A couple of significant things have happened since I broke down the 2014 bullpen in early October: we’ve had a few roster changes, and we’re now armed with some 2014 projections (Steamer and Oliver).  Reviewing how the bullpen lines up breeds two new questions: whether the D-backs can (or should) enter the season with just Joe Thatcher from the left side, and who should close.

In the quick reaction piece on the Reed-Davidson trade, I did include a section on how the new bullpen stacks up (“Fitting into the Bullpen”).  But we can do better.

I’m still assuming a 7-man bullpen to enter the season, so just to recap what’s been already covered: Josh Collmenter is the long guy.  Joe Thatcher will be expected to handle left-on-left matchups.  Brad Ziegler will return to putting out fires.  David Hernandez will be given the opportunity to prove he can be the dominant set-up man he was in 2011 and 2012.  J.J. Putz and new acquisition Addison Reed will pitch at the end of games, and while Will Harris may be used a little differently, he’s got a spot, too.

Barring a significant move or a spring injury, then, we know who the seven guys are.  This is what Steamer expects of them (courtesy of FanGraphs):

Bullpen 2014 Steamer

And, for purposes of quick comparison, the same pitchers’ stats from last year:

Bullpen 2013 Stats


A few things jump out quickly.  It’s extremely unlikely that Collmenter will pitch as few as 40 innings next year; repeating in the 92 IP range is much more likely.  Thatcher is high on innings just because of how he’s probably going to be used, and unless you’re much more of an optimist than me, you’ll probably agree that Putz is probably high for health reasons.

I’ll not re-analyze Addison Reed here (see section from Monday’s post, “Addison Reed: Damned Good Pitcher), but it’s interesting that unlike the other six guys on the list, his FIP was quite a bit better than his ERA.  That suggests some bad luck.  It also suggests that he was rarely used mid-inning, and looking at his game log, that seems right: there is only one appearance before September in which he pitched something other than an even number of innings, and the three 0.2 IP appearances in Sept. were all blown saves.

Why would it matter that he started his own innings?  When a reliever comes into the game with runners on base, there are more out opportunities.  It’s impossible to get a double play without runners on, but it can also be easier to tag or force a runner moving toward other bases (there are pickoffs and caught stealings, as well).  All without any potential ERA harm from those runners.  It’s why guys who are brought in mid-inning more frequently tend to have ERAs better than FIPs.  Brad Ziegler is the most extreme example above, and he was the guy with runners on — unless a strikeout was needed (in other words, a man on third with less than two outs).

By the way, totaling up these WARs is probably not that helpful, either.  FIP does a fine job on balls in play, but the vagaries of small samples still matter on walks and strikeouts.  We’re also unlikely to get more than 75% of the total bullpen innings from this crew of seven.  The D-backs got 518.2 IP from relievers, 101 IP of which came from guys other than the seven with the most innings (and some of those seven weren’t in the top seven to open the season).  Last year the D-backs got 2.1 WAR from their relievers, overall.

Dumping the stats was intended to make the point that even without any Craig Kimbrels or Koji Ueharas, the Arizona bullpen is looking pretty good for next year (and also for the Putz point below).  It was also to make the point that, since the only lefty in the bunch is unlikely to get cut, there are no more obvious candidates for removal.  It’s not a huge surprise to hear Jose Valverde rumors; Marcos Mateo may be the team’s primary insurance against spring injury, but if he doesn’t open the season with the team, he’s no longer an option.  Valverde would be insurance for May or June; after that, the primary insurance policies are prospects like Matt Stites or Jake Barrett.

Where Have All the Lefty Cowboys Gone?

For our Offseason Plan, one focus for Jeff and I was the anticipated need for a second lefty.  Not just any lefty — if having major league stuff and throwing with that hand were the only criteria, we (and the club) would not have written off Tony Sipp.  But a lot of the lefties have come off the board, including the expensive free agents (like J.P. Howell), some bargain guys (Manny Parra), and some trade targets (Tim Collins is no longer available in the wake of the Will Smith trade).

There aren’t a whole lot of candidates anymore, and maybe none who are promising enough to warrant bumping one of the seven guys addressed above.  Oliver Perez might be the best of the bunch, but he would not come cheap.  Mike Gonzalez, Eric O’Flaherty, Scott Downs, Jose Mijares might be the best remaining guys.  Maybe the D-backs can sign Rich Hill and his miraculous curveball on a minor league deal, but I the free agency cupboard is bare.  The trading game of musical chairs is almost complete, as well, with guys like Ian Krol and Jerry Blevins already moved.  I still like Brett Cecil, but it’s hard to justify selling assets to get a lefty reliever when the advantage over the current six RHP is marginal (even against LHB).

