Pitchers 'ahead' of hitters: Testing the narrative

With pitchers and catchers officially reported for spring training, we will soon hear rhetoric of pitchers being 'ahead' of hitters come Opening Day. This is an idea that many baseball fans largely understand to be true and take at face value. Having not really seen any real research to confirm or deny this trope, I decided to take it upon myself.

At first, this seemed like an impossibly easy thing to measure. You take a pitcher's performance early then compare it to a pitcher's performance later in the same season. In a perfect world, there should be an impossibly perfect negative correlation between pitchers' performances and hitters' performances around the league. Alas, the real world -- or at least the way we measure it -- is far from perfect.

So, I first sought to break down the actual phrase into an argument as opposed to the narrative you hear in which we emptily confide. Phrased as an argument, it would look something like this: pitchers perform better than hitters at the beginning of the season. The next step would be to break down pitcher performance. There are a few ways to do this. Hinted above, you could take all hitters' performances at the beginning of the season and compare it to later in the season to see if they 'catch up' to the pitchers. Unfortunately, this would be textbook affirming the consequent. If pitchers perform better than hitters at the beginning of the season, then they are 'ahead' of the hitters. Hitters perform worse at the beginning of the season. Therefore, pitchers perform better at the beginning of the season. It seems like an innocuous error to make -- especially because baseball really is largely about a pitcher against a hitter at most given times -- but it would still be fallacious.

I moved on and decided to look at starting pitchers only. More specifically, ones that made at least 6 starts in April and May of 2014. This may seem like a small sample size, but, if a trope like this is true, it should be true every season. This returned 58 pitchers. After that, I grabbed their month-by-month splits in both xFIP and wOBA-against. While xFIP measures a pitcher's ability to make his team's job easier, wOBA added some variance by measuring the hitter's actual performance against that particular pitcher. I then set the minimum amount of innings pitched in any particular month to 18. This kept in pitchers who made bad, short starts but removed ones that had an injury. That removed 4 pitchers (Jose Fernandez, Martin Perez, CC Sabathia, and Tony Cingrani) totally from the sample pool.

Among those 54 pitchers, the average xFIP was 3.42, and the average wOBA-against was 0.298. Those seem to be pretty sterling numbers off the top. Unfortunately four more pitchers didn't make the cut for the following month (Zach McAllister, Andrew Cashner, Alex Wood, and Dillon Gee). However, in the following month, those 50 pitchers had an xFIP of 3.57 and a wOBA-against of 0.352.

Moving on to June saw Cliff Lee, Michael Wacha, Jordan Lyles, Robbie Ross, Brett Oberholtzer, and Francisco Liriano removed from the sample, while Cashner was re-instated. Those 45 pitchers now had a combined xFIP of 3.62 and a wOBA-against of 0.312. It was definitely beginning to look like narrative prevailed. Not only did pitchers seem to slow down in May, but they got even worse in June.

However, even though the average seems to show us that the narrative is true, it would actually make more sense to look at the average difference between each individual pitcher's scores. Even though this looked relatively conclusive, it was possible that the pitchers who got injured and didn't pitch in the subsequent month were somehow distorting the data set.

On average, the 50 pitchers who made it to May saw their wOBA-against climb by 0.007 points. Not totally insignificant and alongside the averages a number that probably shouldn't be dismissed as random chance. Similarly, their xFIP climbed 0.162 points. That's a growth of 2.3% in wOBA-against and 4.7% in xFIP. Not nearly as decisive as the above numbers suggested yet not easily dismissed. While it is curious that xFIP could be driven up at twice the rate it is not necessarily telling of anything. It is possible that pitchers got worse at inducing groundballs and getting strikeouts (things xFIP likes) and hitters got better at hitting at slightly different rates.

From May to June, the 44 pitchers (Cashner once again omitted because he had no May to compare to) saw their wOBA-against climb 0.009 points and their xFIP climb 0.068 points on average. Alas, all signs point to pitchers indeed being 'ahead' of hitters come Opening Day. All told, on average, starting pitchers saw their xFIP grow 0.212 points (6.19%) and their wOBA-against grow 0.017 points (5.61%) from Opening Day to the end of June.

Just for fun -- and because I gathered the data anyway -- look how much worse those pitchers were in September compared to their March/April starts. 46 of those original 54 pitchers pitched in at least three months of 2014 where one was April and one was September. Of those 46, those pitchers saw their wOBA swell by a whopping, catastrophic, insurmountable 0.00037. They also saw their xFIP shrink by 0.001.

So, there you have it. It definitely seems like the narrative-fiends have this one. As for the last paragraph, I have no idea what to make of that. Perhaps it will turn into its own narrative that pitchers re-catch up with themselves. Or perhaps it was just random chance with this data set. I couldn't help but share it once I saw it though.

. . .

All statistics courtesy of FanGraphs.

Michael Bradburn is a Featured Writer for Beyond the Box Score. You can follow him on Twitter at @mwbii. You can also reach him at michaelwbii@gmail.com