A Note on K/BB

Over at Beyond the Box Score, Glenn DuPaul (@Glenn_DuPaul) recently wrote a great article about the flaws inherent to the strikeout-to-walk ratio (K/BB) stat. It’s one you’ve seen in many pitching-related articles here, and one I’ve generally just used unthinkingly lately due to its comfort and ease of use.

DuPaul wrote:

My primary issue with the statistic is the fact that walks are in the denominator. Strikeouts are just as (if not more) important as walks, however when walks are the denominator, a pitcher who is very good at not walking batters, but not a great strikeout pitcher will end up with a high K/BB. It’s simple math.

I prefer to stick with plate appearances as the denominator. Strikeouts minus walks, then divided by plate appearances — (K-BB)/PA — makes for a much more valuable statistic, in my opinion.

For those who feel like that statistic is too much work when K/BB is already calculated right there for use, think of (K-BB)/PA in a different sense.

(K-BB)/PA is the same thing as K% minus BB% (both statistics that are readily available). It’s simply the percentage of batters that a pitcher strikes out minus the percentage of batters that same pitcher walks.

DuPaul then offers a few examples of how K/BB can be a bit misleading, even citing former Phillie Joe Blanton.

Unfortunately, DuPaul’s more technical stat isn’t widely available on FanGraphs and Baseball Reference (yet?) so it won’t be as convenient to cite, but going forward, I will make a concerted effort to use it instead of K/BB. Hopefully the change leads to more accurate analysis going forward.

What follows is a comparison for Phillies pitchers who faced at least 200 batters during the 2012 season.

Strikeout rate (K%), high to low

Player K%
Antonio Bastardo 36.2 %
Jonathan Papelbon 32.4 %
Cole Hamels 24.9 %
Cliff Lee 24.4 %
Joe Blanton 20.5 %
Roy Halladay 20.4 %
AVERAGE 18.7 %
Vance Worley 18.1 %
Kyle Kendrick 17.2 %

Walk rate (BB%), low to high

Player BB%
Joe Blanton 3.2 %
Cliff Lee 3.3 %
Roy Halladay 5.6 %
Cole Hamels 6.0 %
Jonathan Papelbon 6.3 %
Kyle Kendrick 7.3 %
AVERAGE 7.4 %
Vance Worley 8.0 %
Antonio Bastardo 11.6 %

Strikeout-to-walk ratio (K/BB), high to low

Player K/BB
Cliff Lee 7.39
Joe Blanton 6.39
Jonathan Papelbon 5.11
Cole Hamels 4.15
Roy Halladay 3.67
Antonio Bastardo 3.12
AVERAGE 2.51
Kyle Kendrick 2.37
Vance Worley 2.28

(K-BB)/PA, high to low

Player (K-BB)/PA
Jonathan Papelbon 26.1 %
Antonio Bastardo 24.6 %
Cliff Lee 21.1 %
Cole Hamels 18.9 %
Joe Blanton 17.3 %
Roy Halladay 14.9 %
AVERAGE 11.3 %
Vance Worley 10.2 %
Kyle Kendrick 9.9 %

There isn’t a sea change between K/BB and (K-BB)/PA. Perhaps the biggest beneficiary is Antonio Bastardo, who had the third-lowest K/BB of the eight pitchers listed, but moved up to second-best in (K-BB)/PA. Joe Blanton, as mentioned, isn’t quite as attractive in this new light. Compared to the 11.3% league average, though, most qualified Phillies pitchers performed well.

Leave a Reply

*

13 comments

  1. JM

    November 27, 2012 08:59 AM

    I agree with this change…can we something that compares our staff, includign some of our up and coming young bulpenn arms against what is available in FA this offseason?

  2. Bill Baer

    November 27, 2012 10:44 AM

    @ JM

    Here is a list of free agent relievers and their (K-BB)/PA. I removed a few from the list just because I didn’t think they’d get signed (like Brad Lidge and Mariano Rivera).

