It Doesn’t Matter Who Hits Behind Ryan Howard

There has been a lot of discussion this year in Philadelphia about where Ryan Howard ought to hit in the batting order. Last season, the NL MVP batted fourth behind Chase Utley and in front of Pat Burrell. Many fans, sports writers, and even manager Charlie Manuel (whom I remember fondly leading the Triple-A Charlotte Knights) have pondered changing the line-up in order to produce more offense. The most popular idea was to flip-flop Utley and Howard in the order—having Howard bat third and Utley fourth—to allow Utley to “protect” Howard. I also wonder why Philly fans hate Pat Burrell so much. The Braves would be happy to take him off of your hands.

The popular concept of protection is that a good on-deck hitter affects the way the pitcher throws to the batter. The traditional argument says that pitchers cannot “pitch around” batters, because if he walks the batter he’ll have to face a good batter with a man on. But, the argument isn’t done yet. This allows a protected batter to see better pitches to hit—the pitcher can’t nibble around the edge of the strike zone for fear of inducing a walk—and thus increase his ability to hit for power.

However, I believe one crucial element has been left out of the analysis. Pitchers can alter more than just pitching inside and outside of the zone. Pitchers also regulate their effort from one batter to the next, conserving energy to put on a fastball or hiding a wicked out-pitch for just the right occasion. In a situation where a good on-deck hitter follows the batter, the pitcher has likely saved up his best just for this moment. Even if the pitcher is less likely to walk the batter, the batter isn’t necessarily going to be more productive at the plate.

This is something that Doug Drinen and I addressed in a study on protection (here is an old post on the subject), and I use this study to examine the issue in Chapter 2 of The Baseball Economist (“The Legendary Power of the On-Deck Hitter”). Using play-by-play data and a host of empirical techniques to control for many factors we found that a better on-deck does lower probability that the batter walks, which is the first step in demonstrating protection. However, this does not translate into more productive at-bats. As the on-deck batter improves, the probability that the batter gets hit, extra-base, and hits a home run declines—the exact opposite of the protection hypothesis. But, the magnitude of the effect is so small that it’s best to say that there is no such thing as protection.

Now, when Doug and I wrote this study, we were very proud of the granulated data we used. After all, how can you get more specific than play-by-play? Well, how about pitch-by-pitch data? Steven Levitt at Freakonomics points to a new study by Ken Kovash that breaks the game down even more than we did. He finds that good on-deck hitters induce pitchers to throw more strikes and more fastballs—fastballs are easier to throw for strikes than breaking balls and change-ups. This fits with our finding that the quality if the on-deck hitter impacts walks. However, I’m curious as to how this effect translates into overall success of the batter. One thing that Ken does not report—I’m hoping that he will report it when his schedule clears up—is how the on-deck batter affects the speed of the pitch. Doug and I could only hypothesize as to why the on-deck hitter does not lower the output of the current batter. If the answer is more pitcher effort, we ought to see pitchers putting a little more mustard on their pitches with a good batter on deck.

The moral of the story is this: a good on-deck hitter lowers the likelihood of a batter walking, but it does not translate into more offensive production.

12 Responses “It Doesn’t Matter Who Hits Behind Ryan Howard”

  1. Andrew says:

    I like your comment over on freakonomics. The pitch speed of the fastballs would be great to know because we know that starters bear down on some hitters and lighten up on others.

    And I too would be happy if Pat Burrell were “protecting” Brian McCann in Atlanta.

  2. Mark says:

    Enjoying your book very much.

    The knock against Burrell in Philly is not just about protecting Howard. The biggest complaint is that he strikes out too much with runners on base especially late in games or in close games, i.e., he is accused of being that thing Philly hates the most and seems to have in abundance, a choker.

    Has anyone come up with a stat to show who really is “clutch” and who isn’t?

  3. Pizza Cutter says:

    J.C., I’m treating myself to your book for Easter. Now, I’m really looking forward to it.