We should get used to the idea of having only Thatcher, then.  At least for a while.  The leading candidate for the end of the season is probably Andrew Chafin, who is still working in the minors as a starter; come late July, the D-backs could hand Chafin a relief role if needed without damaging his chances of pitching a full season as a starter in 2015.

In the meantime, I’m looking at Will Harris, as I intimated on Monday.  Harris had a very good year before falling off in September (7 ER in 11.2 IP).  Unlike the other six guys slated for the Opening Day bullpen, the team could option him midseason if need be; I’d guess it would take more than one terrible month, or more than two mediocre months, for him to run out of rope.  And although it might not help Harris’s career, the team might be well served to line him up to face lefties.

Harris was better against LHB in 2013.  The 84 LHB who faced him hit .215, got on base at a .262 clip, and slugged .247.  The 133 RHB who faced him went .273/.338/.420.  And to compare that to our resident lefty, Joe Thatcher’s 97 LHB opponents in 2013 went .241/.323/.310 (that SLG mark is great, but jeepers, that OBP is alarming).

Small sample size?  Most definitely.  But there are two good reasons to take some meaning from a small sample, in my view.  The first is if the phenomenon is so extreme that it’s less likely to be explained by just luck; and a .509 OPS is pretty damned extreme, in my view.

The second reason is if there’s some kind of plausible explanation.  Here, I think we have one: Will Harris has a pretty damned good cutter, which he used to great effect against lefties in 2013.  Against RHB, Harris went cutter 70.4% of the time, and curve 29.6%; against LHB, Harris racheted up the cutter percentage to 78.8% of his pitches.  That’s an agenda, and you can imagine why.  You can also imagine why, if everything is tailing away from RHB, it might be hard for Harris to jam them inside.  (Note: MLBAM considers the cutter a 4-seam fastball, and I think Harris does too — I’m going with brooksbaseball.net, and the point is the same either way.)

Maybe the biggest reason to trust Harris against LHB early in the season is because if his effectiveness against them in 2013 was a mirage… he might not be nearly as good as his 2013 performance would suggest.  Might as well find out, especially since any replacement for Harris that the club might like would either not be ready early in the season, or would cost a fair bit more than him.

Putz Over Reed as Closer

Those Steamer projections above sure do make Putz and Reed look like pitchers of similar quality for 2014.  Maybe they are, maybe they aren’t.  Maybe Reed only has his late-2013 velocity, and his 2014 looks a little more like his 7ER-in-8IP September than the rest of his 2013 season.  Since they’re close, we need not prefer one to the other based on projected performance — just health, and intangibles like that Proven Closer mentality and experience in the ninth.

To risk looking even more ambivalent, I also don’t really care to argue that the two guys be deployed as matchups allow.  I haven’t done the research there on what kind of difference that makes — but I’d at least meekly suggest that if one of the two guys is fresh and the other guy has pitched 3 of 4 previous days… maybe pitch the fresh guy.  But let’s look at a few pros and cons for using Putz as closer over Reed.


If the Trumbo trade was motivated primarily for entertainment- or perceived-value reasons, it’d seem odd to trade for Reed and then not start to put him on program covers and Chase Field banners.  I can’t really discount that.  More revenue helps the team win games, too, and if there isn’t an appreciable difference, why not have Reed close over Putz?  I have to call this a marginal reason to do so, however.

A slightly more compelling reason: Putz is much more likely to get hurt.  If you value having a single guy carry the torch all season, then Reed is probably your man.  I’m not in a position to quantify that kind of thing — which is to say, I’m not putting much weight on it, although I can’t dismiss it outright.


Addison Reed had 40 saves last year, and Putz lost his job as closer.  Easy decision?  I don’t know.  If there’s an x-factor that makes some guys perform better when the chips are down, I have no evidence about it either way.  I’m slightly more inclined to think that there are some guys who perform worse in save situations (paging Octavio Dotel) — that just makes more sense to me.  But it’s not like Putz caves under pressure.  In terms of closing experience, the only reason to prefer Reed over Putz is if you prefer recent closing experience.  We already established that their 2014 abilities might be on similar levels, though, so maybe overall experience matters more… and Putz has 189 career saves to Reed’s 69.  Just like with the two “cons,” I’m not putting a lot of stock in this either way.


