When Will They Ever Learn?

First of all, I’m not sure why Madden feels it’s necessary to bring up Hughes and Nova in this discussion.  For starters, I don’t know about you, but I haven’t seen many 24 year old power pitchers that suddenly lose 2-3 MPH off of their fastball over a winter—well, unless there’s an undocumented physical ailment causing said loss in velocity.  Still, Hughes’s problem appears to be that he’s allowing far too much contact (92.9% contact rate).  Thus far, he’s allowed 19 hits and struck out only three in 10-1/3 innings.  If anything, pitching to contact is killing his ability to be an effective pitcher.

As for Nova, his entire career up to this point has been based on pitching to contact and inducing ground balls.  His problem during Friday’s start wasn’t that he was trying to avoid contact, but that he walked five batters in 4-1/3 innings.  I suppose you could argue that Nova didn’t have command of his pitches or didn’t trust his stuff (whatever that means), but it’s not like you could call the start “out of the ordinary.”  For whatever reason, he walks guys.  It’s been this way his entire career.  During his five year stint in the minor leagues, Nova walked three batters per nine innings on average.  Last season with the Yankees, Nova posted a 3.6 BB/9 rate.  While five walks in one game might seem a bit high; when you post walk rates like the ones I just mentioned, games like Friday are going to happen from time-to-time.  This is especially true when you’re dealing with a raw 24 year old pitcher.

Secondly (and most importantly), Madden waxes about Garcia’s start as if it were a magical, repeatable event.  It wasn’t.  While Garcia might be able to recapture that magic (or at least a reasonable facsimile) over the next four or five starts; eventually, regression will catch up with him.  It always does.  The problem with Madden’s perception of Garcia’s performance was that the actual result doesn’t jive at all with the expected result.  In fact, the two data sets aren’t even in the same ballpark—no pun intended.

  Actual Hits Allowed xBABIP x Hits Allowed
Ground Ball 11 1 .240 2.64
Fly Ball 8 0 .150 1.20
Line Drive 5 1 .730 3.65
Strikeout 1 0 .000 0.00

 

Given the batted ball data expressed in the chart above, we find that Garcia should have been reasonably expected to give up 7-8 hits (7.49 to be exact) over the six innings he pitched on Saturday—not two.  Why the discrepancy?  Well, for starters, Garcia was exceptionally lucky on batted balls categorized as line drives.  With the line drive xBABIP being .730, we should have expected at least three or four of those line drives to fall in for hits.  Instead, only one fell in.  Furthermore, due to the nature of line drives, there’s a strong possibility (under typical circumstances) those hits could have been of the extra base variety; thereby increasing the run expectancy of those situations.

I’m sure that at least a few of you are thinking, “But he didn’t give up those hits for extra bases, so why does this matter?”  I would agree with you.  They weren’t extra base hits.  Those line drives were converted into outs.  Perhaps we should take his performance at face value, right?

Wrong.

The problem with this line of thinking is that it assumes Garcia (and Garcia alone) was responsible for the outcomes of each batted ball.  He wasn’t.  There were several factors outside of his control that played a major role in his overall performance including defense, defensive positioning, and wind speed/pattern (among others).  Can we assume that each factor in each situation will remain identical in his next start?  Or fifth start?  Or twelfth start?  No, and that’s the problem.  Even if the Yankees toted out the same defensive alignment (Jones, Granderson, and Swisher) in every one of his starts, each player would perform differently not only from game-to-game, but also play-to-play.  Furthermore, defensive alignments can change depending on the handedness of the batter at the plate, base-run situation, and game situation (i.e score).  For example, a line drive hit down the right field line could be converted into an out if the right fielder is playing a left-handed hitter to pull.  By the same token, if the RF is playing the same batter to hit the ball the opposite way, that same line drive would likely fall in for a hit to become a double—or even a triple depending on that batter’s foot speed.  Essentially, defense and luck became primary factors in Garcia’s ability to avoid hits.

This brings me to my next point.

“(Garcia’s effort) just goes to show you don’t have to throw 95 if you locate, change speeds, use the whole strike zone and have command of all your pitches,” Joe Girardi said admiringly.

Said Garcia: “I don’t throw hard anymore so I’ve got to pitch to contact … go out and compete. That’s all I can do.”

While Girardi’s statement about pitchers still being effective despite not having plus velocity is true, most successful pitchers typically have plus secondary pitches that allow them induce whiffs and accumulate strikeouts at a rate near the league average.  Garcia no longer has that ability.  While he still has effective secondary pitches (curve and change), they’re no longer “plus” pitches that can be relied upon to put away batters with regularity.  As a result, I’m not sure Garcia’s ability to pound the strike zone will end up being all that helpful in the long run.  In fact, his inability to coax swinging strikes or induce contact on pitches thrown outside of the zone will be the primary reason he’ll have trouble accumulating long strings of outs over the course of the season.  Pitchers who pound the zone with strikes also have a tendency to leave the ball up and over the center-half of the plate.  What do hitters (even the bad ones) do with pitches left over the center-half of the plate?  They either hit the pitch out of the park, or rope a line drive to the outfield.  We’ve seen this kind of stuff repeatedly with the likes of pitchers like Paul Byrd and John Burkett over the years.  Freddy Garcia isn’t any different.

