Tuesday 25 March 2014

Making Lemonade Out of an Eight-Man Bullpen



It's seemed for a while now that the Blue Jays' search for a fifth starter will go right down to the wire, with Drew Hutchison likely having secured one of the two available spots, but with none of the other candidates for the rotation really having done much to separate themselves from the pack. It is also appearing as though, due to a number of pitchers' lack of options, the club may again start the season with an eight-man bullpen. Let me be clear. I don't like the eight-man bullpen. I think the Blue Jays are a better team with a fourth bench player than with an eighth reliever. That being said, I was watching Todd Redmond's start against the Tigers on Saturday and I started thinking that carrying an extra pitcher might be exactly what the Blue Jays need to squeeze every last shred of production out of that fifth starter's spot.

If you recall the game in question, Todd Redmond cruised through the first five innings without allowing a run. In the sixth inning he allowed a lead-off home run to number nine hitter Ezequiel Carrera. As the lineup returned to the top, Victor Martinez hit a double and was pinch run for by Hernan Perez. A fielding error allowed Steve Lombardozzi to reach, and advanced Perez to third. The next batter, a young upstart by the name of Miguel Cabrera, scored Perez on a sacrifice fly. Clean up hitter Don Kelly smacked a double that advanced Lombardozzi to third, and then Tyler Collins smacked another double that scored both Kelly and Lombardozzi. Redmond retired the next two batters to get out of the inning, having allowed four runs on six straight batters reaching base, albeit one on an error. He was replaced by Steve Delabar in the top of the seventh.

I'm not faulting anyone for this situation. It's spring training, so it makes all the sense in the world that they would allow Redmond to try and pitch his way out of trouble in the sixth inning. I'm sure that in a game that actually counts for anything there would have been a much quicker hook. In any case, drawing conclusions from one spring training start would be idiotic. Still, it got me thinking, and led me to tweet this:



That led to a bit of a back and forth with intelligent Blue Jays fan @AHume92 about the practicality of such a policy, and while I'm not sure if I convinced him, I'm pretty sure I convinced myself that there's a way that the Jays can at least make carrying an eighth reliever worthwhile. Again, I'm not saying an eight-man pen is a good move, but if we accept that we live in a world in which the 2014 Blue Jays might at least start the season with eight men in the bullpen, I'd rather see them leverage the situation as effectively as possible than just give everyone an extra day off here and there. Basically, if they're going to have an eight-man pen anyways, I'd love to see them set up a pretty strict piggyback situation in the number five spot in the rotation. Quite frankly, I'm not even all that concerned with who the particular pitchers are, because whoever they choose is likely to benefit greatly from never, ever seeing a lineup for the third time, regardless of how many innings he's pitched, pitches he's thrown, or even how well he's been doing that game.

If you're not familiar with it, my thinking is rooted in the "times through the order penalty" (henceforth referred to as TTOP), which basically dictates that pitchers perform worse the second time they face a lineup than the first time, and perform even worse the third time through. For pretty much everything you could possibly want to know about the TTOP check out this piece at Baseball Prospectus, by Marshall Lichtman, a.k.a. MGL, co-author of the sabermetrics bible The Book: Playing the Percentages in Baseball.

According to the research in The Book, which is based on data from 1999-2002 when the run scoring environment was more favourable to hitters than it is today, pitchers allowed an average opponent wOBA of .345 the first time through the lineup, .354 the second time through, and .362 by the time hitters got their third look. According to their research, the driving factor is batters gaining a familiarity advantage the more looks they get in the course of a game. Of particular note is their breakdown of the numbers by good pitchers (<.320 wOBA against) and bad pitchers (>.340 wOBA against), because let's face reality here. All the pitchers the Blue Jays are realistically considering handing that fifth spot to are pretty likely to fall into that "bad pitcher" bucket.  Basically, the TTOP effect is the same for "good" and "bad" pitchers, about 20 points of wOBA between the first and third times through the lineup. The thing with the "bad pitchers" though, is that while while "good pitchers" go from excellent to above average over the course of a game, the bad ones go from below average to replacement level scrubs. A team can live with Clayton Kershaw giving up 20 points of wOBA by the third time through because he's still an above average pitcher. Tack the same 20 points onto some lucky-to-have-a-job fifth starter and the picture gets a whole lot uglier that third time through.

Imagine if in a game like Saturday's, where Redmond has managed to get through the lineup twice with his 80-90 m.p.h. slop, rather than send him back out there to throw batting practice his third time through the order, they replace him with Jeremy Jeffress to blow the lineup away with his upper 90s gas for a turn, or possibly two. Again, it doesn't have to be Redmond and Jeffress, I just chose them because of the dramatic contrast between their stuff. Regardless of who the specific pitchers are, the starter benefits from only ever facing a lineup twice in a game, and whoever relieves for the third time through reaps the benefits of facing the whole lineup for the first time, when pitchers statistically perform the best. It also provides some redundancy to the fifth spot. If the guy slated to start the game gets shelled his first time through, well then maybe he gets pulled early and the second guy gets two turns through the lineup instead.

All of the candidates in the fifth starter mix would stand to benefit significantly from being put in the best possible situation to succeed, and I'm not sure that means expecting any of them to go out there and provide a full six innings of quality pitching. They're all capable of going multiple innings, but asking a bunch of marginal starters to turn a lineup over three times is asking for trouble. Happ has been lousy this spring, and his tendency for high pitch counts often prevents him from going deep anyways. Redmond and Rogers are similarly mediocre but may have the inside track due to their lack of options. I don't know about you, but I don't feel awesome about their ability to provide a season's worth of quality starts. McGowan is probably the only one of the lot with even league average upside, and given his injury history he might benefit from the shorter outings as much from a health perspective as the others would from a performance standpoint.

