- FanGraphs Baseball - http://www.fangraphs.com/blogs -

The Baltimore Orioles and Fleeting Bullpen Greatness

Posted By Jack Moore On February 14, 2013 @ 10:35 am In Daily Graphings | 30 Comments

Over-Under Day — the day the first sportsbooks release their win-loss over-under totals for all 30 MLB teams — is one of my favorite days of the lead-in to baseball season. I’m not much of a gambler — I stick to risking my money on fantasy sports, personally — but Las Vegas is as good a projection system as we have, and although the numbers here will likely be revised between now and Opening Day, they provide as good a barometer for current team strength as you’ll find anywhere.

The Orioles, unsurprisingly, have the biggest drop-off from last year’s win total to this year’s over-under — 93 wins in 2012 to just a 76.5 over-under for 2013. Should Baltimore perform to Vegas’s projection, it will be just another example of the fleeting greatness of particularly clutch units, like the Orioles’ 2012 bullpen.

The idea of fleeting greatness is nothing new, of course. For example, only three of last season’s eight playoff squads from 2011 reached the postseason in 2012 despite the expanded format. Specifically, though, Baltimore’s greatness came on the back of a great — and particularly great in the clutch — bullpen. This greatness has, in the past, been the toughest to cling to. Baltimore’s plus-13.86 WPA in 2012 was the highest mark a bullpen has posted since 1974 (the extent of our data); its plus-7.19 clutch score was the highest recorded as well.

Observe, the fates of teams with the next 25 best bullpen WPAs (all plus-8.50 or greater):

bestbullpens

The results match up well with research I did at Disciples of Uecker during the season looking at the 25 worst bullpens by WPA. Whereas those clubs typically gained 13 wins or so (67 wins to 80 wins) the next season, the best bullpens by WPA dropped 10.6 wins (from 96.2 to 85.6) the following year.

Part of the effect is the impossibility of repeating such incredible performances. The 2001 Mariners went from 116 wins to 93 the next season; the Big Red Machine dropped from 108 wins in 1975 to 102 in 1976 to 88 in 1977. The list goes on, and certainly applies to specific bullpens (or offenses, or rotations) as well.

Intuitively, specifically clutch performances tend to regress harder to the mean. But how much harder? As before, all data comes from the 1974 season onward:

Statistic Wins Next Year Wins Change Avg. Clutch
Position WAR 98.5 92.9 (5.6) (1.0)
Position WPA 102.1 92.8 (9.3) 2.3
Reliever WAR 87.4 82.8 (4.7) 1.1
Reliever WPA 96.2 85.6 (10.6) 3.8
Starter WAR 94.2 86.2 (8.0) (0.4)
Starter WPA 97.6 93.4 (4.3) 1.3

For both position players and relievers, teams in the top of 25 of the WAR leaderboard dropped about 40-60 percent as many games as teams in the top 25 of the WPA leaderboard. Unsurprisingly, there was a significant gap in the average clutch scores for both of these groups — 3.3 wins for position players, 2.7 wins for relievers. This matches up with the intuition — teams regress to the mean, but the combination of regression and the ever impermanent nature of clutch performance leads to a doubly hard fall.

For starters, things get a little dicey. The top 25 teams on the WAR leaderboard actually regressed less than the top 25 teams on the WPA leaderboard (and the same thing happens with RA9-Wins substituted for FIP-based WAR). I would surmise this is because starting pitchers typically deal with far fewer high-leverage situations than hitters and relievers, but it remains an interesting result perhaps worthy of deeper research.

The 2012 Orioles’ combination of the highest recorded bullpen WPA and bullpen clutch score is just one of many statistical indicators pointing more towards 76.5 wins than 93 in 2013. One can also point to their Pythagorean record (82-80), ZiPS projected WAR (33, per this post) and the fact they were outscored (511-504) in the first six innings of games, for example.

Nothing about Baltimore’s relievers suggests its clutch bullpen dominance will continue. Jim Johnson became just the second closer ever to record 10 fewer strikeouts than games saved (Danny Kolb, 2004) and neither his 2.49 ERA nor 3.25 FIP particularly stand out among closers. Pedro Strop posted a 2.44 ERA but struck out under three more batters per nine innings than he walked (which led to a 3.59 FIP). What the pair did last season in recording 73 shutdowns to just 16 meltdowns was phenomenal, but like most phenomena, it is unlikely to repeat.

Baltimore did little to improve their team in the offseason and will go into 2013 with a very similar roster to the one which finished off 2012. And like 2012, Baltimore will be facing long odds to contend, and likely long odds to even finish .500 unless bullpen lightning can strike twice.


Article printed from FanGraphs Baseball: http://www.fangraphs.com/blogs

URL to article: http://www.fangraphs.com/blogs/the-baltimore-orioles-and-fleeting-bullpen-greatness/

Copyright © 2009 FanGraphs Baseball. All rights reserved.