Swinging With Doc Halladay

WAR ’06-’10
Name

Contact % |

WAR |
Roy Halladay 81.6% 32.7
C.C. Sabathia 76.1% 31.6
Dan Haren 78.5% 26.0
Justin Verlander 80.0% 25.1
Felix Hernandez 78.7% 24.6
Tim Lincecum 75.0% 24.0
Cliff Lee 84.0% 23.5
Johan Santana 76.8% 22.9
Zack Greinke 79.7% 22.2
John Lackey 81.0% 21.5

Chip’s chart:

Name Contact % K/9 ERA FIP xFIP
Cole Hamels 74.9% 8.54 (5) 3.54 (14T) 3.77 (19) 3.51 (7)
Tim Lincecum 74.9% 10.07 (1) 3.04 (4) 2.86 (1) 3.15 (1)
C.C. Sabathia 76.1% 8.00 (16) 3.13 (5) 3.25 (4) 3.49 (6)
Johan Santana 76.8% 8.34 (8) 2.93 (1T) 3.52 (10) 3.61 (11)
Jake Peavy 77.1% 9.18 (2) 3.39 (10) 3.34 (6) 3.56 (9)
Javier Vazquez 77.1% 8.57 (4) 4.24 (38) 3.85 (22) 3.71 (15)
Scott Kazmir 77.3% 8.74 (3) 4.16 (33) 4.20 (34) 4.25 (33)
Chad Billingsley 78.0% 8.14 (12) 3.51 (12) 3.72 (16) 3.98 (27)
Jered Weaver 78.2% 7.82 (19) 3.55 (16) 3.75 (17) 4.12 (29)
Ubaldo Jimenez 78.4% 8.11 (14) 3.54 (14T) 3.59 (13) 3.87 (19)

The main reason for comparing these charts is that when commentators speak of active pitchers who personify the wisdom of “pitching to contact,” they will almost always turn to Cliff Lee and Doc Halladay as examples.  One of Chip’s conclusions is as follows:

“If pitching to contact is such a great concept, then why don’t its benefits show up in the results-oriented statistics?  It’s simple.  The concept is wrong.  It doesn’t actually work. Whereas striking a batter out guarantees an out, allowing the hitter to put the ball in play creates uncertainty around the situation.”

In principle, I agree completely.  However, of the 70 pitchers who have pitched 700+ innings in the last five seasons, Halladay ranks 42nd and Lee ranks 56th in Contact%, much nearer the bottom than the top.  Why have they been able to succeed in the face of the uncertainty created by batted balls?  To answer this question, I’d like to address the vaunted Philadelphia Phillies rotation.  It’s useful here for a few reasons.  Firstly, all five pitchers in it have logged an abundance of innings over the last five years, so it provides us with a decent-sized data set.  More to the point, however, the Phillies cover something close to the full spectrum of contact rates.  I think it’s possible that, as a group, they provide something of a synecdoche for more wide-ranging statistical trends.

Name | Contact % | WAR | ERA | K/BB | GB% |
Cole Hamels 74.9% 18.2 3.53 3.62 41.3
Roy Halladay 81.6% 32.7 2.96 4.86 53.1
Roy Oswalt 81.7% 21.4 3.29 3.37 48.5
Joe Blanton 83.5% 15.0 4.44 2.46 43.6
Cliff Lee 84.0% 23.5 3.61 3.87 39.8
Name | Swing % | Z-Swing% | Z-Contact% | O-Swing% | O-Contact% |
Cole Hamels 49.8% 68.7% 81.8% 28.8% 56.6%
Roy Halladay 49.2% 66.3% 90.2% 30.7% 61.4%
Roy Oswalt 49.2% 66.9% 87.3% 26.9% 64.3%
Joe Blanton 44.8% 63.1% 90.3% 24.6% 64.3%
Cliff Lee 49.5% 66.7% 88.4% 27.4% 70.3%

Z = Pitches in Zone
O= Pitches Outside Zone

The conventional wisdom about Roy Halladay is that “pounding the strike zone” and “pitching to contact” (at least situationally) are his secrets to success, the reason why he consistently maintains low pitch counts (per inning), pitches deep into games, and leads the league in wins, innings, and complete games.  I’ve been watching Halladay religiously for years and I can’t tell you how many times I’ve heard announcer decree, “he’s not afraid to throw strikes” or “you better go up there hacking.”  But according to the data, Doc spends less time in the zone than any of his Philadelphia teammates.  The fact is, over the course of his career, Halladay has been “attacking the zone” less and less.  Since his first full season in the Jays rotation, 2002, the percentage of balls he’s thrown in the strike zone has been progressively descending, to the point where he threw to the zone only 48.6% of the time in 2010, compared to 55% of the time during his first Cy Young season (2003).  You see, Halladay’s secret isn’t throwing strikes, it’s getting hitters to create their own strikes.  As you can see, he has an extremely high O-Swing%, and it’s been even higher in the last three seasons.  Hitters “hack away,” and, as a result, he turns a third of the pitches he throws outside the zone into strikes.

