts.

In fact, it seems to be that the more hits and walks you give up, the less strikeouts you would have because some of those outs would then be made through double plays.

I assure you that I'm completely prepared for the inevitable "I'm such an idiot" feeling when this is explained. I'm just not smart enough to figure it out myself.

Thanks.

Hi Rich, you forgot to include the guy who was 3rd in baseball, with 158 Ks in 2240 pitches (198 IP):

Ben Sheets 7.05

Oops, I take it back. 2240 pitches was from 2007; in 2008, Sheets threw 3051 pitches, for a K/100 of 5.18. Sorry!

Hmmm...I'm with Kramer on this one. That statement sounds quite odd. In fact, I could see the opposite being true. As a pitcher faces more batters in an inning, he's bound to get tired, which would decrease the number of strikeouts per inning.

In response to the skepticism over the statement I made (quoted four comments above), the more batters a pitcher faces, the more likely he is going to rack up additional strikeouts, walks, singles, doubles, triples, and home runs, at least as measured per inning pitched. His K/BF will be negatively affected but his K/IP has the potential of being positively affected because he will have more chances to face opposing pitchers, bad hitters, and those who strike out more often than the average. Ergo, it's more likely that the outs recorded will be via strikeouts.

There is a difference between striking out the side and striking out the side in order. The latter is a much more difficult achievement, yet both result in the same K/IP rate. As I stated in the article, K/IP is "nothing more than strikeouts per out."

A wild pitcher like Edinson Volquez is going to look better using K/IP than K/BF or K/P. Why? Because he walks a lot of batters and, therefore, faces more batters and gets more opportunities to strike out batters, again as measured per inning pitched. In other words, he is not as proficient at striking out hitters as another pitcher who gets the same number of Ks per inning but faces fewer batters. Another way to think about it is if a pitcher like Volquez chose to challenge hitters more often when working behind in the count, he would most likely produce more batted ball outs and fewer strikeouts.

I'm new to the website. Great stuff. I may be missing it, but was the link to the spreadsheet ever posted?

Yes, Todd. At the bottom of the article, click on "2008 K/100P Rankings" next to the word "Update." The spreadsheet will give you the K/100P rankings of 142 SP with 100 or more innings.

Thanks!