April 16, 2010
Checking the Numbers
Braves Bullpen Makeover
As spring training came roaring to a close, it was impossible to discuss the Braves without mentioning Jason Heyward’s emergence or the team’s final season with Bobby Cox as manager. Earlier in the spring, however, conversations about the Braves tended to focus on the acquisitions of Billy Wagner and Takashi Saito and the departures of Mike Gonzalez and Rafael Soriano. In what felt like one fell swoop, general manager Frank Wren essentially swapped late-inning relievers. The four relievers were similar, having experience closing games out and spending their faire share of time on the disabled list, with both Wagner and Gonzalez undergoing Tommy John surgeries in the last four years.
Another commonality exists in that each pitcher represents another aspect of risk, a topic I began to cover at the beginning of the year. Previously, discussions of risk dealt with performance and how pitchers were generally binned based on their consistency or volatility, terms easy to personify through Jon Garland and Joel Pineiro, respectively. The main conclusion was that performance volatility for individuals—different from a volatile individual like, say, Paula Abdul—meshed well with performance volatility for teams. In other words, Pineiro would have been a solid fit for the Mets, who could go 74-88 or 88-74, a wide range of possible outcomes. The high end of his performance distribution could put them over the top while the risk that he comes closer to the lower end would not be crippling to their chances. Along similar lines, a pitcher like Garland is better suited for the Cardinals, who appeared like locks to win their division and would be more in search of a guaranteed level of performance, even if it falls below the high end of a Pineiro-type.
When discussing Wagner, Saito, Gonzalez, and Soriano, the risks are a bit different for two reasons:
1) They are relief pitchers, a group already known for their statistical fluctuations;
2) All four have injury histories, so there is no guarantee they will last the entire season.
The first point deals more with the inherent statistical noise in their performance lines, as 220 innings for a starter isn’t really a reliable sample, let alone the 52 1/3 innings a closer might amass over the course of a season. And by the time relievers get to the point that analysts can start to become confident in what they bring to the table, they are usually in their mid-30s, or nearing 40 like both Wagner and Saito, when the declines associated with aging mitigate any semblance of confidence in their abilities.
The health risk is more obvious: Wagner could whiff 15 batters in five frames, but if he misses the rest of the season after those innings, he isn't really helping the club. Free agent signings aren’t, or at least shouldn’t be, as cut and dried as deciding a player is good and offering him a contract. These risk factors come into play quite a bit when deciding how to invest a pretty penny, as well they should; while everyone would love a tremendous, consistent and durable pitcher, this archetype does not simply fall off of trees. Pitchers fitting this bill are usually referred to as Halladay, Sabathia or Oswalt, or Rivera for relievers. Since those players are not easily reproducible, personnel decisions are scrutinized from a number of different angles, but before analyzing the swap of late-inning relievers for the Braves under the angle of risk, let’s first revisit what makes relievers so frustrating in the first place.
Relief pitchers are inherently risky because they are much tougher to evaluate. In fact, aside from a fielding metric for catchers, valuing relievers statistically is probably the most controversial or heated topic in this field. How much can you know about a pitcher after 45 innings, especially if he is used in a specific role that could differ from how he might be employed by another team? Even those durable relievers who consistently throw 65-80 innings per season are not mortal locks to come anywhere close to repeating performance from year to year. From a mathematical standpoint, the year-to-year correlation of park-adjusted ERA for relievers with 40 or more innings pitched in consecutive years is very low and signals a lack of reliability in reliever performance. From a qualitative standpoint, find me a fan comfortable with the way his team’s bullpen is constructed and I’ll find you someone whose pants are on fire.
The best example is Brad Lidge, who certainly was not as good as his sub-2.00 ERA in 2008 suggested, but did anyone really think he was going to post a 132.35 ERA last season? Okay, his ERA wasn’t that high, but it sure felt like it. A lot can happen in a very short span of time to muddy the statistical waters, but relievers don’t pitch much more than a very short span of time. Additionally, a reliever can make 10 appearances, perform admirably in nine of them, but stink up the joint in his last game and walk away with what would be considered pedestrian or unimpressive numbers. Making matters worse is how some relievers are used in the specialist capacity, facing a mere batter or two per game for what equates to maybe half of a standard season of relief.