Cliff Lee and Roy Oswalt really do throw strikes.  They rank #1 and #2 in percentage of pitches thrown to the strike zone over the last five seasons, both at greater than 55%.  The distinction I think is most important to observe from this chart, is that it isn’t throwing strikes, as Oswalt and Lee do, that’s most important, nor is it creating contact, as Lee does, nor even avoiding contact, as Hamels does.  The Phillies four Aces have all been extremely productive pitchers over the last five seasons, though their approaches vary rather wildly, because in one category they are remarkably similar.  The top four pitchers in the Phillies rotation all rank in the top six in Swing% over the last five seasons.  Not contact rate.  Not Zone%.  Not strikeout rate or K/BB ratio.  Not GB/FB ratio.  The Phillies Aces don’t pitch to provoke contact, they don’t pitch to avoid contact, nor to get a particular type of contact.  They pitch to provoke attempts at contact.

One would expect that when the batters takes a cut, it’s a bad thing for the opposing team.  That’s how offense is created, right?  But Swing% statistics suggest, quite to the contrary, that the more often the hitter takes the bat off his shoulder, the better it is for his opponents, or at least for the long-term success of the starting pitcher.  Halladay, Lee, Oswalt, and Hamels are all lumped together, provoking the batter to swing close to half the time.  Blanton, on the other hand, the least successful pitcher in the rotation, provokes a swing less than 45% of the time.  As we all know, hitters fail far more often than they succeed.  The more often you induce a batter to action, presumably, the more often you can take advantage of percentages which tilt in your favor.

You can see further circumstantial evidence of this fact when you look at the swing percentages of great hitters.  Swing% for batters is one aspect of that critical skill we often refer to as “plate discipline.”  Of course, it was famously difficult to get a hitter like Barry Bonds to flinch.  Analyzing Swing% doesn’t necessarily have to imply high walk rates or pitches per plate appearance, although I suspect they frequently do correlate.  Regardless of their approach, hitter frequently do better when they swing less.  For Prince Fielder‘s first three seasons, his Swing% was consistently above 47%, in the last two he’s decreased that to around 44%, which means he takes about 100 less swings over the course of the season.  The result, if you assume a correlation, is that he’s had huge upswing in walks and OBP, as well as increasing his Park-Adjusted Batting Runs Created from 29.3 per season to 45.2 per season.

This is, I want to emphasize, a very cursory and selective look at the subject, but one which I think exposes the extent to which over the long haul attempting to wholly avoid contact isn’t any better than actively coveting it.  It is by no means a defense of the “pitch to contact” philosophy.  What I’m suggesting, rather, is that the more practicable version may be “pitching to the possibility of contact,” which can be done in a number of different ways, and doesn’t necessarily have anything to do with throwing strikes, as the pitchers in Philadelphia’s rotation have shown us.

Matt teaches at The University of Alabama. Roll Tide. He specializes in American Literature and Rhetorical Economics. Fate chose for him the peculiar perdition of rooting for the Chicago Cubs and the Los Angeles Clippers.

8 thoughts on “Swinging With Doc Halladay

  1. Mark

    Insightful. The best pitch a pitcher can throw is one that gets weakly grounded int an out right? One pitch out is better than a 3 pitch strikeout when it works, so it would make sense that the best pitchers get swings on balls and pitches to the corners of the strikezone.

  2. Wow. This was so good, I don't even care that you disagreed with me! In all seriousness, great piece.

  3. LarryAtIIATMS

    OK, let's debate. The thinking here is that great pitchers induce a high swing percentage?

    Here are the ten pitchers (out of the 91 pitchers listed when you search 2010 starters on FanGraphs) that induced the highest swing percentages in 2010: Scott Baker, Cliff Lee, Johan Santana, Phil Hughes, Roy Halladay, Ted Lilly, Matt Cain, James Shields, R.A. Dickey and Hiroki Kuroda. This is a good list of pitchers, but by no means great. Only two of these pitchers were in the top 10 last year in fWAR.

    Here is the list of 12 top starting pitchers by fWAR, with their rank in swing percentage in parentheses: Lee (2), Halladay (5), Justin Verlander (31), Josh Johnson (15), Ubaldo Jimenez (77), Felix Hernandez (32), Adam Wainwright (73), Francisco Liriano (23), Jered Weaver (26), John Lester (71), Zack Greinke (49) and Tim Lincecum (37). Again, the correlation between great pitching and high strike percentage seems to be there, but it also seems to be weak.

    Not surprisingly, swing percentage correlates strongly to the percentage of pitches thrown as strikes. 6 of the 10 pitchers on the high swing percentage list also appear on the list of top ten pitchers by strike percentage, and two others just missed: Phil Hughes (12th highest strike percentage in baseball), and James Shields (15th highest strike percentage in baseball).

