Pitcher performance after batting success

It began in the bottom of the third inning of Game One of this year’s National League Championship Series. St. Louis Cardinals pitcher Joe Kelly, behind 2-0 after a rocky top of the frame, came to bat with two outs against L.A.’s Zack Greinke, who so far had not allowed a baserunner. Kelly pulled a 2-1 pitch through the 5.5 hole for a single. This sparked a rally that pulled St. Louis even with the Dodgers, in a game the Cardinals would win 3-2 in 13 innings.

Kelly’s hit was a key to the team’s success that inning, obvious long before extra innings arrived. But that’s not what got the attention of color commentator Ron Darling.

The former pitcher focused instead on Kelly’s work in the top of the fourth. For the first time that night, Kelly set down the Dodgers in order, after three earlier innings in which he had yielded five hits, two walks, a hit batter and a wild pitch. Darling believed he knew why Kelly had settled down: because he had singled in the previous inning. Darling opined that getting a hit makes a pitcher feel more a part of the game, gets him into the rhythm of play.

There are reasons why one would like to take Darling at his word. He was a pretty decent pitcher in his day, so he has the experience one would want backing up such words. He attended college at Yale, so his head is demonstrably not full of rocks. Along with Gary Cohen and Keith Hernandez, he forms a TV booth for the New York Mets that is probably the best-regarded broadcast team in the game not comprised entirely of Vin Scully. He ought to know what he’s talking about.

But we are a more skeptical, if not indeed a more cynical, bunch of baseball fans than that. We don’t take what the talking heads say at face value. If we’re thinking clearly, we may recall that the conventional wisdom points instead in the other direction. Isn’t getting on board and running the bases supposed to be bad for a pitcher? Why else would he need the warm-up jacket they give him at first base? And an inning in which he gets a hit should be a longer inning than usual anyway, giving his arm more time to cool off and stiffen up.

Of course, we don’t fall in line with conventional wisdom, either. So we’ve got two competing theories, Ron Darling’s and pretty much everybody else’s, and so far nothing but anecdotes to choose between them.

Time to bring in some data.

One-hit (or more) wonders

Baseball-Reference credits 623 hits to pitchers during the 2013 regular season: 597 in the National League, and 26 for American League pitchers. I looked up every hit a pitcher made this year, and came up instead with 624, the extra hit coming in the NL. There is actually one further pitcher hit credited in the team-by-team numbers, but this is due to Pittsburgh Pirate Wandy Rodriguez being used as a pinch-hitter in an April 27 game against St. Louis. As we’re looking at how hitting affects pitching performance, this is an irrelevant event for us, however cool it was for Wandy.

There were also three occasions this season when position players, who had gotten hits earlier in the game, were moved to the mound to eat some at-bats in blowouts. They were Dodgers second baseman Skip Schumaker on April 29, Yankees shortstop Alberto Gonzalez on May 15, and Astros emergency catcher Jake Elmore on Aug. 19. (All three managed scoreless stints.) They don’t count for this survey, and wouldn’t have even if any of them had gotten a hit while in as pitcher. (Schumaker came closest, coming to bat after pitching the ninth. He struck out swinging.)

Since the theory being tested is that pitchers get into the flow of the game and pitch better after getting a hit, I am using the first hit a pitcher gets in a game as my dividing point. Subsequent hits will not be looked at. In case you think this might be an event roughly as common as a confirmed Bigfoot sighting, there were 53 occasions in 2013 when a pitcher had two hits in a game, and 11 where he got three. Adam Wainwright and Mike Leake each managed the latter trick twice, which involves not just hitting well but lasting long enough on the hill to get at least three at-bats.

This leaves 549 games in which a pitcher got at least one hit. I will look at the before-and-after in two ways. I will compare the inning the pitcher threw before getting his hit with the inning after; I will also compare the pitcher’s whole game before the (first) hit to what he did after that hit.

I could have tried measuring the before or the after against the pitchers’ ERAs or league ERAs, but there are possible biases in that method. I’m selecting for games where pitchers get hits. That could mean, because of park or weather effects, that it was a higher-run environment that day. Conversely, you’d expect pitchers to bat better against poor teams, and such teams would bat poorly against him. (I’m confident there is a significant correlation between quality of offense and quality of pitching and defense for major league teams.) Rather than wade through all that, I’m keeping the comparisons direct.

