Two blown saves in extra innings
Posted by Andy on August 5, 2011
Yesterday the Blue Jays blew two saves in extra innings and lost the game. It made me curious to find other recent games where a team blew more than one save after the 9th inning.
- In this memorable game 5 weeks ago, the Nationals blew a 3-run lead in the bottom of the 9th, and then blew 1-run leads in both the 10th and 12th inning before finally scoring 4 times in the 14th and having it hold up. That's 3 blown saves, of which 2 came in extras.
- About a month before that, the Rockies blew 2 saves in extras against the Brewers.
So that's 3 such games in 2011. I went back and checked all of 2010 and 2009 and there were none...amazing.
August 5th, 2011 at 7:57 am
It's a pretty narrow set of circumstances that allows an extra inning blown save to happen at all:
1) obviously, an extra-inning game;
2) road team takes the lead in the top of the inning:
3) new pitcher takes over for the road team in the bottom of the inning;
4) new pitcher gives up enough runs to not only tie, but lose the game in the bottom of the inning.
The perfect game (and near perfect game) cluster from last season tended to remind us that the improbable can happen--but I was surprised that the Rockies/Brewers games apparently took place in Milwaukee (since a higher run scoring environment such as even humidor Coors provides would tend to produce more scoring in any particular inning).
August 5th, 2011 at 8:02 am
Ooops, strike the last one--obviously, allowing the tying run to score is enough to blow the save, even if the game continues.
August 5th, 2011 at 8:32 am
I thought this game from 1995 would qualify for sure, but even though the Astros lost the lead in the 10th and the 11th, the pitcher in line for the win actually blew it in the 10th so no blown save: http://www.baseball-reference.com/boxes/CHN/CHN199509280.shtml
August 5th, 2011 at 8:37 am
How about 2 blown saves by the same pitcher? Guy comes in with the lead gives up the tie run, moves to right field. Comes back and pitches the next inning and does it again.
August 5th, 2011 at 8:48 am
I just saw where Bill James wrote a book on the Boston strangler. Is he going to have some new statistics for the type of kill performed? Like WAR, but related to strangling?
August 5th, 2011 at 8:59 am
Andy, thanks for drawing yesterday's painful experience for Blue Jays' fans to everybody's attention. 🙁 🙂
Actually, full credit to Tampa Bay for getting it done in high-leverage at-bats, but the game was agonizing to follow. If you cared about the teams, it was like being on a roller coaster of extreme elation and depression.
I suppose multiple blown saves in extra innings could happen at "random" to any team, even a good-bullpen one, but Toronto's bullpen absolutely reeks.
Surprisingly, their 55% save percentage (19 blown, 23 saved, including last night) isn't the worst in the majors. It must hurt to watch Houston's bullpen at work.
Also surprising to me that the Angels and Cardinals have such poor save efficiences for playoff contending teams.
August 5th, 2011 at 9:16 am
Actually Neil, Toronto's bullpen is pretty decent; the 3.70 bullpen ERA is right around average (8th in AL, 16th in MLB, better than the 3.82 AL average). This despite having to throw a lot of innings and
The problem is that John Farrell has no idea whatsoever how to run a bullpen, and isn't ever able to find a right role for his pitchers. For most of the year, the good relievers have been pitching in sixth-, seventh-inning roles or mopping up, while the awful relievers have been given the eighth- and ninth-inning roles. Farrell's misuse of Francisco, Rauch and Dotel has been appalling. Farrell seems unable to look past a guy's salary number when trying to figure out if he can pitch.
August 5th, 2011 at 9:32 am
@7
Thanks, Craig, for talking me off the ledge about the Toronto 'pen. However, you are using ERA as the only measure of effectiveness, not the ability to close out games they are leading in the late innings.
And don't forget that a significant portion of the Blue Jays' decent bullpen ERA was Marc "Scrabble" Rzepczynski and Jason Fraser. After them, the relief consistency has been pretty thin.
What is a more important indicator of bullpen quality? To have a one-man wrecking crew like Rivera or to have middle-relievers and set-up men keep you in the game after your starter has faltered?
August 5th, 2011 at 10:38 am
@3 This game is one of my all-time favorites! It was tied at 2-2, 5-5, 6-6, 7-7, 9-9, 10-10, and 11-11. There were also four times when a team blew a 83%+ WPA (the Cubs in the 4th, the Astros in the 8th, 10th, and 11th). Wonder if that's the most for a game or a team! Amazing how they exactly matched runs from the sixth to tenth!
