Reliever Market Changing – Maybe For The Better

One recently signed free agent deal just blew me away. “What were they thinking?”, I said to myself. “That is just a ridiculous amount of money for a guy like that”, “they’ll raise the bar all around!”. I wasn’t thinking of Zack Greinke‘s big deal, or Josh Hamilton‘s or Anibal Sanchez‘ – I was thinking of the deal that the Philadelphia Phillies gave Mike Adams.

August 11, 2012; Arlington, TX, USA; Texas Rangers relief pitcher Mike Adams (37) pitches to the Detroit Tigers at Rangers Ballpark in Arlington. The Rangers won beating the Tigers 2-1. Mandatory Credit: Jim Cowsert-USA TODAY Sports

Mike Adams is a reliever. Specifically, he has been a setup man and a very effective one at that. The Tigers caught no small amount of flak for the contracts that they gave Joaquin Benoit following the 2010 season and Octavio Dotel following the 2011 offseason – these were two other guys considered to be among the very best setup men in baseball. This one raises the bar a tad over those (and remember that most people thought the Tigers’ generosity ridiculous, especially when it came to Benoit). Adams will be getting $6 million per year for two years with a vesting option for a $6.5 million third year. For all intents and purposes, this is a 3 year, $18.5 million deal if he stays healthy and a 2 year $12 million deal if he doesn’t.

The thing is: he isn’t being asked to go to Philly and close games – he’s being asked to continue to pitch the 8th. What seems to be going on is that the gap between what teams are willing to pay closers and what they are willing to pay non-closers is narrowing. Joel Hanrahan and Chris Perez will be paid in the neighborhood of $7 million – but their own teams consider them overpaid and they look to be difficult to move because of that salary. There is no market for Rafael Soriano whatsoever right now, though he is the best reliever available (probably in trade or in free agency). It isn’t just that he wants more money than anyone is currently willing to offer a closer (or other reliever) it’s that he would cost the signing team it’s first round pick. [Personally, I’m not sure that Soriano will get any contract offer this year – unless it’s a tail-between-his-legs one year deal with the Yankees. Even if he’d sign for 2 years and $14 million, no one wants to give up that pick.]

To a stat-head all this makes perfect sense (though it, no-doubt, makes no sense at all to Jim Leyland) so maybe what we’re seeing is just the continuing stat-headization of MLB front offices. Pitcher talent is “controlled” by the pitcher, and should therefore be paid it’s market value. Leverage is “controlled” by the manager, not the pitcher, and should therefore not impact salary offers. If (value-wise) the only thing that separates Soriano-closer from Soriano-setup-man is when you choose to put him in the game, why should one be able to demand more money than the other? For years, teams have paid for saves by making the implicit assumption that non-closers don’t have the mental toughness to close – in the same way that third basemen don’t have the agility to play shortstop.

In a way, it makes sense that a “closer” would get paid a lot of money simply because of talent. The value that any reliever has is dependent upon leverage, but you have to find somebody good to pitch in those high leverage situations and value is amplified by that leverage. The additional value that the closer gives you because of the “amplification” is real, and should – in a sense – make the very, very best closers disproportionately more valuable than the rest. What it shouldn’t do is create a widening gulf between non-elite closers and elite non-closers. But… it did. The arbitration process has reinforced this, by comparing guys that save games to other guys that save games for the purposes of salary determination. And though this process has gone into reverse in free agency, it’s still going on in arbitration and will continue to for at least a few more years no matter what happens in free agency.

The past 3 years, Joaquin Benoit has a 2.71 ERA, a 0.989 WHIP and 10.4 K/9. That’s pretty darn good. Mike Adams has a 2.06 ERA, a 1.048 WHIP and 9.0 K/9. Though a small part of that could be due to the offense killing environment in San Diego, that is still pretty darn good. Lefty Sean Burnett, who signed a 2 year, $8 million deal with the Angels a few weeks back has a 2.76 ERA, a 1.231 WHIP and 7.8 K/9. They haven’t accumulated many saves, though. Joel Hanrahan has a 2.73 ERA, 1.172 WHIP and 10.4 K/9. Chris Perez 2.84, 1.137 and 7.9. Soriano 2.50, 1.063 and 8.6. There isn’t really any talent gap between these to groups, thought there is a gap in saves and (still – despite the “outlandish” contracts) in pay.

The kicker here is that if you determine a guy’s value as “wins” that he adds to the team and determine your own willingness to pay him accordingly, if you don’t use him in high leverage situations he isn’t going to add much in value no matter how good he is when he pitches. A “garbage man” can go a whole season without allowing a hit and his contribution to the team will still be trivial. The question is whether a team has enough leverage to go around to make it worthwhile to pay two semi-elite relievers $6 million apiece. And for that… how about some statistical tests?