  4. Brant says:

    Since he’s been up, Howard has had 318 AB’s in the #6 spot. I know that’s a small sample size, but I bring it up because he probably didn’t have all that much “protection” from the #7 hitter during those AB’s, but still posted a line of .292/.361/.563. Howard’s a good hitter no matter where he is, and to tweak things after he’s had just 37 plate appearances in 2007 seems like a bit of a knee-jerk reaction to me (that, as JC pointed out, would be irrelevant anyway).

    By the way, I’d love to have Burrell in Atlanta too.

  5. Jeffrey Beaumont says:

    I’m happily ready to agree with your points on protection, but what I don’t understand is the more larger issue that it seems to make more sense for your better hitters to be generally higher up in the order so that they can get more at-bats. On this point I am in particular confused with the decision of Pittsburgh to have Jason Bay, (to me) inarguably their best hitter, hit fifth in the lineup. How can in any context it make sense to have out-machine Jack Wilson hitting second and Bay fifth (behind Freddy Sanchez and Adam LaRoche)? Not because of anyone being more naturally suited to hitting second or fifth (or third or fourth for that matter), but just because i’d imagine that Batter #2 will pick up an extra 30-50 ABs over a year than Batter #5. To me this seems like common sense, but are there any studies to corroborate this logic?

  6. Isabel says:

    Phillies fans don’t like Burrell because we don’t like anybody. (Also, see Mark’s comment above, which is probably a more accurate answer.) But remember, this is the city that booed Santa Claus.

  7. John W says:

    Jeffrey – Try a Google search using the terms “batting order” or something along those lines…I’ve seen those studies before and your hunch is true. The best hitters should hit toward the top of the lineup.

    Mark – Try Fangraphs for stats on Clutch. I think The Hardball Times has a few things as well.

    JC – I saw the Freakonomics blog post but didn’t read the comments, so I’d been waiting for your take on this. It’s good to see you make the distinction between batters getting more strikes and overall offensive production, since it seems like people are coming to the conclusion that your studies have opposite findings.

  8. Brant says:

    Jeffrey/John…I agree. Looking at 2006, the average total PA for hitters #1-9 (per team) across the Majors was: 767, 749, 732, 714, 697, 680, 661, 643, 625, respectively. To continue with the Howard example, he averages a home run every 13.6 PAs, so you can guess at the difference it would make using this limited data. Perfomance is the same, but more PAs are always nice for a good hitter.

  9. Ron says:

    I’d take Burrell as an upgrade in left field, but don’t for a minute delude yourself he would provide protection for McCann. Bobby won’t be moving Francoeur out of the 6th spot in the order.

  10. tangotiger says:

    Each slot gets 18 PA (162/9) more than the next. The 3-5 spots get more runners on base than the other slots. Typically, your best hitters should hit 2 or 4, but that’s not set in stone. David Ortiz is the best clutch hitter of 05/06, as you can guess. Check Fangraphs.com. Also check my site for our take of pitching around.

  11. Kenny says:

    LaRussa’s strategy of hitting Duncan ahead of Pujols would seem to fit your finding well. Duncan hits fastballs well, in front of Pujols he will see more fastballs therefore his average should be higher in front of a good hitter than in front of lesser hitters. Maybe the issue is not so much where a Pujols or Howard hit in the lineup as it is whether there is a fastball hitter ahead of them.

  12. Jonathan says:

    I know I’m a little late to this party – but I was wondering if the study was controlled for the skill level of the pitcher. Particularly the fact that in a tight game a good manager will often use his best reliever(s) against the best hitters, in a bid to preserve a lead/keep a deficit small.
    I really haven’t thought about this all the way through, but if it could be demonstrated that they (the good hitters) face better quality pitchers more often than weaker hitters, then that in itself may be a good reason for the lower walk rates.
    I don’t know if I spelled that out as well as I’m seeing it in my head, but just wanted to mention it as something to think about…