On rare occasions, a pitcher gets removed promptly after his hit, so that he faces no post-hit batters. I will exclude them from consideration, lowering the total by four to 545. There was also one instance in 2013 where a pitcher did not face a batter before getting his first hit. On Aug. 28, Homer Bailey‘s Reds batted around against St. Louis in the top of the first, and Bailey knocked an RBI single. He went on to throw 7.1 scoreless innings (and pop another single), but without a “before” to balance against the “after,” I can’t use his numbers. His great effort becomes a mere anecdote.

Fractional innings, before or after the hit, do count for my purposes. Instances where a pitcher records no outs after his hit, but does face at least one batter, also count: I consider that to be a partial inning. These judgments allow, among other things, for relievers to get their full due in this survey. There were 16 hits by relievers in 2013, and while we lose one for being pulled right after his hit, we preserve five others by allowing fractional innings before or after.

There are several wrinkles that pop up in making these comparisons, and the biggest one is fairly obvious. Pitchers lose effectiveness the deeper they go into games. If batting success has no effect on pitching prowess, we will see pre-hit performance beating out post-hit performance. It could take a substantial hit benefit to erase that natural tilt. It could also be tough to judge whether a small pre-hit advantage means the hit is actually having a positive effect.

As it turns out, this didn’t end up being a problem.

Making a hit, taking a hit

I’ll start with the full-game numbers for successful hitters heading back to the mound.

                         IP     R    ER    H     BB    K
Game before first hit  1789.2  587  544  1510   481  1450  
Game after first hit   1594.2  725  695  1549   444  1254

                        R/9  ER/9   H/9  BB/9   K/9  K/BB
Game before first hit  2.95  2.74  7.59  2.42  7.29  3.01
Game after first hit   4.09  3.92  8.74  2.51  7.08  2.82

The best that can be said for this set of figures is that the results most under the pitcher’s control, walks and strikeouts per nine innings, don’t deteriorate badly. The K/BB ratio drops by about 6.3 percent from pre-hit to post-hit. I compared this to data for starting pitcher performance given in The Book, by times through the order. From first time to third time, the authors found a drop of 8.4 percent in K/BB ratio.

The pitchers counted in this survey went an average of 6.167 innings per appearance, starting and relieving. This would make about 3.08 innings between midpoints of the pre-hit and post-hit stretches. (This stays true wherever the dividing line is.) The pitchers examined in The Book were in a high-run environment, so I’ll generously estimate they faced 4.5 batters per innings, as opposed to 4.22 for 2013. That means their 8.4 percent K/BB drop takes place over about four innings. 6.3 percent divided by 3.08 innings is roughly equal to 8.4 percent divided by four innings. There’s no obvious difference there.

It looks a whole lot more obvious in other categories. Hits leap by 1.15 per nine innings, or by over 15 percent. The increment recorded in The Book between first time and third time through the order is 7.7 percent. The rate of growing hit vulnerability effectively doubles, or more.

As for runs and earned runs, The Book gives us no help. We’ll go to 2013 National League statistics (because most of the pitchers getting hits play in the NL), but we need to be selective. The highest ERA for any inning in 2013, by a wide margin, is the first inning. There’s a plain reason: batting orders are set up to produce in that inning. For that reason, measuring from the first inning to some other endpoint would be deceptive.

Instead, I’ll use the difference between the second inning and the fourth. Going past the fourth inning gets us into territory where relief pitchers start making serious inroads, and since the pitching splits at Baseball-Reference don’t give us inning-by-inning numbers broken out for starters alone, the numbers would start skewing. Second-inning ERA was 3.51; fourth-inning ERA was 3.93. From this, we can roughly estimate that pitchers (starters, really, but I’ll stretch the point to cover the few relievers in the survey) worsen by roughly 0.20 points of ERA per inning.

That would mean we could expect something like an 0.6 to 0.65 rise in ERA from the pre-hit to the post-hit sections. What we see instead nearly doubles that amount. My calculation was rough, but not rough enough to explain away that difference.

This may have been the long way round to get somewhere that a quick glance at the leftward columns seemed to get us promptly, but calculation beats intuition, if you can make a decent calculation. Pitcher performance takes a substantial fall beyond normal deterioration after that pitcher gets a hit, over the full course of the game.

What happens in the shorter term, from the inning before to the inning after that first hit a pitcher gets?

                           IP    R    ER   H    BB   K
Inning before first hit  544.1  205  194  493  154  448
Inning after first hit   528.2  245  233  525  136  415

                          R/9  ER/9   H/9  BB/9   K/9  K/BB
Inning before first hit  3.39  3.21  8.15  2.55  7.41  2.91
Inning after first hit   4.17  3.97  8.94  2.32  7.06  3.05

The pattern we saw before repeats itself even more extremely. Walk rates actually fall in the short term, and while strikeouts do as well, the K/BB ratio ends up rising.