The best stat available for situational usage for a reliever is “gmLI”, which is a measure of leverage (how important what is about to happen is for the ultimate outcome of the game) calculated when the reliever enters the game, as opposed to re-calculated for every pitch thrown, etc… Among “qualified” relievers over the past 3 years (values for non-qualified relievers tend to be low, these are guys that weren’t trusted enough or healthy enough to be used as regularly) values range from 2.11 for 2011 Jordan Walton (who was closing for the Angels) all the way down to 0.38 for Rhiner Cruz of the Astros in 2012 (whose 6.05 ERA certainly didn’t warrant his inclusion when the game was on the line, even if Houston didn’t have any better options for 7th man in the ‘pen).

It is rare for non-closers to come anywhere near the top in gmLI. Of the top 25 (over the span 2010-2012) only 2 were non-closers (Daniel Bard in 2010 and Scott Downs in 2011). Out of 26-50, 7 were non-closers and out of 51-75 only 6. Closers typically have gmLI scores of 1.50 or more (though not all – 2010 Jose Valverde had only 1.27) while trusted setup men tend to be between 1.10 and 1.40. When a team has both an established closer (not a revolving door) and a trusted setup man, there have been cases where the setup man wound up with a higher gmLI than the closer (like the 2010 Tigers with Ryan Perry (setup) and Jose Valverde (closing) or the 2012 Indians with Vinnie Pesano (setup) and Chris Perez (closing)) this is pretty rare. A more typical case would be the 2010 Tampa Bay Rays, where Joaquin Benoit and Rafael Soriano were almost equally dominant. But… Soriano (who closed) had a gmLI of 1.90 and added 4.09 wins by Win Percentage Added. Benoit (who did not close) had a gmLI of 1.27 and added 1.89 wins by WPA.

Benoit – one of the best and most trusted setup men in the game from 2010-2012 – had gmLI numbers of 1.48 and 1.33 in 2012, higher than in 2010, but it is still a decent crude estimate to say that the contribution that the same guy is able to make as a closer (provided he has talent) is roughly double what the same guy can do in a setup role. Jose Valverde contributed 6.03 wins by WPA from 2010-2012 despite one of the lowest gmLI scores for a full-time closer while Joaquin Benoit contributed only 4.72 wins despite one of the highest gmLI scores for a full-time setup man. This is true despite the fact that Benoit was unquestionably the better pitcher over that span, with a lower ERA, a lower WHIP, fewer walks and more strikeouts. Teams pay – it is reckoned – somewhere between 4 to 5 million dollars for an improvement of one “marginal” win. That would suggest that Jose Valverde was “worth” about $2 million per year more than was Benoit, because he closed. Of course, Benoit himself might have been worth $6 million more per season had he been in a closer role.

In a sense, the second best relief pitcher on any team sees his value drop by roughly half simply because he isn’t the best. As far as whether Mike Adams, Joaquin Benoit and others like them are worth the money they are being paid… yes they probably are. It’s harder to make this case using numbers when we’re looking at middling relievers, since adjusting a leverage-based stat like WPA for “replacement level” is basically impossible but just the wins that Benoit adds over a zero-WPA (above replacement level) baseline have been worth between $6 to $7.5 million per season. The second kicker is that while leverage amplifies your positive contributions when you pitch well it also amplifies your negative contributions when you pitch badly. As a result, a mediocre closer (who contributes virtually nothing in terms of WPA) is much farther from a good closer even if the mediocre closer is just a good closer having a bad year like Francisco Cordero in 2010. A terrible closer can do untold damage to a team even if he doesn’t hold onto the role – like Heath Bell for the Marlins in 2012 with his -3.03 WPA.

Given that uncertainty is the name of the game in major league bullpens, it probably doesn’t make sense to pay big bucks (and here I mean Soriano contract big bucks) and expect exceptional performance (rather than decent performance) from more than a handful of relief pitchers. Chris Perez was solid if unexceptional as a closer for the Indians last year: added 0.25 wins. John Axford cost the Brewers 1.59 wins despite pretty filthy stuff. It only makes sense to pay vast sums of money to a great closer if he is a mortal lock to remain far above the field. Like Mariano Rivera, who has accumulated an amazing 54.65 career WPA. [Hall of Fame reliever Goose Gossage accumulated only 31.4 and had 7 seasons with a negative WPA, Rivera hasn’t had one since his rookie year.] In summation – I’d say that guys like Joaquin Benoit and Mike Adams are – definitely – worth those $5.5 or $6 million per year contracts, even if used in setup roles. I’d also say that the reason that they’re worth it is because they’re such damned good relief pitchers… and that begs the question: why shouldn’t they close?

And before we get started on that – I do understand the very good reasons why you wouldn’t want to throw Benoit into the closer role right now (way, way too many home runs in the second half of 2012). But I do wonder why nobody wanted him to close in 2011…