Really the only reason to ask a below average fifth starter to turn over the lineup more than twice is to spare the bullpen, but if we're assuming an eight-man pen I don't think there's much reason to worry about gassing them by turning the game over to them early. Two pitchers would simply be designated as the "co-starters" (or something), leaving the rest of the bullpen relatively unaffected unless both of them are completely ineffective, which is less likely than either one of them struggling individually. By the time the co-starters have combined to turn the lineup over three times they should be just about at a point in the game where they'd be turning to the pen anyways. Using the eighth reliever in a piggyback role shouldn't tax the bullpen any more heavily than if they only had seven in the pen, but would theoretically help two pitchers combine for a better "start" than one would. If they pitch well enough or have a big enough lead, the two could even conceivably combine for a complete game and give the bullpen the game off altogether.

For me, the biggest obstacle is probably the human element, by which I mean the egos involved in this unorthodox arrangement. Pitchers and managers are largely conditioned to think in terms of innings pitched and pitches thrown, and to value performance according to those measures. Asking them to mostly disregard inning and pitch counts and to think in terms of times through the lineup would be a big change, the significance of which can't be ignored when dealing with human personalities. Most pitchers hope to go out there for six strong innings and a hundred pitches or so. In this scenario, their goal would be to turn the lineup over twice and then pass the baton to the next guy, regardless of how well the first two times through the order have gone. Whoever starts the game would have to accept that no matter how well things have been going so far, a new pitcher has a better chance of getting through the third time than the starter does, simply because by that point the hitters have gotten too much information on the starter.

There's a bit of a flip side to that as well though. Sure, no starter wants to be told that he's not being trusted to get through the lineup a third time. That's a pretty understandable blow to his ego and his perception of himself as a starting pitcher. However, his numbers would probably end up being better than if he ever went through a lineup a third time which might inflate his value in the eyes of some prospective future employers. Also, it might give some purpose to the eighth man in the pen. We're always hearing about how relievers like having defined roles, and "co-starter" (or whatever you want to call it) seems like much more of a defined role than "the extra guy kicking around the bullpen who might not even be there if not for the fact that he's out of options". There's no question it would require some serious buy-in from the pitchers, given that they'd asked to accept fairly unorthodox roles, but it would be a distinct role nonetheless. Why not give the eighth reliever a bit of dignity by assigning him a valuable role?

From the manager's perspective, when something goes wrong it's much easier to defend a conventional approach than an unconventional one. Using Saturday's example again, it's much easier for Gibbons to explain letting a cruising Redmond go back out there to face the lineup a third time, even he gets smashed the way he did, than it would be had he put in a new pitcher who promptly allowed the same four runs, even if the numbers say that a new pitcher of similar quality was likely to outperform Redmond in that situation. You can bet that there would be all sorts of uncomfortable questions from the media the first time the piggyback arrangement blew up in their faces, and I'm not so sure ol' Gibby's job security is such that he'll be eager to stick his neck out with a bit of creative bullpen management. That being said, managers are constantly having to explain bullpen moves gone wrong. How often do you hear a manager say something something like "Well the numbers say you want a lefty in that situation, it just didn't work out this time." Is it really that big a stretch to imagine them saying "Well the numbers say you don't want to let professional hitters see a crappy pitcher three times in one game, it just didn't work out this time"? They'd presumably find a more diplomatic way of saying it, and I'm sure it's harder to imagine than I'm letting on, seeing as how the platoon advantage is far more accepted in baseball orthodoxy than the times through the order penalty. Still, the concept is at least intuitive enough that, if it were clearly explained, I feel like all but the dumbest of fans would be able to understand the rationale and perhaps even appreciate the attempt to wring maximum production out of what's shaping up as a pretty dismal fifth starter situation. 

So that's my proposal. As long as the Blue Jays are going to run with an eight-man bullpen (still not a done deal), I'd like to see them designate two guys as co-starters. The combination of two pitchers should be able to get through the order that third (and perhaps fourth) time more effectively than one would, at least given the quality of the candidates the Jays are considering. They're kind of in the ideal situation to try something a bit wacky. Anthopoulos' asset management fetish means there's a distinct possibility that they'll at least start the season with an eight-man bullpen anyways, and probably one very shaky spot in the rotation. They've got plenty of guys that can give them multiple innings, but who you probably don't want to be counting on to go six strong every fifth day. If they're carrying an extra arm anyways, why not try and Frankenstein two of their fringe starters into one competent one by making sure that neither one of them ever sees a lineup more than twice?

Would it be worth sacrificing the additional offense and flexibility that a fourth bench spot would provide? That's a question for someone far smarter than I, but personally I'm doubtful. I hope that if they do start the season with an eight-man pen the situation resolves itself into a more standard seven-man set-up sooner rather than later. At the very least, using a couple of pitchers in the way I'm proposing could be a way to get a regular, multi-inning look at some of the candidates for the fifth starter spot once the season is underway, or to let McGowan continue to build up his pitch count slowly. If someone takes the job and runs with it all the better. In the meantime, as the saying goes "if a marginal asset obsessed general manager gives you an eight-man bullpen, make lemonade".


Follow Blue Jays Blackboard on Twitter @BJaysBlackboard
Like Blue Jays Blackboard on Facebook   





No comments:

Post a Comment