Obviously, anecdotal evidence is nice, but it only takes you so far.  In order to drive home my point, I decided to look for statistical evidence that showed in most cases (there are always exceptions) pitchers who allow lower rates of contact tended to be more successful than those who allow greater levels of contact.  I ran a query on Fangraphs for pitchers with the highest contact rate between 2006 and 2010, and compared it to their K/9, ERA, FIP, and xFIP.  To limit potential outliers due to small samples, I’ve included only pitchers that pitched at least 700 innings during that time period.  There were 63 pitchers that qualified, and I’ve included their ranking in parenthesis next to their stats.  Here are the ten pitchers with the highest contact rate:

Name Contact % K/9 ERA FIP xFIP
Livan Hernandez 87.7% 4.53 (62) 4.94 (63) 4.79 (60) 4.94 (62)
Joel Pineiro 87.5% 4.82 (56) 4.56 (51) 4.25 (36) 4.07 (28)
Aaron Cook 85.9% 4.00 (63) 4.26 (46) 4.29 (39) 4.27 (35)
Zach Duke 85.8% 4.59 (61) 4.80 (61) 4.45 (47) 4.41 (46T)
Jason Marquis 85.7% 4.79 (58T) 4.86 (62) 4.94 (62) 4.87 (59)
Jeff Suppan 85.7% 4.72 (60) 4.68 (56) 5.05 (63) 4.84 (57)
Joe Saunders 85.7% 5.15 (54) 4.25 (45) 4.57 (51) 4.55 (53)
Paul Maholm 85.7% 5.57 (48) 4.58 (54) 4.30 (41) 4.25 (32T)
Mark Buehrle 85.6% 4.79 (58T) 4.10 (31) 4.36 (43) 4.40 (45)
Jeremy Guthrie 85.3% 5.47 (51) 4.03 (30) 4.69 (57) 4.63 (56)

 

Before we make any conclusions about the above chart, let’s take a look at the flipside of the spectrum.  Using the same parameters I used in the first example, let’s take a look the ten pitchers with the lowest contact rates.

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)

 

After reviewing the two charts, it’s abundantly clear to anyone with eyes and a couple of functioning brain cells that pitchers with low contact rates not only strikeout batters at a higher rate, but also produce lower ERAs, FIPs, and xFIPs than pitchers with higher contact rates.    There are, of course, several other factors that play into pitcher performance (ability to avoid walks, GB/FB ratio, tendency to give up home runs, etc.), but they typically won’t sway a low contact pitcher’s performance from above average to below average and vice-versa.  While there are a few exceptions to the general rule (Buehrle, Guthrie, Kazmir, and Vazquez for example), the correlation holds up pretty well.

So this begs the question:  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.*  As I mentioned earlier, out conversions on balls in play are dependent on the talent of the defense, defensive positioning, defensive routes taken, park factors, game situation, and many others (like the luck dragon) that reside outside of the pitcher’s sphere of control.  Literally, anything can happen on a ball-in-play, so removing that possibility from the equation (via the strikeout) will typically lead to fewer hits; and therefore, fewer runs allowed on average.

*  GB xBABIP – .240; FB xBABIP – .150; LD xBABIP – .730; K xBABIP – .000

There’s a perception that pitchers who allow greater rates of contact have lower, more efficient pitch counts than those who strike batters out.  This is simply not true.   While allowing contact may lead to lower pitch counts, it’s by no means a guarantee.  Why?  Because we’re assuming that (1) contact always occurs early in the count, and (2) balls put into play are being converted for outs.  These two things may or may not be true.   The outcome of those events are largely reliant on external factors. 

Furthermore, the factor that drives pitch count more so than anything is not strikeout rates, but strike/ball counts.  If you have a pitcher who either walks batters at a high rate, or allows hitters to work them into deep counts, allowing high rates of contact won’t keep said pitcher’s pitch count low.  By the same token, pitchers like Cliff Lee, Roy Halladay, and Greg Maddux have historically shown they have/had the ability pitch a complete game, while making only 100-110 pitches and striking out 10+ batters.  Could these three pitchers be exceptions to the rule?  Absolutely, but it drives home an important piont.  All three have/had a tendency to not only get in favorable pitcher’s counts, but also limit walks.   Perhaps the pitch to contact crowd would be better served if they stopped railing against strikeout pitchers (like Francisco Liriano), and instead focus on those who exhibit below average control.

I’m certainly not suggesting that pitchers go to the mound with the intent of striking every batter out and avoiding all contact.  That’s not only unrealistic, but also leads to pitchers overthrowing and making additional unnecessary mistakes over the plate.  Additionally, there are some guys that will never be strikeout pitchers.  Instead, I’m suggesting that pitchers go to the mound without a specific agenda for how they’ll get batters out.   The best way for pitchers to get batters out is by setting them up and attacking/expanding the edges of the strike zone.  For example, if pitcher A sets up batter B, and gets him to swing and miss at a slider sweeping out of the zone, great.  By the same token, if he sets up batter C, and gets him to hit a broken bat grounder on a cutter on the hands, that’s fine too.  The point of pitching is not for pitchers to pitch to contact and trust the defense, but instead keep hitters off balance and use the full range of their repertoire to get hitters out. 