Technically, I guess you could say that money is an intangible — but it’s a whole lot easier to quantify.  And it’s easier to analyze, too: saves factor into arbitration awards.  They can factor into market prices for free agents, too.  But we don’t give a damn what Arizona does in terms of lining Putz up for a 2015 payday… chances are he won’t be getting paid by the D-backs.  We certainly do give a damn what Reed will make, because the 2015 will be paid by the team, and the 2015 salary will be the platform from which Reed’s 2016 and 2017 salaries jump.

Counting stats factor heavily into the arbitration process, and barring a contract extension, Reed’s 2017 salary is fairly likely to outpace his actual value to the D-backs.  That math might catch up to the team as early as 2016, quite frankly.  Once a player becomes a non-tender candidate, control over that player is an asset that is essentially worthless (what could Towers have gotten for Tony Sipp in trade at the end of September?).

John Axford may not be the best parallel — he was non-tendered by St. Louis in lieu of a ballpark $5.7M arb award because he was not worth it, but the discrepancy between value and perceived value had a lot to do with all of the saves Axford had logged as the primary closer in Milwaukee for parts of three seasons.  He’d be a parallel if Reed got bumped from the closer role in mid-2015, maybe — in that event, the pitching juice would not be worth the arbitration squeeze.

More instructive, maybe, is the situation of the Cleveland Indians, who released closer Chris Perez rather than go to arbitration a third and final time (MLBTR’s Matt Swartz projected a $9M salary).  It may be that Addison Reed improves as a pitcher.  If he doesn’t, he’s not that different a guy from Perez, and I’d guess we’d face a similar situation.  I’d guess it was something of a judgment call for Cleveland to tender Perez a contract for 2013 (he made $7.3M), considering he almost certainly would have made less on an annual basis in free agency; it may be that hopes of trading Perez tipped the balance in favor of a tender.  Perez did get shopped, by the way — no takers.

I agree with Jeff that the Reed trade may have been better than it looked at first blush, but the value-for-value aspect of the trade looks worse if Addison Reed has little to no value in the last one or two years of team control.  Maybe it’s inevitable that Reed gets traded or non-tendered before his final arb year, 2017.  Or, who knows, maybe he’d be worth the salary he might command at that time through the arbitration process ($12M?).  Looking ahead, though, I’m concerned that his 2016 value might be outpaced by his 2016 salary.  Since Putz is around, and since there aren’t strong, substantive reasons to have Reed close in 2014 instead, I think it’s worth limiting his counting stats.  It’s worth trying to push back the sun in hopes of keeping away the rain for one extra season.

Tagged with:

13 Responses to Why RHP Will Harris Should Be the Second Lefty Reliever, and Why J.J. Putz Should Close

  1. Jeff Wiser says:

    Great recap. Only thing I’d throw in: Oliver Perez has really weird reverse platoon splits over his recent career as a reliever, so although he’s a lefty, he’s not an ideal match up against left-handed batters.

    Otherwise, carry on…

  2. Paul says:

    One glaring omission from the article is Matt Reynolds (LHP), I would suspect that he would be in the bullpen ahead of Will Harris. This, however, would make the story about the second lefty in the bullpen not necessary. Otherwise good article.

    • Ryan P. Morrison says:

      Had Reynolds in the section with Chafin, then left him out. Reynolds had Tommy John on Sept 24 of this year. In the first draft, I had him a potential late-season guy. Even though relievers generally return from TJ a little quicker… the quickest I can remember is 10 months. Given the team’s recent experience with Hudson… maybe a few innings in September?

  3. Tristan says:

    “More instructive, maybe, is the situation of the Cleveland Indians, who released closer Chris Perez rather than go to arbitration a third and final time ”

    The Indians released Perez not only because of his contract, but because he was awful the final months of the season. Why pay a middle reliever that much money? He was no longer a closer. Also the reason he wasn’t traded.

    Also, on Putz, it is worth noting that most of his innings later in the season that were scoreless, we not the 9th. He had less pressure during the year and posted decent numbers, but when he was the closer, his numbers were pedestrian.

