18 September 2014

On the Orioles' Rotation: Better to Be Both Lucky and Good

According to FanGraphs' version of wins above replacement, three O's starters have been about the same -- Wei-Yin Chen (2.3), Chris Tillman (2.2), and Kevin Gausman (2.1) -- followed by Bud Norris (1.5) and Miguel Gonzalez (0.7). Ubaldo Jimenez (0.3) has barely been above replacement level. Per Baseball-Reference's version of WAR, Tillman leads (2.1), followed by Gonzalez and Chen (1.6 each), Norris (1.4), and Gausman (1.0). Jimenez also rates much worse (-0.7). Interestingly enough, two relievers, Darren O'Day and Zach Britton, lead O's pitchers in bWAR, at 2.4 and 2.1, respectively. (Britton (1.0) and O'Day (0.9) are t-fifth (with Andrew Miller) and seventh, respectively, in fWAR.) So neither fWAR nor bWAR rate O's starters highly.

A future Camden Depot post, I'm sure, will focus on the disappointing debut season for Jimenez in Baltimore. This is not that piece. And, hey, the Orioles clinched the American League East title Tuesday night, after all. So let's stay positive and focus on the five best starters in the rotation for the Orioles in 2014, who all happen to have earned run averages under 3.75. Let's also note that O's starters have pitched much better in the second portion of the season:

First half (more Ubaldo): 6.54 K/9, 3.30 BB/9, 4.09 ERA, 4.56 FIP, 4.36 xFIP
Second half (less Ubaldo): 7.30 K/9, 2.48 BB/9, 2.93 ERA, 3.66 FIP, 3.81 xFIP

As a reminder, fWAR for pitchers is based on FIP, which focuses on pitchers' strikeouts, walks, and home runs. bWAR for pitchers, however, uses Runs Allowed (earned and unearned runs) and innings pitched. Among all qualified MLB starters, Chen is 45th in fWAR. Among all qualified MLB starters ranked by bWAR, Tillman is tied for 40th. (Gausman and Gonzalez do not have enough innings to qualify.)

O's starters don't fare better in either metric for a few reasons. The strikeout leader of the five is Norris (7.36), tied for 19th in the AL. Chen has been great at limiting walks (1.66 BB/9, fifth in the AL), but the rest have walked between 2.69 and 3.22 per nine -- which is fine, but not outstanding. And they aren't skilled at inducing ground balls. They are all around 40% -- Norris is the best at 42.5% and Gonzalez the worst at 37.1%. So it's not surprising that just one starter (Gausman) has a lower FIP (3.31) than ERA (3.57). (As a reminder, ground balls are more likely than fly balls to become base hits, but a grounder has no chance of leaving the ballpark.)

This season, the average AL starter has posted the following numbers:

7.32 K/9, 2.73 BB/9, 0.93 HR/9, 43.1 GB%, 3.94 ERA, 3.86 FIP, 3.87 xFIP

And here's how the O's five best starters have done (in no particular order):


Two ways to determine "luck" are a low BABIP and a low HR/FB rate. Maybe Tillman and Gonzalez have been fortunate with low BABIPs, but they've done so throughout their careers. Gonzalez has posted similar overall numbers over three seasons, and he has managed to outpitch his FIP by more than a run in that period of time. Tillman, meanwhile, has outpitched his FIP by nearly half a run. Gausman has been the best at limiting home runs, but his 4.6% HR/FB rate isn't sustainable long term. (His average in parts of two seasons is 8.6%.) Tillman (8.5%) is below the league average (9.5%) by a percentage point, while Chen, Gonzalez, and Norris all exceed that number.

But one thing O's starters have been able to do well as a group is post higher than normal strand rates. The average AL starter in 2014 has a 72.1% strand rate. Gonzalez has a strand rate of 84.8%, followed by Chen (79.6%), Norris (77.5%), Tillman (77%), and then Gausman (72.7%). Among qualified AL starters, Chen (3rd), Norris (7th), and Tillman (8th) are in the top 10 among best strand rates. Lots of ace-type pitchers are in this group, including Chris Sale, Felix Hernandez, Jon Lester, and Max Scherzer, but you likely wouldn't put any of the O's starters in that class.

Gonzalez and Chen have been great at stranding runners throughout their careers, while Tillman has been above average, Norris about average, and Gausman below average. High-strikeout pitchers (like the names mentioned above) are typically better at stranding runners because of the ability to limit contact, but as mentioned above, the O's rotation does not rack up a ton of strikeouts.

O's starters have not really been fortunate BABIP or HR/FB wise, but they have been strand rate wise. So how has the rotation been able to maintain such a high strand rate? As with most things, several factors likely come in to play. First, they've been able to stay healthy. Athletes are, obviously, more likely to be effective if they are in good health. The Orioles have used only seven starters this year, and one of them (T.J. McFarland) was needed for a single spot start. The only significant injury was Ubaldo Jimenez's sprained ankle, and while it was unfortunate, the O's rotation has been better without him. Second, the O's defense has been outstanding. The O's rank first in UZR and second in DRS, and that has partly been fueled by Adam Jones and Nick Markakis performing better defensively (at least according to the advanced defensive metrics) than in recent years. That's pretty remarkable considering the O's only got half a season out of Manny Machado, who may be the team's best defensive player. A big part of the O's defense also may be defensive shifting, which is something that's deployed by most major league clubs. Third, the addition of pitching coach Dave Wallace and bullpen coach Dom Chiti perhaps has played a part.

