Author Archive

Delayed Overanalysis of Casey Janssen

The Nats signed reliever Casey Janssen, formerly of the Blue Jays, to a one-year, $5-million contract a few weeks ago (feel free to stop reading now to avoid the existential dread associated with over-analyzing Casey Janssen). Overall, it’s hard not to like this pick-up. One year and five million dollars is basically nothing (except when it comes to signing a second baseman), and the Clippard trade certainly left a hole in the bullpen. There was also a recent stretch of time when Janssen was quite good. From 2011-2013, Janssen averaged 57.1 IP, 8/9 strikeouts per 9 innings, and a sub 3 FIP. WAR isn’t the best way to measure relievers, but he averaged 1.2 WAR a season over those three years, which put him squarely in the pretty damn good category of relief pitchers.

So why did a recently good closer sign for a seemingly below market sum? Because 2014 was mostly terrible. Strikeouts were way down (5.5 Ks/9 in 2014 compared to 8.5 in 2013), Homers were way up (1.2 HR/9 in 2014 compared to 0.5 in 2013), and his groundball percentage dropped from 48% to 34%. These are all fairly alarming trends for a relief pitcher that is 33 and doesn’t throw very hard (2014 average fastball velocity: 89.3 miles per hour). Every analysis for relief pitchers should contain small sample size warnings in all capital letters, but important indicators trending that strongly generally indicate something wrong happening.

In July of last season, Janssen came down with a particularly awful bout of food poisoning, and he probably came back too quickly. And looking at the mid-season splits, there’s a case to be made that it was the (negative) turning point for the rest of Janssen’s season. Let’s compare:

1st half: 22 IP, 1.23 ERA, 0 HR, 14 Ks, 1 BB, .218 wOBA against
2nd half: 23.2 IP, 6.46 ERA, 6 HR, 14 Ks, 6 BBs, .378 wOBA against

In the first half, Janssen made opposing hitters look like Austin Kearns. In the second half, they all looked like Yasiel Puig. His numbers did take a nosedive in July when he was sick, but got worse in August when one would have expected him to be feeling better (or put on the DL to recuperate). It’s impossible for anyone to really know how he was feeling, and if food poisoning actually was the main cause of Janssen’s second-half struggles. But, his velocity didn’t change from the first half to the second half, and his strikeout rate remained about the same. The uptick in walks and home runs in the second half are troubling, but maybe first-half Janssen was a fluke based on a year over year decrease in velocity (lost about .8 MPH on his fastball from 2013 to 2014)  and a decrease in strikeouts. For comparisons sake, here is an unnamed reliever’s 1st and 2nd half splits in 2014:

1st half: 37 IP, .97 ERA, 1 HR, 36 Ks, 11 BBs, .208 wOBA against
2nd half: 25 IP, 6.48 ERA, 3 HR, 23 Ks, 8 BBs, .375 wOBA against

This reliever? Rafael Soriano. There wasn’t an injury narrative to fault for his falling off a cliff bad second half, but he stunk nonetheless. Screwy things can happen in small samples, which is why we try to avoid over-analyzing them. Janssen may have just had impeccable timing, and his new true talent level as a command relief pitcher is that of a 4.00 ERA. But unlike with Soriano, there is a realistic narrative for Janssen that fits the timeline of his struggles. Here’s another 1st half/2nd half comparison

1st half:

2nd half:

While his K rate was basically the same from the first half the second half, these charts show that his whiff rates weren’t. Janssen had much more success both down and up in the zone earlier in the season in terms of swings and misses, so while his velocity was the same between the first half and second half, it appears that his stuff wasn’t.

Again, in such small samples, it’s impossible to draw any definitive conclusions. It’s true that first-half Janssen looked pretty similar to 2011-2013 Casey Janssen, while second-half Janssen looked more like Brian Bruney. It’s reasonable to look at the splits and say that Janssen’s bout with food poisoning ruined what looked to be a promising season. It’s also reasonable to look at his decrease in velocity and strikeout rate and think this was money not well spent. But for a paltry (in the context of the MLB) five million dollars, it’s not that much money anyways, so why the hell not?


What the Hell Happened to Rafael Soriano?

Now that the division title belongs to the Nats, and the race for the number one seed in the NL is pretty much locked up, there are still a few reasons to watch the rest of the regular season games (if you are a Nats fan). If I were an unbiased observer, I would find the whole Rafael Soriano situation fascinating. He was having a fantastic first half, and while his ERA was beating his peripherals by a decent margin, his peripherals were still pretty strong. There was reason to expect regression, but not reason to expect a full-on collapse. But Soriano has picked up over two runs on his ERA during this second half and gone from closer to “cross your fingers mop up guy.” While watching another mentally exhausting Soriano “save” on Sunday, I wanted to figure out what exactly had happened to a season that started out so promising.

One thing that is important to remember is that relievers are volatile, and a few bad outings can throw things out of whack. In September, Soriano has given up 7 ER in 7.2 IP. That’s awful, but it’s also only seven innings. You could find quite a few SPs this season who have had a stretch of 7.2 IP giving up 7 runs. Strikeouts haven’t been the problem either, as he is averaging over a K per inning, and a 3:1 K-BB ratio. And despite the recent blow up (by recent I mean the entire second half), Soriano is still sporting a .59 HR/9 ratio for the season, which is much lower than his career of .86 HR/9. His BABIP is the exact same as last year, his strikeouts are up significantly (6.89 in ’13 versus 8.70 this year), and while his walks are up too, the overall K-BB is stronger. Not to mention that he has the second best SwStr% of his career after posting a career low in the same metric last year. So with all these seemingly positive things happening, what’s the deal? Where has this implosion come from if it doesn’t stem from gophers or a high BABIP against?

I think the answer is two-fold: extra-base hits, and a lack of infield fly balls. Below is a chart from 2013 of hit types against Soriano:

Here is 2014:

There are two important takeaways from this chart. One, even though Soriano gave up more dingers in 2013, he has given up significantly more extra-base hits this season. By my count, he gave up 15 extra base hits last year, and 21 this year (including home runs). Six may not sound like a lot, but that’s a 40% increase. When you only throw 60 innings a season, that makes a huge difference.

Two, look at the location of the outs in 2014 compared to 2013. Notice how there are way more silver dots in the infield in 2013. As a pitcher, infield fly balls are the second best thing to strikeouts. They are an out basically 100% of the time and runners can’t advance on an out like they can on a deep fly ball or a grounder. Soriano went from a 16.3% infield fly ball percentage in 2013 to 7.4% this year. A pattern is forming here. For a guy who pitches with runners on base fairly frequently, infield fly balls and strikeouts are a fantastic way to get out of a jam. Even though Soriano has more Ks this year, he also has far fewer IFFB, which almost offset one another. A lower IFFB% despite a higher overall fly ball percentage from 2013 explains a lot of what’s happening here. More balls in the outfield leads to more extra-base hits or even runners advancing/scoring on an out.

I came up with a quick metric I’ll call “Nearly Automatic Out Percentage” to illustrate my point. Soriano has faced 248 total batters this year compared to 277 last year. He has 59 Ks and 6 IFFBs this year (65 total nearly automatic outs) for a NAO% of 23.5%, compared to 51 Ks and 14 IFFBs in 2013 (also 65 nearly automatic outs), a NAO% of 26.2%. These numbers are closer than I would have thought considering how much better Soriano has been with Ks this season. But when you factor in the additional extra-base hits and a few additional walks/HBP, it explains how the end result in 2014 can be so similar to 2013 (nearly the same WAR, ERA, and xFIP in 2013 and 2014) in two completely different ways.