Is the Royals’ contact rate responsible for their World Series championship?
It’s time for everyone’s favorite offseason game. Let’s look at this year’s World Series winner, find some fun fact about it, and assume that fun fact was the reason that it won the World Series! Now entering the studio are this year’s contestants. A contact-making team from Kansas City, Missouri: The Royals!
Even as the Royals entered the World Series, people were wondering in an "irresistible force meets immovable object sort of way" what would happen when the Royals, who led the league in the percentage of the team’s swings that resulted in contact, met the Mets, who had several pitchers who specialized in getting hitters to swing and miss. Of course, the Royals won the Series, so it must have been the contact, right?
We do know a few things. The Royals, either through design or sheer luck, did lead the majors in contact percentage with 81.9 percent of the swings that they took making landfall on the ball, even if just to foul it off. Should teams start seeking out hitters this offseason who are good at making contact? Maybe now would be a good time to point out that the Royals finished in a virtual tie in their team contact rate with the Oakland A’s (also at 81.9 percent). Given that the A’s finished with a 68-94 record, I don’t think that we can call contact rate the #NewMoneyball yet.
But there is a certain logic to the whole thing. After all, we live in an era where a lot of teams are loading up on power arms, especially in the bullpen, with a goal to make hitters swing and miss. Strikeouts are up. Why not find players who can foul pitches off, even if they are 96 mph, and then hit the inevitable pitch that gets just a little too much of the plate? Even if it’s just putting the ball in play for an out, a groundball could advance a runner while a strikeout can’t. On the other hand, maybe the Royals succeeded because they had good hitters who, on a somewhat unrelated note, happened to make a lot of contact.
Let’s put it to the test.
Warning! Gory Mathematical Details Ahead!
The math in this one is going to get a little complex.
The first question worth asking is the one about the irresistible force and the immovable object. Suppose that a pitcher who – in general — induces a miss on 30 percent of swings matches up against a hitter who only misses 10 percent of the time when he swings. If we know that the batter will swing at the next pitch, what percentage chance should we give him of making contact? The intuitive thing is to say "well, splitting the difference, we’d expect a 20 percent chance of a swing and miss." But that’s not how things actually work.
Instead, we’re going to use something called variance decomposition. The best way to explain it is to picture a glass of water into which you pour both red and blue dye (and then shake it up). The result will be purple, but what the exact shade of purple depends a lot on how much red and how much blue went in. Will it be more of a reddish-purple or a bluish-purple? Now, suppose that all you had was the purple stuff. Could you deconstruct and figure out how much red and how much blue went into it? That’s essentially what variance decomposition does.
For the initiated, I classified each swing that a player took between 2010-2014 as either a swing-and-miss (0) or contact (1). I then took the pitcher’s contact rate allowed for that year (minimum 250 batters faced), and the batter’s contact rate (minimum 250 plate appearances) for that year, as well as the overall league rate. (For the super-initiated, I converted them into natural log-odds terms.) I then ran a binary logistic regression using those three factors to predict back to the individual cases. I used the change in -2 log likelihood as my indicator of variance accounted for.
I told you it was about to get #GoryMath in here.
The idea is that if it were an extreme case where the pitcher has total "control" over whether a batter would swing and miss, then batters would always swing and miss at a rate that was consistent with the pitcher’s overall rate. And any differences between batters would be related to the quality of the pitchers that they faced, since they are powerless to stop it. Nothing is ever quite that extreme, but it’s always interesting to see what shakes out of these analyses. Sometimes you find that some outcomes are much more in the control of someone that you didn’t expect.
In this case, it turns out that the recipe is that a swing-and-miss (or a swing-and-contact) is actually about two-thirds the fault (or credit) of the batter and one-third of the pitcher. (The league rate picked up a very small portion of the variance.) In practical terms, it means that if a good contact hitter is at bat against a guy with swing-and-miss stuff, the results that we would see are more likely to be tilted toward (although not completely) the batter’s tendencies. In the Royals’ case, it means that if they were specifically assembling a team made entirely of contact hitters with the idea to try to blunt the rise of the smoke-filled, swing-and-miss bullpen, then they were correct that they’d be able to make more contact than they expected. But would that make a difference?