There's also the Buck Showalter factor. I pointed out the following on Monday:

Showalter's first full season as O's manager was in 2011. I doubt Showalter has some secret formula for stranding baserunners, but the Orioles have steadily improved in that department under his watch.

Regardless, what the Orioles have been able to do is put together a group of good but not necessarily great flyball starting pitchers for little cost. In hindsight, they probably did not need Jimenez, but there was no way of knowing they'd pitch this well as a group. Norris is making $5.3 million; Chen about $4 million; Tillman $0.55 million; Gonzalez $0.53 million; and Gausman $0.50 million. Norris was acquired in a trade in 2013 with the Astros; Chen was a shrewd free agent signing; Tillman was picked up by Andy MacPhail (along with Adam Jones) in the often celebrated Erik Bedard trade; Gonzalez was an unheralded minor league signing in 2012; and Gausman was a first-round draft pick in 2012.

Norris has one more year of arbitration before he can become a free agent in 2016. Tillman and Gonzalez are both arbitration eligible next season but cannot become free agents until 2018. Chen has a $4.75 million club option that will assuredly be picked up before becoming a free agent in 2016. And Gausman is under team control until 2020. And then there's Jimenez, who still has three years left on his four-year, $50 million contract. This group will be a bit more expensive next season -- Norris, Tillman, and Gonzalez will all receive substantial raises -- but it also may not be quite the question mark many figured it was before the season. Then again, Jimenez isn't going anywhere, so someone like Norris may have to go. The O's aren't paying Jimenez all that money to sit in the bullpen and pitch a few innings every few days. But there's plenty of time to worry about that later.

Stats as of September 17. Photo by Keith Allison


Unknown said...

RE: the Buck Showalter effect. Just an idea, but one reason Buck could be (at least) partially responsible for better strand rates over the years is the coaching staff's insistence on the pitchers being quick to the plate with runners on base. Combine that with Caleb Joseph throwing out a bunch of runners, and you're strand rate will probably improve.

It's easier to strand a runner on first than it is to strand a runner on second or third.

Matt Kremnitzer said...

Sure, that could be part of it. O's catchers have been better than league average at throwing out opposing baserunners from 2011 on. A lot of that is probably due to Matt Wieters.

Philip said...

I respect Billy Beane for having the guts t dump Jim Johnson when it became clear that he wasn't what Beane had hoped.
50 Million is a lot more than ten, but what are the chances the orioles decide to cut their losses and dump Jiminez?
they'd get little for him, and probably have to eat some saary to boot, so it's unlike to happen before the season, but the notion of dumping a successful, inexpensive pitcher just so you can keep the expensive failure is really foolish.

BTW, Matt, regarding Wieters: Doesnt caleb Joseph have a better caught stealing rate than Wieters?

Matt Kremnitzer said...

It's not impossible, and if Jimenez only had one year left, sure, I'd say the O's would consider dumping him. But with three years left, that would be a lot of money and I just don't see it happening.

And I mentioned Bud Norris departing because he'll be in his last year of arbitration and won't be cheap, but he's just one example. It will be worth revisiting.

Yes, Joseph is at 42% while Wieters is at 33% (career). Sample size should be factored in. But Joseph wasn't around before this season, obviously.

m3tan said...
This comment has been removed by the author.
m3tan said...

The pitchers being healthy and the O's having an excellent defense would have no bearing on strand rate. Those attributes are true whether the bases are empty or not. My observation is that most of the HR are solo shots - not sure if Dave Wallace has taught them to throw strikes when the bases are empty and not fear solo home runs and change their approach when men are on base. One thing is obvious - All the O's pitchers have a lower ERA than their HR, K, and BB rate would suggest. Either the staff has been incredibly lucky collectively for 3 years or something else is going on. 4200 innings is not a small sample size any longer...

Joshua Josephs said...

I wonder if it is possible to get leverage statistics and see if the Orioles starters strike out rates improve in high leverage situations such as RISP. My gut watching the Orioles particularly Chris Tillman and Chen says yes. Second, I wonder if the strand rate has something to do with the fact that the Orioles have often played from ahead this year so teams dont try small ball to score a single run, but rather take the risk/reward of hitting a homer/extra base hit.

Joshua Josephs said...

A quick look at Orioles starters with runners on base yields an interesting stat, that I have to see if it holds for the whole league. But for Tillman, Chen, Norris, and Gonzalez, once there is a runner on base, the SO/W rate is vastly better than when bases are empty, suggesting they are able to really focus down in that situation. Also the Orioles strand numbers are probably lucky to an extent; Gonzalez has given up only one hit with the bases loaded this year.

Matt Kremnitzer said...

I could have worded that section better. Health and good defense have benefited the entire staff, certainly. In a future post, I may examine this topic further.

EastCoastMatt said...

1 reason for high strand rates has to be great fielding.