    • Ryan P. Morrison says:

      No question, Perez was not great at the end of the season. Three particularly memorable meltdowns pulled down the Aug and Sept stats. And you’re clearly right that performance had something to do with it — if he was Kimbrel, $9M doesn’t seem so bad, and he’d probably have been kept. Like how global warming is exacerbated by both more greenhouse gases AND deforestation, performance and expected salary were both causes of the release.

      But note that the arbitration award was supposed to compensate Perez for who he was — which included the bad last two months. So it’s really the fact that saves affect the arb process so much. If he were lined up for a near-market AAV based only on his other stats, you can bet he wouldn’t be projected for $9M.

      Your “no longer a closer” point is a very good one. It’s the same as with Axford. And if they’re the same pitchers, and can offer as much help, should they be “worth” less just because they’re not expected to pitch in the ninth?

      On Putz, I just don’t buy the pressure thing until someone convinces me otherwise. When he was closer, he wasn’t throwing his slider, and he wasn’t very effective. When he came back midseason, he didn’t get the closer job back, but he was throwing his slider. Putz thought it was the slider, and that makes a whole lot of sense to me. It’s a good pitch.

  4. Sam says:

    The analysis of Harris versus Thatcher seems a little lacking to me. Last year, wouldn’t Gibson have been more likely to use Thatcher against stronger left-handed batters, and left Harris in there against the weaker ones? Thus, their stats might be a bit skewed by managerial strategy. Is there another stat that compares hitters’ general performance with their performance against that pitcher?

    • Ryan P. Morrison says:

      You raise a very good point — I did not control for the level of competition Harris faced. A few things, though:

      I’m not sure Harris was ever used to match up with LHB in 2013. Switch hitters make a general split difficult, but just using a couple of other relievers as examples, 39% of batters faced by Josh Collmenter were LHB (151 of 384) and David Hernandez faced LHB 48% of the time (125 of 263). Will Harris faced them 39% of the time, so I’m not sure he was ever deployed against LHB on purpose.

      Harris only faced 84 LHB, so we’re dealing with a really small sample. Since you’d expect a platoon split in general, though, that discrepancy between a .224 wOBA (vs. LHB) and a .333 wOBA (vs. RHB) is really quite remarkable.

      You’re right — the success Harris seemed to enjoy against lefties in 2013 could seem to soften if we somehow provided for strength of competition. Still, I’m not suggesting that Harris get used as another one-and-done lefty next season — I’m just suggesting that Arizona could get by with just one lefty, if it worked to neutralize some other LHB with Harris. Maybe, a greater percentage of the time, Harris could start innings when two of the first three expected hitters were LHB. Or, he could be brought in mid-inning with two outs to face a LHB, especially if he could stay in the next inning and face another one.

      I was just spitballing, really. With the possible (but not probable) exception of David Hernandez, the D-backs now have six RHP relievers who look like they could be pretty damned good. Just seems like a shame to bump one in favor of a second lefty, when there are mix and match options with Harris, and when Chafin and Reynolds could be the guys in 2015 anyway.

      I’m not aware of a stat like that, and after thinking about it, I see what you’re saying. If we were to use a stat like that to gauge performance against Harris against performance by the same hitters against other pitchers… maybe you’d want the “other pitchers” to be the same handedness. I’m not sure what difficulty would be posed by facing some hitters many times, either. I’ll ask around. The closest thing I know of is a paid service that compares a hitter’s stats to groupings of pitchers with similar profiles (because a hitter’s career stats versus a single pitcher are worse than useless, generally).

  5. […] it as a long term issue, and buying a second guy like Thatcher seems foolish. As I wrote when I broke down the composition of the bullpen, I’m perfectly happy for the team to treat Will Harris as the second pen lefty, which is […]

  6. […] been a while since I argued that Will Harris should be treated as the second lefty in the bullpen, someone who could pitch a whole inning with multiple left-handed batters due up while Joe Thatcher […]

  7. […] risky and I’ve never been a fan of bullpens with only one lefty. There’s a chance that Will Harris will maintain his reverse platoon splits but he has a short major league track record and I don’t want to pin my hopes on reverse […]

  8. […] At AZ Central, Zach Buchanan covered the Reed-Davidson trade as well.  Interesting comments from Reed himself, who clearly wants to be the closer.  As I wrote last week, it’d be Putz at closer if the decision were up to me. […]

  9. […] the alternatives (although there is also the financial question, which was part of the reason my favored closer was J.J. Putz, as Jeff also […]

Leave a Reply

Your email address will not be published.