Consider the bullpen of the Giants my latest fascination.
Why? Because the Giants’ bullpen has been great this season, right? I wrote about it yesterday, here. (That post has since been updated at the end.) Except, I think it’s possible that it’s been a bit more than just great. It’s been spectacular. In fact, it almost seems like the ‘pen is able to hold a 1-1 tie for 18 straight innings, allowing the Giants’ “offense” time to accidentally score a run with an error, a squeaker through the 4-3 hole and some aggressive base running. That’s the formula.
Boom! Giants win. How else do you go 62-3 in one-run games? (I think it’s actually 27-13, but you get the point.) Here’s how…
(Please go to the bottom of this post for the definitions of the Win Probability statisitics I’m using, which are plucked from the FanGraphs glossary. Find also the link to an article on Bruce Bochy and one-run games in his career.)
Basically what I’ve stumbled on is the historical “pace” of the Giants’ bullpen. I’m mainly looking at the relievers’ total Win Probability Added, Win Probability Added/Leverage Index (or how many wins are accumulated in a context neutral environment), and shutdowns per meltdowns, which isn’t an official FanGraphs statistic but one I’ve adopted myself.
The Giants relievers have a WPA of 7.00 with 60 games to go. Which means they are on pace to post a WPA of 11.12. They’ve also already posted a WPA/LI of 4.95 and are on pace to hit 7.86. Finally, they are averaging 3.05 shutdowns for every meltdown (113 SD and just 37 MD). That’s quite high.
Compare that to last year’s best bullpen, the Padres, which had a WPA of 8.42 in 162 games. Those Padres also had a SD/MD of 2.22 and a WPA/LI of 5.79. As it turns out, that’s nothing particularly special.
So I basically set the WPA at 10 or above, the WPA/LI at 7 or above, and the SD/MD at 2.75 or above, and I did a test to flag those teams that hit those benchmarks, going all the way back to 1992 (for the strike-shortened season of 1994, I projected all the figures to 162 games).
For WPA, only nine teams between 1992-2010 had a WPA of 10 or better. They were the 1995 Indians (10.73), 1997 Orioles (10.22), 2001 Yankees and Mariners (10.47 and 10.23), 2003 Dodgers (10.59), 2002 Braves and Twins (11.22 and 10.10), 2004 Rangers (10.04) and the 2006 Twins (10.36). The Giants are on pace for 11.12, having already hit 7.0 WPA.
For WPA/LI – I’ve found the averages for these bounce around a lot, over the years — only four teams have had a figure of seven or better. They were the 2006 Twins (7.9), 2004 Cardinals (8.18), 2003 Astros (7.02) and the 2001 Mariners (7.05). The Giants’ bullpen is on pace to post a 7.86 WPA/LI, having already hit 4.95.
And for the shutdown to meltdown ratio, only five teams have finished with a ratio of 2.75 or better. They were the 1992 Blue Jays (2.83), 2001 Yankees and Mariners (2.76 and 2.87), 2003 Dodgers (2.77) and the 2008 Angels (2.85). The Giants currently have a ratio of 3.05, while the Braves and Phillies have also done quite well in ’11 at 2.79 and 2.82.
The Giants would be the first to hit in all three categories at my established thresholds (again, 10+ WPA, 7+ WPA/LI and 2.75+ SD/MD) since the 2001 Mariners. Yes, those Mariners; the ones that won 116 games.
They won 116 games but their win-loss Pythagorean was just 109 wins. Another great ‘pen, the 2001 Bombers, won 95 with an 89-win Pythagorean. The 2003 Dodgers, and Mr. Eric Gagne? 85 wins with an 83-win Pythagorean. 2006 Twins: A really sexy set of relievers. They had 96 wins and a 93-win Pythagorean. So is there a precedent for a team with a dominant stable of relievers winning more (presumably close) ballgames than they should, thereby beating their Pythagorean win-loss record? You bet.