    Player PA BB K (K-BB)/PA
    Jason Grilli 244 22 90 27.9%
    Sean Burnett 239 12 57 18.8%
    Hisanori Takahashi 212 14 52 17.9%
    Guillermo Mota 91 8 24 17.6%
    Brandon Lyon 258 20 63 16.7%
    Vicente Padilla 218 15 51 16.5%
    Jason Frasor 191 22 53 16.2%
    Michael Gonzalez 151 16 39 15.2%
    Francisco Rodriguez 305 31 72 13.4%
    Matt Lindstrom 200 14 40 13.0%
    Jose Veras 300 40 79 13.0%
    Jon Rauch 233 12 42 12.9%
    Mike Adams 228 17 45 12.3%
    Randy Choate 168 18 38 11.9%
    Jonathan Broxton 238 17 45 11.8%
    Matt Capps 120 4 18 11.7%
    Livan Hernandez 292 16 48 11.0%
    Todd Coffey 83 9 18 10.8%
    Takashi Saito 60 5 11 10.0%
    J.P. Howell 203 22 42 9.9%
    Brett Myers 272 15 41 9.6%
    Mark Lowe 162 13 28 9.3%
    Kyle Farnsworth 120 14 25 9.2%
    Juan Cruz 162 19 33 8.6%
    Chad Durbin 257 28 49 8.2%
    Jeremy Accardo 162 16 29 8.0%
    Jamey Wright 306 30 54 7.8%
    Jose Valverde 294 27 48 7.1%
    Will Ohman 112 5 13 7.1%
    Francisco Cordero 192 18 31 6.8%
    Kevin Gregg 200 24 37 6.5%
    Dana Eveland 95 5 11 6.3%
    Ramon Ramirez 277 35 52 6.1%
    Chad Qualls 231 14 27 5.6%
    LaTroy Hawkins 178 13 23 5.6%
    Jeff Manship 98 7 12 5.1%
    Kyle McClellan 83 9 11 2.4%
    Brad Penny 133 9 10 0.8%
    Shane Loux 112 9 9 0.0%

    And here is a list of Phillies relievers from 2012.

    Player PA BB K (K-BB)/PA
    Raul Valdes 104 3 33 28.8%
    Jonathan Papelbon 284 18 92 26.1%
    Antonio Bastardo 224 26 81 24.6%
    Jose Contreras 56 3 15 21.4%
    Jeremy Horst 125 14 40 20.8%
    Michael Schwimer 147 16 36 13.6%
    Jake Diekman 131 20 35 11.5%
    B.J. Rosenberg 90 12 21 10.0%
    Josh Lindblom 107 17 27 9.3%
    Phillippe Aumont 65 9 14 7.7%
    Joe Savery 108 8 16 7.4%
    Chad Qualls 140 9 19 7.1%
    Justin De Fratus 44 5 8 6.8%
    Kyle Kendrick 63 10 10 0.0%
  3. JM

    November 27, 2012 10:58 AM

    Thanks Bill…with the exception of a guy like Grilli, your point about letting our young arms play really shows up here. Especially our young lefties. I keep saying they are commodities…

  4. pedro3131

    November 27, 2012 11:26 AM

    Food for Thought, Koji Uehara was good for a 30.8% (k-bb)/pa last year, with his worst stat being a 28.7 back in 2010.

  5. Phillie697

    November 27, 2012 06:53 PM

    Bill,

    Here is a question, since Glenn seemed to be on a jihad to challenge conventional wisdom… Is his premise that Ks are just as important, if not more so, than BBs, accurate? The name of the game is getting people on base while consuming as few outs as possible, so K is great in that it’s an automatic out, and BB is not so great in that it’s an automatic base. But is Ks really “just as important, if not more so,” than BBs?