    So as I see it, pitchers that throw a lot of strikes induce a lot of swings, but inducing a lot of swings correlates weakly with success. Lee and Halladay induce a lot of swings, but King Felix and Lincecum induce just a bit more swings than average, Zack Greinke is about average in inducing swings, and Lester and Jimenez induce fewer swings than average.

    So for the moment, I'm not seeing anything more than a weak correlation between swing percentage and pitching effectiveness. Moreover, a weak correlation is to be expected: good pitchers throw more strikes, inducing more swings.

    So … I think this piece is outstanding, but I'm not buying your argument here.

    • BrienJackson

      I don't have time to do this right now, but as a hypothesis I'm fairly confident about, I think we'll see a stronger correlation with fWAR if we look at O-Swing%.

    • Yea, the major thing is that if you swing…you can put the ball into play. the FIP equation is based on the idea that limiting BIP is beneficial to pitchers. I think the run value of a ball put into play is set at something like .05 runs. I have to agree with you that the swing% thing [probably] has to do with these pitchers throwing a lot of strikes.

      Regardless, it's good to have you on board Hippeaux. Welcome!

  4. Thanks for all the feedback. First of all, I don't see this as a "disagreement" so much as an addendum. As with most advanced stats, Swing% doesn't tell the whole story. As Larry points out, when you look at Swing% for any given season there are always a few notable outliers. For instance, Scott Baker. We can probably all surmise that the reason Scott Baker has a high Swing% is that he's pretty hittable. However, my suspicion is that as we get larger and larger sets of data the correlation between Swing% and success gets a little better. Sadly, FanGraphs only has this info back to 2002, which doesn't even cover a full baseball career, but I do like the results somewhat. Forgive this dirty table:

    Name – Swing % – FIP Rank
    Brad Radke – 52.1% – 45
    Curt Schilling – 51.5% – 2
    Johan Santana – 51.2% – 5
    John Smoltz – 51.2% – 1
    Paul Byrd – 50.7% – 85
    David Wells – 49.9% – 38
    Cole Hamels – 49.7% – 24
    Roy Oswalt – 49.5% – 8
    Randy Johnson – 49.4% – 9
    Roy Halladay – 49.1% – 4
    Cliff Lee – 48.8% – 25
    Ramon Ortiz – 48.8% – 104

    Sorry I'm not going to have a chance to check this list against fWAR for the moment, though I'm very curious about the results. (By the way, 104 pitchers have thrown 900+ innings since 2002, so Ramon Ortiz is dead last in FIP among qualifiers, a real fly in my ointment.) However, as you can see, in the top 12 in Swing% since '02 we've got several Hall of Famers, several borderline cases, all four Phillies Aces, and three innings-eaters. Before we get outside the top third we'll add the likes of Greg Maddux, Pedro Martinez, Javier Vazquez, Jason Schmidt, Jake Peavy, C.C. Sabathia, Mark Buehrle, etc.

    So, what do we do with somebody like Brad Radke? I was surprised to see him atop this list, but reminded thereafter that Radke was a damn fine pitcher, whose career totals are probably better than you remember. He has five seasons with a WAR above 5.0 and eight seasons above 3.0. My hypothesis is that in cases like Radke and Byrd, these pitchers may have succeeded beyond what their stuff warranted, because they induced the batter to action. Obviously, this is a very rough hypothesis.

    Again, thanks for the lively discussion. I'm looking forward to many more.

  5. LarryAtIIATMS

    Hippeaux, again I think this is outstanding work, which I look at as something more like work in progress. If you can show a strong correlation between swing percentage and pitching success, you might then jump to ask if one can be said to cause the other. Does getting a batter to swing more often MAKE a pitcher more effective? Or do hitters swing more often knowing that they're facing an effective pitcher?

    Or is there a third factor at play that is the driving force? Maybe what makes a pitcher effective is throwing strikes and avoiding bases on balls … and that pitchers that throw a lot of strikes also happen to get a higher swing percentage? If that's the case, then the swing percentage may be a tag-along fact … something like noticing that a guy who hits a home run tends to run the bases more slowly than a guy who hits infield singles, but there's nothing about a slow trot to first that leads to home runs.

    I think you also need to consider the significance of the difference you're seeing in swing percentages. Let's say that pitcher A tends to get hitters to swing 10% more often than pitcher B. Is this enough of a percentage to comment on? In a typical game where a pitcher might throw 100 pitches and face 20 batters, that 10% difference amounts to about one extra swing for every two batters. I THINK this should be significant, but maybe not significant enough to explain all you might want to explain.

    I'm not sure where all this will take you, but I look forward to finding out.

    • BrienJackson

      I think maybe we're playing a fool's game here in trying to find some strong correlation to the very best of pitchers. The more I think about it the more I think the common variable is fairly obvious; those guys all have a lot of talent. And when you have that much talent, there's probably a number of different ways you can be effective.

Comments are closed.