We’ve seen that the Royals and A’s had roughly the same high contact rate, although the Royals finished seventh overall in run scoring, while the A’s finished 14th. The Blue Jays, who led everyone in run scoring finished 11th in making contact. Contact is only good if the ball either falls where a fielder isn’t or goes over the fence. But perhaps having a high contact rate can make it a little more likely that a hitter can fish a good outcome out of the prize box.
What we do know is that contact hitters usually have specific outcomes that they tend to favor. For example, using data from 2010-2014 (among batters with more than 250 PA), we find that contact rate is negatively correlated with strikeout rate (more contact, fewer strikeouts; the correlation coefficient, which is abbreviated "r", is -.897, which is exceptionally strong.) Cutting down on strikeouts is a good thing.
However, contact rate is not generally associated with walks (r = -.181). Contact hitters instead tend to favor singles (r = .582) and being out on balls in play (r = .786), although their home run rates tend to suffer (r = -.525). Some of those strikeouts are probably turning into grounders to second, and while those are still outs, it’s been pointed out elsewhere that they at least have the chance of advancing a runner on base, while a strikeout does not.
While those findings probably don’t surprise anyone, there’s a hidden secret in there. Before assuming that all contact hitters are slappy singles hitters with no power, note that the correlation between contact and singles is +.58 and the correlation between contact and home runs is -.52. The positive sign in front of the singles correlation means that more contact is associated with more singles, and the negative sign in front of the correlation for home runs means more contact, fewer home runs. What you really want to pay attention to is the number afterward.
In statistical terms, those are moderately strong correlations. (In correlations, zero means that there is no relationship between the two variables. One means that they track each other perfectly.) A correlation around .50 means that contact and home run power don’t track each other perfectly. There’s some room in there for a guy with good contact skills and who has some home run power. They are out there. On the flip side, it means that there’s a guy who can hit home runs, but who isn’t a just swinging through everything hoping that he might hit it, like we normally stereotype power hitters to be.
They might not be plentiful, but they exist. And if no one is looking for them specifically, they might be cheap relative to the value that they can provide. In other words, a market inefficiency.
I Don’t Think We Were Ready For That Royal Jelly
I don’t know whether the Royals specifically sought out contact percentage when putting together their team. Maybe it was in there somewhere, but it turned out to be a pretty handy evolution for the baseball ecosystem in which we now live. In a world where everyone seems to have swing-and-miss stuff, the bad effects of that can be blunted (although not completely erased) by hiring batters who don’t like to swing and miss. Pitchers, it seems, do not have the ability to throw the ball by hitters at will. They need the cooperation of the batter.
The genius of the contact-based approach is that, if a pitcher is reliant on making hitters miss as part of his game plan, a contact-based team will have an antidote to his poison. Now it’s a negotiation between the pitcher and batter in terms of how good the pitcher is at inducing weak contact and how skilled the batter is at making good contact. If teams are selecting for swing-and-miss stuff and ignoring whether the pitcher is also skilled at inducing weak contact, then teams that emphasize a swing-and-hit approach and can find players who make decent contact will have plenty of guys to pick apart.
If fighting a battle under one set of rules isn’t working, change the rules. These effects aren’t the only reason that Royals fans can wear "Defending World Champions" shirts next year. I’m told by well-placed sources that they had a decent bullpen. It’s also not the case that the contact approach is the only way to win baseball games. But smart fans would do well to pay attention to the natural evolution of the game. For a few years, we have worried about the strikeout scourge and the drop in offense. Perhaps this is just the counter-move, and it was thrown into the limelight … or perhaps the royal blue light by some gentlemen from Kansas City who are now holding a trophy.