Jeff’s post last week about how bad the Diamondbacks have been with runners in scoring position got me thinking. He explained that there is no such thing as clutch hitting. I wholeheartedly agree. I wondered if good relievers could make a team’s pitching more clutch. My idea was that better relievers would hunker down, especially in close games, and it would be reflected in a team’s record in one-run games.

This idea is especially important for the D-backs, as their bullpen has been and continues to be, in a state of flux. Just last year the main guys were supposed to be J.J. Putz, David Hernandez, and Heath Bell. (I think I speak for D-backs fans when I say I do not miss the ‘experience’ that was Heath Bell.) None of these three men is currently part of the bullpen. Bell was traded away due to his ineffectiveness. Putz is losing a battle against Father Time, who is undefeated, and was DFA’d. Hernandez tore his UCL and should resurface in 2015. Now, the top three relievers are Addison Reed, Brad Ziegler, and Evan Marshall. More change is on the horizon as Matt Stites has been called up and he and Jake Barrett are expected to be integral parts of the bullpen in the future.

Assuming the bullpen improves, does this mean the D-backs will perform better in close games? Dave Studeman compiled much of the data on this topic, but the gist of it is that for the most part, teams’ W-L record in close games tracks a team’s overall record. However, teams can outperform their talent level in one-run games with small ball and good pitching.

It’s hard to say how much good pitching can actually help, but there is a positive correlation between Holds-Saves Percentage (holds plus saves divided by holds plus saves plus blown saves) and a team’s record in one-run games.

Let’s use the D-backs last few years as an example. In 2012, they were 15-27 in one-run games, and 81-81 overall. Their Holds-Save Percentage was 84.25, which was below the average of 87% that year. Only six teams were worse. The poor record in close games seems to be in line with the poor performances from the bullpen. In 2013, they were 34-21 in one-run games. Despite the stellar record, the hold-save percentage was a miserable 77.8. The league average was 86%, and only the lowly Astros were as bad as the D-backs’ bullpen. The wide discrepancy is confusing, but the team’s insane 17-8 record in extra-inning games helps explain it. The offense bailed out the bullpen. This year, they have gone 12-11, with a HS% of 85.29, which is right in line with the league-average of 86.15%.

Each pitcher’s success in high-leverage situations also seems instructive in W-L record in close games. Addison Reed has been quite bad in high-leverage situations this season, with batters putting up a .387 wOBA (his wOBA against in high leverage situations last year was .193). Ziegler has been relatively bad in these situations; batters have a wOBA of .332. They’re under .300 in other situations. Marshall has been superb in high leverage situations, holding batters to an OPS of .643, versus .833 and .831 in medium and low leverage situations. All of these numbers come from small sample sizes, though.

I do want to point out that the D-backs combined record in one-run games since 2012 is 61-59. In 2012 and 2013 they finished .500. So again, the real thing driving one-run record is probably the overall runs scored and runs allowed. But, as the D-backs bullpen begins to change shape and improve, look for a slightly better record in one-run games — and a slightly better record overall.

Tagged with:

6 Responses to Bullpens and One-Run Games

  1. Jeff Wiser says:

    This is fabulous! I agree that we get some wonky outputs in the numbers with the small samples, as you mentioned, but the three-year trend is essentially average. The ‘pen has been up and down over that time, but when it all shakes out, they’ve been simply “okay.” With Marshall, Stites and eventually Barrett up in the majors, it should improve, as well as get cheaper.

    Another great point that you mentioned is how intertwined things like runs scored and runs allowed, especially by starters, impacts the work of relievers. To me, it’s critical to try to separate these components whenever possible so they can judged independently. For example, the ‘pen faced a ton of extra inning-games and blew some leads, but they didn’t have a lot to work with in the first place. Inheriting one-run leads is tough on any bullpen.

    So my conclusion is this: if you want to make the bullpen better, score more runs offensively and give up fewer of them with your starters. That will help alleviate this mess for everyone.

  2. Rod says:

    Thanks Jeff! Yeah I agree with everything you’re saying. Everything is so intertwined. It’s interesting how it all ends up being driven by the team’s talent level. You’d think a really good or bad bullpen would create different results.

    • Jeff Wiser says:

      For sure, it would seem to reason that a good or bad ‘pen could drive results, and I guess that’s true to a degree, but where it gets really difficult, and I know you know this, is separating out true skill and luck. When the samples are so small, and situations the ‘pen gets placed in are largely out of their hands, defining truly “good” and truly “poor” bullpens gets tough. I know the best brains in baseball have tore this apart, but I’ve never really been truly satisfied with any of the reasoning. So much noise in the data!

  3. Paulnh says:

    I definitely think our bullpen will be better in the coming years than it has been. Plus, it will be way cheaper. If KT is no longer in charge, I would not be surprised if the DBacks non-tender David Hernandez (if Towers is still our GM, that will never happen because David is too gritty and was actually acquired by KT). I know Hernandez has great stuff and really believe that he will be really good down the road, but he is just now entering his arbitration years and will be getting expensive very quickly. Plus, he is very expendable. The best depth in our organization is in the bullpen. We have several guys coming back off injury (Reynolds and Hudson), and even more guys in the minors. Obviously Stites, Barrett, and Marshall are guys that will be quality relievers, but I still really like Jimmy Sherfy. He strikes out a ton of guys, and has just been bitten by the long ball so far this year in Mobile. I also really like Eury de la Rosa. I was really impressed with his stuff in his short major league stint last year, but like Sherfy, gave up an unrealistic number of homers during that time (5 homers in 14.2 innings). I think that he could be our lefty in the pen down the road (he has also done really well in Reno so far this year with a 2.55 ERA and 33 k’s in 35.1 innings). I definitely think we need to trade some of our older bullpen arms at the deadline and go with a young, inexpressive group that I believe will be better anyway. How about a bulen next year consisting of Josh Collmenter (long guy), Daniel Hudson (closer), Matt Reynolds and Eury de la Rosa (lefties), Jake Barrett, Evan Marshall, and Matt Stites with Jimmy Sherfy in AAA? That could be the best and cheapest bullpen on baseball.

  4. […] then had Rod look at bullpens and one-run games. He noted that there is a correlation between holds and save percentage to team winning […]

  5. […] this site: We’ve tried to define “clutch” in terms of LOB%; We’ve looked at bullpens and one-run games; And we’ve also introduced the idea of using WPA/LI, to list a few. Our very own Ryan […]

Leave a Reply

Your email address will not be published.