With that in mind, how do we evaluate Wren’s decision to bring aboard Wagner and Saito while waving goodbye to Soriano and Gonzalez, given that it’s fairly tough to have any real clue about the true talent level of relievers? Well, it is clear that the Braves were going to be cost-conscious in the offseason, having unloaded Javier Vazquez to the Yankees in a five-player deal for a return “highlighted” by a fourth outfielder/SHINO in Melky Cabrera and a prospect still far from the majors in Arodys Vizcaino. Additionally, Soriano’s decision to accept arbitration was reported to have irked the team, and they wasted no time cutting ties by trading him to to the Rays. Gonzalez is certainly a welcome addition to anyone’s bullpen—please, Orioles fans, it’s only been a week—but it became clear that he was going to command a fairly lucrative deal on the market, making it no surprise that the Braves let him walk.
But the decision to replace relievers carrying the additional risk of health with pitchers carrying that same form of risk serves as an interesting case study in filling out a roster. It seems that the Braves, realizing that they were not a lock for a playoff spot, decided that the two bullpen spots in question should go to pitchers with a better chance at putting them over the top. Just like the high-end Pineiro could push a hypothetical team with a wide range of win distributions into the playoffs, the Braves wanted pitchers in these two relief spots with the chance to help them surpass the Phillies. However, the rising costs of both Soriano and Gonzalez reduced their levels of attractiveness for the roles in question. Why pay $13.25 million per year for their services and be obligated to commit more than a year or two when guys like Wagner and Saito could be had for a combined $10 million with options controlling the commitment? For that to make sense the team would need to be convinced that the lame ducks were definitively superior to the incumbents; with relievers, there is virtually no such thing as certainty. The Braves decided to carry some risk but go for cheaper alternatives.
Only, from a pure qualitative standpoint, they did not really lose much in the way of performance, points in their favor for such a series of moves or non-moves. So if all four were injury risks and solid performers, why were Wagner and Saito cheaper than Gonzalez and Soriano? Aside from age—Wagner is 38, Saito is 40, Gonzalez is 32 and Soriano is 30—both Wagner and Saito returned from injuries last season and pitched in limited action. In Saito’s case, appearances were considered by many to have been “protected,” as in he did not appear in many real crucial situations, making his 2.43 ERA in 55 appearances less impressive. Wagner, meanwhile, split time between the Mets and Red Sox—the latter of whom also employed Saito—putting together a 1.72 ERA and 3.25 K/BB in 17 appearances. The age issues combined with the more recent injury problems to lessen the market value for both pitchers, but the Braves wisely realized that it was more cost-effective to take a chance on these players than to simply dole out big money and years to who they originally had in those spots.
How does the risk of health compare to that of performance volatility? In reality, these are merely different sides of the same coin, and the preference is in the eye of the decider. Some teams may decide that they want to get 15 great starts of Ben Sheets and fill the remaining 20 starts if and when it gets to the point he can no longer throw. Others might prefer a full season of Joel Pineiro, who will stay healthy, but might only look like a front of the rotation pitcher in a quarter of his games. The Braves, in making these decisions, likely had a ton of inside information on the health histories of all four pitchers and erred on the side health risk as opposed to the risk of performance volatility which, I suppose, would have been evident if they sprung for someone like Bob Howry. The Braves kept the aspect of risk intact in their bullpen, but cut costs, leaving them more wiggle room to help increase the high end of their win distribution while increasing the win variance as well; if they were considered to be a team capable of winning 77-86 games, perhaps now their range was 75-89 wins.
Regardless, this is all just another way of analyzing the same concept of risk, in that it is more beneficial for teams considered underdogs for post-season qualification to spring for riskier players with a large projected variance in possible outcomes. Having Sheets for 15 starts and Rodrigo Lopez for the remaining 20 probably is not too far off of good-Pineiro for 17 starts and bad-Pineiro for the remainder. Put yourselves in Wren’s shoes: which two of these four would you love to have in the bullpen and why? Are the costs of Soriano and Gonzalez—really just $3.25 million more than what Wagner and Saito eventually signed for that prohibitive? Which duo would instill more confidence in you as a fan?