    I think everyone can agree that a hitter with an OBP of .400 is likely a fantastic hitter. If you strike out said hitter, you increase the chance that such hitter generate an out from 60% to 100%, no question. However, if you walk said hitter, you decrease the chance that such hitter generate an out from 60% to 0%. In that little way-too-simplistic scenario, isn’t avoiding BBs more important than Ks?

    Of course, I fail to credit what else Ks prevent, namely extra base hits, and the ability for a base-runner to advance on a BBIP that generates an out anyway. On the other hand, I also understated the importance of BBs by using OBP; using AVG is actually more indicative of the relative values of Ks and BBs. In a game where a .300 AVG is considered a very good outcome for a hitter, how was it determined that Ks are “just as important, if not more so,” than BBs?

  6. LTG

    November 27, 2012 08:04 PM

    So, I should be reading about takings right now, but it is putting me to sleep. Instead, procrastination!

    Phillie, wouldn’t the better approach to answering your question be to figure out the different average expected run differentials between a BB-event and a K-event? Hasn’t this already been done?

    Or we could look at pitchers like Blanton (average K-rate or worse and high K/BB) to see whether their ERAs tend to be higher than their RA estimators (or whatever we call the category housing FIP, xFIP, SIERA, etc.).

    I have no idea how to evaluate the validity of your thought experiment. But I’ll go out on a limb and say it is too abstract. Why not just say that since 3 Ks ends an inning but 4 BBs earns a run, Ks are more valuable?

  7. LTG

    November 27, 2012 08:48 PM

    /Sadface
    www.fangraphs.com/blogs/index.php/carlos-ruiz-suspended/

    But this was produced in the comment section:
    “Who’s to say Ruiz isn’t actually a Soviet-era biological experiment cut lose on the world? The fact of the matter is we just don’t know. Where was Carlos Ruiz in 1988? Was he involved in the Soviet retreat from Afghanistan? Did he, in fact, train Osama Bin Laden? The fact of the matter is it could be so and he just hasn’t gotten caught. We can try to correlate stats with communist mad-science-ery but we don’t really know what those red bastards were doing. But what we do know is that we don’t know what Ruiz’s role in all of this cover up is.”

  8. Dan K.

    November 28, 2012 12:21 AM

    @LTG,

    The suspension is for Adderall. I’m not sure how that helps one perform better in baseball other than, you know, “curing” (masking?) ADHD… Seems stupid to have it as a banned substance, but maybe it makes your reflexes superhuman or something that I’m unaware of.

  9. LTG

    November 28, 2012 09:24 AM

    I know it’s for Adderall. I’m still sad that a) a Phillie knowingly violated MLB’s rules and b) Ruiz will miss 25 games to start the season.

    The suggestions are that it makes you feel better than you would otherwise and so improves performance and that it can be used to decrease your T-cell count, which would allow you to pass your first round of steroid testing and avoid later ones.

  10. Phillie697

    November 28, 2012 11:15 AM

    @LTG,

    Except Ks aren’t the only way to get outs, and there are more than one run at stake when you walk someone since you put a man on base without consuming an out. Talk about being even more abstract :)

  11. LTG

    November 28, 2012 10:10 PM

    Yes, the absurdity of my thought experiment was to highlight the absurdity of yours.

  12. Philllie697

    November 29, 2012 04:39 PM

    Except still no one has answered my question.

  13. LTG

    November 29, 2012 08:48 PM

    Here’s a start:
    www.tangotiger.net/customlwts.html
    Which event is more valuable depends on the run environment, and the breaking point is somewhere between 5 and 6 runs.

    Of course, there is also the simple mathematical fact that equal amounts of change in the numerator and denomenator have unequal effects on the ratio. The pitcher with a 40/21 looks worse than the 39/20 pitcher. So, the ratio already favors pitchers who suppress BBs to pitchers who promote Ks.

    And still we would have to investigate whether the “Blanton correlation” exists. Maybe guys like Blanton tend to underperform their RA-estimators. That would make K/BB a suspect metric for straight comparison.

Next ArticleThe Future is Unwritten: Cody Asche