The 2006 Twins finished games with Joe Nathan (284 ERA+), Dennys Reyes (507 ERA+) and Pat Neshek (206 ERA+). The 2003 Dodgers saw dominant seasons from Eric Gagne (337 adjusted ERA or ERA+), Guillermo Mota (205 ERA+) – Hello, you! – and Paul Quantrill (232 ERA+). The 2001 Yankees had some guy named Mariano Rivera (192 ERA+) and one Mike Stanton (175 ERA+). How about the 2001 Mariners? They had Kazuhiro Sasaki (129 ERA+), who was pretty good with 45 saves. But they also had a young Arthur Rhodes with a 243 ERA+ and nasty Jeff Nelson (152 ERA+).
The 2011 Giants have really licked their Pythagorean, too, with 59 wins to the expected 53. They’ve been pretty great top to bottom bullpen-wise, so look at these flashy figures: Brian Wilson (129 ERA+), Jeremy Affeldt (120 ERA+), Ramon Ramirez (167 ERA+), Javier Lopez (175 ERA+), Sergio Romo (198 ERA+) and Santiago Casilla (200 ERA+). And, let’s not forget our old friend Mota, and his respectable 92 ERA+. There’s not a one unreliable face in there. Not one. There was, but his name was Dan Runzler and they sent him to Triple-A to learn to start in a terrible place called Fresno.
What makes this group special is the fact that they don’t really have a weak link. They have a weaker link, but not a weak link. Wilson hasn’t been as dominant this season, but he’s been good, and maybe his second half is better than the first few months. And whatever his flaws may be (too many base runners, primarily), he remains in an elite class in terms of supressing home runs. Don’t bother crediting AT&T Park, which is far too often used to “explain” the Giants’ real talent in this area, because Wilson has had a better (lower) HR/9 on the road (0.45) than at home (0.53) in his career. (Check out Wilson’s HR/9 graph below.)
Other than that, Bochy mixes, matches and gains the platoon advantage relatively often, using Lopez/Romo primarily for the eighth and Affeldt/Ramirez/Casilla primarily for the seventh. Most nights, each of them shut it down. One by one.
The final thing that I found interesting about this as that the Giants’ dominant bullpen innings haven’t been all that well-placed. The 2011 Phillies, for example, have a Clutch figure of 3.08, when the Giants have a figure of just 0.42. Many ‘pens of the past have juiced up their WPA with well-placed quality innings, but the Giants haven’t leaned on old Clutchy Clutcherson too much this season. They’ve mostly just been good, just about all the time. So when they do cough it up, it’s terribly surprising (and unsettling).
Will it continue? Can they continue to rack up the positive WPA? I don’t know, but damned if it hasn’t been one heck of an impressive 300 innings. As fellow staffer Otis Anderson pointed out to me, WPA is not a very predictive stat, as it depends on usage, situation and performance. So, as great as the bullpen has been and as talented as the group appear to be, it’s going to be tough to sustain their current pace with a good portion (around 40 percent) of the season to go.
If they do manage to keep their current, furious pace, they might end up producing one of the better bullpen seasons in the past 20 years, if not the very best. And if they should keep this current pace and pitch a little better in high-leverage situations (see: be more clutchier), then my head might explode.
Other than the bullpen, could there be another reason why the Giants have played so well in one-run games? How about Bruce Bochy? In this piece at Baseball Prospectus that coincidentally goes exceptionally well with what I’ve written here, Geoff Young writes, you decide.
- – - – - -
WPA (win probability added): WPA is the difference in win expectancy (WE) between the start of the play and the end of the play. That difference is then credited/debited to the batter and the pitcher. Over the course of the season, each players’ WPA for individual plays is added up to get his season total WPA.
WPA/LI (context neutral wins / game state linear weights): How many wins a player contributes to his team with the Leverage Index aspect removed, invented by Tom Tango.
Clutch: A measurement of how much better or worse a player does in high leverage situations than he would have done in a context neutral environment.
A Shutdown is when a reliever accumulates greater than or equal to 0.06 WPA in any individual game.
A Meltdown is when a reliever’s WPA is less than or equal to -0.06 in any individual game.
SD/MD is shutdowns per meltdowns.