While a pitcher will occasionally have success while posting low strikeout totals; more often than not (provided he produces acceptable walk and home run rates), he’ll have far more success when he’s striking hitters out a league average rate or higher.  The law of averages demands it.

26 thoughts on “When Will They Ever Learn?

  1. In a few months, after Garcia regresses, Madden (that ultimate bandwagon-jumper) will be vilifying the signing.

  2. Madden was also saying this weekend that Bonds should not go into the HOF — even if you believe he did not use PEDs — solely because he was convicted of a felony. Funny, but isn't Maddon among those who has championed GMS III's HOF candidacy? And wasn't GMS III convicted not just for a felony, but for the EXACT SAME felony that Bonds was convicted on? Hmmmmm . . . . .

  3. I think the thing I liked most about Madden's column is the way he makes an actual effort to avoid citing Garcia's 4.67 ERA from 2010, even though it's much easier to cite than "number of games pitcher X allowed fewer than 3 runs per start." I dare say Madden was deliberately attempting to mislead his readers.

  4. This is my first time on the site and I can tell Chip and Brien have never played baseball and do not know much about it. 1. pitching to contact is an effective method, especially with a pitcher like garcia who tops out at 87. The bit about maybe those line drive outs will be hits in starts 5 or 12 is ridiculous. That is why you play the game. You can't just say well next time that might be a hit, who cares. Remember, there are 9 defenders and 1 hitter, probability says the ball will go at someone. Think boys. And the fact that Hughes' contact rate is up is because he lost mph's on his fastball. It's not as tough to hit when its 90 instead of 96 hence guys are making contact and he is not blowing the ball past anyone. If randy johnson threw 90 i bet his contact rate would be up as well.
    To base this whole article on the fact that you can't understand or agree with "pitching to contact" is kind of ridiculous fellas. Think about it.
    Shaken not stirred,
    Tom

  5. On average, who throws more pitches in a start: Those who pitch to contact, or those who are strikeout artists? Can you run numbers for me, please? I'm interested in efficiency between those who get the outs in the field versus those who get the strikeouts. Maybe you could use the two sets of pitchers that are shown in your data in the article. Thanks.

  6. The only guy I can remember who "pitched to contact" and was successful was Wang – back when he had his first couple of monster seasons. Before he was injured, before the league batters figured him out. But he was a sinkerball pitcher, who depended as much groundball outs as Mo depends on his cutter.

    In his case, pitching to contact worked — for a while. For more normal guys, contact = at best line drives, but quite possibly, home runs. So – if we're talking Wang of 4 years ago, sure – pitch to contact. But Garcia – I sure hope he misses a few bats too.

  7. Dave Duncan has been successful in St. Louis with some pitchers like Suppan, Penny, Marquis, Pineiro by having them increase their GB% and reduce their BB/9, even though their K/9 showed no significant increase. They, and other GB pitchers such as Wang, Lowe, etc., do so using a heavy two-seam low in the zone, which by its trajectory, often has the hitter hit the ball on the ground. To me, that seems like they are trying to make "their pitch" (as all pitchers are), while at the same time trying to place the ball in a place where they are minimizing their BB/9 and HR/9 and increasing their GB%. I think some of the argument is syntactical, but I would consider that "pitching to a place in the zone where GBs and not LDs or HRs are encouraged"). It seems to be that's what people mean when they say "pitch to contact."

  8. In no way shape or form am I saying that allowing a high LD% (like Garcia) is a recipe for sustainable success. It's not. Since your defense (which may not be good) actually has to field the baseball, I'm also not saying that giving up a lot of groundballs is better than striking a bunch of dudes out. It's not. But to echo what Dave Cameron (and many others) wrote recently about Kyle Lohse in St. Louis (article on ESPN.com), increasing your GB% and reducing your BB/9 and HR/9 are in fact sustainable ways to be successful without missing as many bats.

  9. Very interesting read, though I do think that some pitchers have the ability to pitch to poor contact (see Mo). The interesting thing about Hughes is that he seems to have no problem getting 2 strikes, the problem seems to be having a pitch or pitches that look like strikes, but end up just outside the zone, where good contact is more difficult. Hopefully, Garcia will show the ability to entice batters to hit ball just out of the strike zone, resulting in poor contact. I also would suggest that there are different types of line drives, ground balls and fly balls as far a how hard they are hit- I'm not sure the raw data really accounts for that.

  10. Late to this, and I haven't been to this site often, nor did I play organized baseball above the high school level. :)

    My opinion on where "pitching to contact" is effective is only in one single area – avoiding walks. If you are pitching to contact, you are around the strike zone and you aren't going to be walking as many folks. You're going to give up balls in play (and trust your defense, but you're going to give up hits too), but the only people you walk are going to be folks you are pitching around. Being at work, I don't have time to look them up, but I would be very curious what the BB/9 is on those pitchers in the top chart, compared to the BB/9 on the lower chart.