This would back Darling’s assertion, if the other numbers weren’t so strongly in opposition. The hit rate, which we’d expect to creep up about two percent, leaps by nearly 10 percent. Runs and earned runs, which ought to rise by about a fifth of a point per nine innings, instead soar by over three quarters of a point. The erosion of pitcher effectiveness in hits and runs is greater in the narrow focus than over the long haul.

The drop-offs don’t occur where we’ve lately been conditioned to expect the core of pitcher performance to show up, walks and strikeouts, but they happens where it matters, in runs. To dismiss the result because it isn’t shifting the “true outcomes” is, to my thinking, missing the forest for the trees. Granted, pitchers aren’t supposed to have strong control over how many hits they yield, but what is the alternative? Are their fielders slacking off once the pitcher gets a hit? Ludicrous. Are opposing batters taking umbrage at a pitcher who dares to successfully hit, and bearing down on him in retaliation? Laughable.

There’s something quite interesting going on with the hit rates, which apparently drives the parallel rise in run rates. I could speculate about it—maybe pitchers start feeling invincible, begin coming right after hitters, and in consequence groove a few pitches too many—but there’d be precious little information behind my guesses. I might worry about sample sizes, but at over 1,500 innings for the longer look, it seems okay. The conventional wisdom about pitchers losing oomph after a spell on the basepaths looks like it’s correct.

Conclusion

Ron Darling is a dope.

Conclusion that’s not unfair and rude to a respected broadcaster

Ron Darling was wrong. That much isn’t changing. But he came to this mistaken conclusion for a reason, and I want to discover it. The most obvious hypothesis to pursue is that he is reasoning from personal experience. If he got more into a game as a pitcher after getting into the game as a hitter, he may have projected this onto the rest of baseball.

This jump in logic would be modest in a way: It would assume that he was nothing special among pitchers, that he was experiencing something common to them all. It would also be self-centered in a way: It would assume that he could generalize from himself to all his fellow pitchers.

It may have sprung from taking pride in his hitting, which he himself has acknowledged. He was a good enough batter in college to DH for himself, and to play the field on days when he didn’t pitch. In the major leagues, for whatever reasons, this skill faded, and he was just somewhat better than average at the plate for a pitcher.

So did he experience what he says other pitchers do? I looked up the appropriate numbers for Ron Darling’s career, and did the same things with it as I did for pitchers in 2013. Darling got hits in 65 separate games, for 76 career hits. He got all 65 as a starting pitcher, and always faced at least one batter both before and after that first hit of the game. This made my calculations straightforward. How do they stack up for him, and against current pitchers?

                         IP     R    ER    H    BB    K
Game before first hit   221     67   61   173   84   163    
Game after first hit    231.2  114  100   222   86   148

                        R/9  ER/9   H/9  BB/9   K/9  K/BB
Game before first hit  2.73  2.48  7.05  3.42  6.64  1.94
Game after first hit   4.43  3.88  8.62  3.34  5.75  1.72

For the full game, the numbers are an even bigger wipeout for Darling than for 2013 pitchers. Runs, earned runs, and hits make a bigger leap after the ice-breaking hit, and strikeouts take a much bigger plunge. Darling’s walks inched down rather than up, which did not prevent his strikeout-to-walk ratio from dropping by over 11 percent, a bigger fall than for the class of 2013.

It’s a disaster. If anything, Darling should have been learning the exact opposite lesson and becoming an evangelist for the designated hitter, so pitchers don’t ruin themselves by getting hits. Only if the inning-to-inning numbers are completely different could it be otherwise.

                          IP     R  ER   H  BB   K
Inning before first hit   65    22  22  53  23  36  
Inning after first hit    62.1  19  18  54  15  41

                          R/9  ER/9   H/9  BB/9   K/9  K/BB
Inning before first hit  3.05  3.05  7.34  3.18  4.98  1.57
Inning after first hit   2.74  2.60  7.80  2.17  5.92  2.73

Well, son of a gun. The inning-to-inning numbers are completely different.

That’s a slight exaggeration. His hits per nine innings go up about six percent: one would expect roughly two percent from natural weakening from one inning to the next. All the other numbers, however, get better, and markedly so. Most impressively, he drops a walk per nine, gains a strikeout per nine, and his K/BB ratio goes through the roof.

