A Quick Note On Michael Pineda’s Splits

Since the swap of Michael Pineda for Jesus Montero became public last night, a few talking points have become pretty commonplace. Some of them – Pineda’s reliance on his fastball and slider due to a subpar change-up, for instance – are definitely true, and are stories supported by the evidence. However, there have also been a few points that have been raised that don’t stand up to closer scrutiny, and those mostly stem from cursory looks at Pineda’s splits.

Split data, by its nature, generally consists of small sample sizes. In breaking a season down into smaller slices, you’re necessarily introducing greater uncertainty into the numbers. It’s important to not draw too many strong conclusions from what appear to be trends in split data, and at the same time, to make sure you’re looking at the entire picture.

With Pineda, two of his 2011 splits are most commonly cited as reasons for the Yankees to have some concern about his future performance – his home/road splits (specifically, the 2.92/4.40 ERA numbers) and his first half/second half splits (3.03/5.12 ERA). From these numbers, questions have been raised about how well Pineda will do outside of spacious Safeco Field and whether he’ll be able to hold up over a full-season and still be able to pitch well for the Yankees in the playoffs.

In both cases, however, looking a little deeper than simple ERA shows that these concerns are probably overblown.

Let’s start with his home and road numbers. For context, here are the relevant numbers for both:

Split BB% K% GB% HR/FB BABIP LOB% FIP xFIP
Home 9.1% 26.6% 35.4% 10.5% 0.220 77.5% 3.62 3.51
Away 7.0% 23.5% 37.0% 7.8% 0.286 64.4% 3.26 3.55

His walk and strikeout rates are both slightly higher at home, though the numbers are close enough to essentially be equal. His GB% is slightly higher on the road, but again, there’s not a big enough gap to really read into that at all. In those three core categories, there’s really nothing to show that Pineda actually pitched much better at home.

In fact, if there was one area where you’d expect Safeco’s park factor to show up, it would be in HR/FB rate. Safeco’s a pretty big ballpark and heavily suppresses home runs for right-handed batters, but Pineda’s HR/FB rate was actually lower on the road than it was in Seattle. While this might be a bit surprising, the reality is that Safeco actually isn’t all that beneficial to right-handed pitchers, as the right field porch is fairly short and the ball carries pretty well that direction. Left-handed pull hitters – the ones most likely to take Pineda deep – do just fine in Safeco. This is one of those times that a general park factor just isn’t all that helpful, because the park’s asymmetrical dimensions create very different effects for lefties and righties.

So, why was his ERA at home so much lower? Well, the chart should make it pretty obvious. Not only was his BABIP 66 points lower at home than on the road, but his road strand rate was just absurdly low. A pitcher who posts a 3.26 FIP and a .286 BABIP, as Pineda did away from home a year ago, should strand something close to 75% of his baserunners. Pineda stranded just 64%, the kind of total that is so low that you’d expect improvement from even the worst pitcher in baseball.

Essentially, his road ERA was just artificially inflated because of a strand rate that has no real predictive value. And there’s no reason to believe that Michael Pineda is going to strand runners on the road again next season. That’s one of those numbers, like a player’s batting average in Thursday day games, that is basically just trivia rather than anything useful. This is a case where home/road ERA isn’t giving us any insight into how a pitcher’s home park helped him perform better. It’s a number that’s best ignored, honestly – it doesn’t really tell us anything about what we should expect in 2012 and beyond.

Now, for the second half fade narrative. Here’s the same table as above, just with his performances by month:

Split BB% K% GB% HR/FB BABIP LOB% FIP xFIP
Mar/Apr 9.5% 23.8% 30.9% 0.0% 0.262 76.5% 2.26 3.90
May 5.6% 28.6% 36.6% 10.8% 0.231 82.0% 3.15 2.98
Jun 8.3% 21.2% 26.5% 7.5% 0.243 76.1% 3.90 4.28
Jul 9.3% 28.8% 40.6% 14.3% 0.294 46.1% 3.74 3.14
Aug 7.4% 24.2% 46.8% 19.0% 0.262 65.4% 4.20 3.09
Sept/Oct 6.7% 22.7% 45.3% 9.5% 0.275 71.4% 3.41 3.44

Again, we’re looking at a scenario where ERA isn’t really giving us an accurate picture of what happened. As you can see by looking at the numbers, the spike in runs allowed was basically due to a regression in his BABIP and HR/FB rates from the first half of the season. His K/BB ratio is nearly identical from the first half to the second half, and Pineda actually started generating far more ground balls as the season wore on. Those numbers do have predictive value, and don’t support the idea that Pineda was “wearing down” in the second half of the season.

Still, it’s easy for people to see increases in hit rate and home run rate as evidence that Pineda was simply not pitching as well. After all, both a spike in BABIP and HR/FB rate could be explained by batters just hitting the ball harder. And they very well might have been – we don’t really have the tools necessary to eliminate that as a possibility.

But, we also have to be realistic in our baselines. Pineda’s first half BABIP was .247, and his HR/FB rate was 7.1%. Those numbers aren’t really sustainable for anyone, and so, no matter who the pitcher was or how many years he had in the big leagues, we would always expect him to perform worse in the second half after posting those numbers. And we’d expect anyone watching him regress to report that he “looked worse” than he did previously. It’s almost impossible to watch a guy’s performance regress to the mean and not think that he’s pitching worse, as our opinions are certainly influenced by the results that a pitcher gets – even if we understand that those results might not actually be telling us anything about how well he’ll do going forward.

Of course Pineda looked worse in the second half than he did in the first half – he was performing at a level through June that only guys like Roy Halladay can keep up, and he’s not that kind of dominating ace. The fact that he established a ridiculous standard in the first three months of the season doesn’t mean that is the standard we should actually hold him to.

But, because the story of a young pitcher tiring under a workload he’s never carried before is such an easy one to tell, and because ERA seemingly confirms that narrative, this is the story that is being told about Pineda. And, for whatever reason, people have decided to add in a drop-off in velocity to make it sound more convincing. Well, here’s Pineda’s velocity chart from last year:

His velocity was down in his final start when the team gave him four innings after a 10 day rest. But in his second to last start, his fastball average matched his total for the season, and he ran it up into the 97-98 range as usual. His velocity did jump around from start to start, but there’s simply no clear downward trendline anywhere in that graph.

The pitcher-will-struggle-leaving-Safeco story and the young-pitcher-wears-down-in-second-half story are easy ones to tell, and have been true of other pitchers in prior years, but we should demand more than a basic look at split-data ERA before we jump to those conclusions. In looking at Pineda’s case specifically, there’s not much evidence that either story is true here.

The Yankees got a really good arm who had a really good year. There are legitimate causes for concern about his future performance (as there are with any young pitcher, really), but they won’t be found in looking at his 2011 home/road or first half/second half splits.





Dave is the Managing Editor of FanGraphs.

72 Comments
Oldest
Newest Most Voted
Inline Feedbacks
View all comments
Scott Candage
12 years ago

I watched quite a few of Pineda’s starts last year. I’m sure there’s data to back this up, but there were starts that he didn’t come out firing at 97 until a few innings had passed. I’d be willing to bet that those games are represented by the lower velocities on the last graph. I also think his ERA and such suffered because he was left in some games too long. He had starts where he was in trouble in the 6th or 7th inning. On a team with a better bullpen he would have been pulled. Because Seattle didn’t have the best pen, he was left in and gave up runs that he might not with the Yankees.