This does come from a small sample, but in this case that doesn’t much matter. This was Ron Darling’s whole major league career, seen through one narrow lens. If Darling was paying attention to what happened through this lens, he could easily have reached the conclusion that, in the short term if definitely not overall, getting a hit made him a better pitcher.

And, just maybe, it did. Perhaps there’s some individual variation that made him respond differently than most other pitchers. There’s no telling. At least, not yet.

Conclusion, and this time I mean it

The entrenched conventional wisdom has it right in this case. A pitcher’s success at the plate sets him up for failure back on the mound, at least in the current game. Ron Darling’s own experiences in the inning after getting his first hit of a game may well have blinded him to the greater trend.

As regular readers here may have observed, I have an appreciation, or at least an interest, in how pitchers fare when batting. While I tolerate the designated hitter, I would be sorry to see pitcher batting pass out of the game. Here, though, we have a surprise argument against pitcher batting. Not only does concentration on the specialty of pitching worsen a player’s hitting, but the successes they do manage with the bat come back to haunt them on the mound.

I cannot say I like this result. It’s more aesthetically pleasing to me for someone’s play to be an organic whole, success in one aspect supporting success in the others. That might be the source of the myth that defenders who make spectacular plays to end one inning lead off the next more often than chance dictates. That player is primed to perform in one task, and comes through in another. It seems natural.

The numbers above are not impressed with what I like and dislike, what seems natural or unnatural. And I can’t bury them just because I don’t like what they say.

It would be nice to understand the whys and wherefores beneath them, but that may get into the psychological aspects of the game, and those are more slippery than anything else in baseball, not to mention maddeningly resistant to being frozen into statistics. I might try looking into some other possible synergies between components of play in the future, to see if the phenomenon holds in other ways.

For now, though, we may have to settle for hoping the pitcher just bunts the runner over whenever he comes up.

References & Resources
Pitching and hitting data, as usual, came from Baseball-Reference. Overly familiar comments about Ron Darling’s hitting were based on his memoir, The Complete Game, and on the classic Roger Angell article “The Web of the Game,” in which Angell recounted the equally classic Yale-St. John’s playoff pitching duel between Darling and Frank Viola.

Print Friendly
 Share on Facebook0Tweet about this on Twitter0Share on Google+0Share on Reddit0Email this to someone
« Previous: Houston hoodoo ‘62
Next: Player-A-Day: Tim Lincecum »

Comments

  1. Bob Burpee said...

    Shane,

    Thank you for an informative article.  I wonder what a similar study including all other times that a pitcher gets on base (walks, fielder’s choices, etc.) would reveal.  I suspect it would show the same.  But if it didn’t, we would have to look for a different explanation.

  2. MGL said...

    First of all, those are such typical, ludicrous claims from ex-player broadcasters, that it is a near-waste of time trying to verify or debunk it.

    Next time you or anyone else wants to test one of those “commentator wisdoms,” please tell me before hand and I’ll allow anyone to wager a large amount of money that they (the commentators) are wrong, AND I’ll offer very long odds (like 10-1).

    Anyway, did you consider controlling for the batters in the before and after samples? I would have to assume that if a pitcher gets a hit, he obviously had to come to the plate, and that would bias the slot in the order for the other team. Maybe the bias is small, I don’t know, but I think I would want to know that…

  3. MGL said...

    “That could mean, because of park or weather effects, that it was a higher-run environment that day. Conversely, you’d expect pitchers to bat better against poor teams, and such teams would bat poorly against him. (I’m confident there is a significant correlation between quality of offense and quality of pitching and defense for major league teams.) Rather than wade through all that, I’m keeping the comparisons direct.”

    Oh, and I would definitely not worry about that. A pitcher getting a hit contains MUCH more noise than any signal associated with the things you mention. I can almost guarantee that you are not going to find any significant differences in the environment in games in which a pitcher gets a hit versus games in which they don’t, especially if you don’t include Coors games!

  4. MGL said...

    On MORE thing:

    People talk about how much they like and respect Darling (and Hernandez). I guess how much respect you have for a commentator is all relative (to the other commentators), but both Darling and Hernandez make stupid comments, observations, and “insights” like that ALL the time. Since those guys ARE some of the best and smartest, again, relatively speaking, I suppose that is why I hate ALL these guys…

  5. ajnrules said...

    As a Rangers fan I’m all too aware of how Cliff Lee fell apart in Game 1 of the 2010 World Series. Back then it was speculated that the exertion from the base running led to the decreased efficacy. It’s good to have the numbers to back things up.

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>

Current day month ye@r *