August 5th, 2011 at 10:45 am
Here's (a painful) one from 2003. Also featured a BS in the 8th inning, so three total in the game.
http://www.baseball-reference.com/boxes/PHI/PHI200306210.shtml
August 5th, 2011 at 10:52 am
What exactly is the record for most different extra innings where a team scores at least one run?
August 5th, 2011 at 11:53 am
Just realized that the aforementioned Nationals-White Sox game also featured four times blowing a 83%+ WPA. Before winning the Nats actually blew it when they had WPAs of: 98% in the 9th, 89% in the 10th, and 95% in the 12th! If they lost it may have been one of the most unlikely in history.
I've actually been trying to come up with a baseball analogy to the US women's soccer team losing the World Cup final the way they did. I think this game, had the Nats lost, would have been a good one.
August 5th, 2011 at 12:00 pm
@3 @9
Jay and Eorns, thank you. Your posts actually got me to look at the box of that game. Four blown saves and two holds for the Astros!! That takes the cake.
Again the Astros-Cubs game. Fifteen of the twenty two half-innings were scoring innings, resulting in 9 lead changes and 6 ties, as you pointed out, Eorns. More lead changes and ties than most basketball games. 🙂
The Wrigley faithful who hung in for the whole almost-five-hour game must have needed a nervous system transplant. But they went home happy!
August 5th, 2011 at 1:53 pm
Ha, well put Neil 🙂
Elias pointed out a game that had even more blown chances, a 7/24/98 Phillies-Marlins game. The Marlins blew leads in the 9th (95% Win Expectancy), 10th (85%), 11th (83%), and 12th (81%) innings. For good measure, the Phils blew an 84% opportunity in the 11th. There were actually five plays of at least 35% WPA in this one game! Not sure if you can really calculate it this way, but the chances of those five events occurring (.05*.15*.17*.19*.16) is 1 in 25,800.
(Pardon my previous misuse of WPA. I meant Win Expectancy.)
August 5th, 2011 at 2:01 pm
Just noticed that the crazy Phillies-Marlins game was the second game of a doubleheader. Would you believe that in the first game the Marlins blew the lead in the ninth (83% WE) and lost it in 12? That's basically five blown saves and two losses in one day! Has a pitching staff ever had a more frustrating day??
August 5th, 2011 at 2:29 pm
@6, regarding the Cardinals' poor save %, Ryan Franklin blew 4 out of 5 save opportunities before they ditched him. Salas is their main closer now, and he's 19 of 22 in save situations.
And of course set-up men also get blown saves if they give up the lead -- maybe they should be called blown holds? 🙂
August 5th, 2011 at 3:14 pm
Another way of identifying blown leads is using WPA. Here are the worst 16 games in terms of team pitching WPA in the play-by-play era:
http://bbref.com/pi/shareit/N0VKa
The first game on the list was the only one with team pitcher WPA below -2.0; the Cubs gave up four leads: one 1-run lead, two 2-run leads, and a 3-run lead, all in just 9 innings. Montreal pitching had a -1.595 WPA in the same game, the worst by a winning team (by definition, it has to be exactly .500 better than the Cubs' value, since Montreal won).
The fifth game on the list is the 1995 Astro-Cub game mentioned above. The fourth game on the list is the one in which Art Shamsky hit 3 HR as a sub in a losing effort, the highest single-game WPA by a player (+1.503).
August 5th, 2011 at 3:45 pm
Great find! I wonder if there's been a more roller coaster inning than the 9th of that Cubs-Expos game. After making the first out in the top of the frame, Chicago's Win Exp was only 4%. After scoring 4 runs and retiring the first batter in the bottom of the ninth, it rose all the way to 95%. As they allowed 3 runs and lost it went all the way back down to 0%. They did not go home happy on that one 🙂
Henry Rodriguez hit a 3-run go-ahead homer in the top of the 9th, getting .75 WPA. Is this the highest value you can get in one PA without winning the game?
August 5th, 2011 at 4:24 pm
Confirmed that that is indeed the case. I found two two-outs-in-the-top-of-the-ninth-down-by-three grand slams, and the WPA is .75. The last two times it was done was by hall-of-famers: Dave Winfield in 1992 and Andre Dawson in 1991.
Dawson's was pretty amazing in that his Cubs actually still managed to lose the game (this thread is turning into Cubs' stomach-chruning games!). The Cubs were down 3-0 having managed just one infield single and two walks through 8.1 innings against the Pirates' Vincente Palacios and Stan Belinda. A walk, a double, a strikeout, and a HBP later, up steps Andre Dawson to pinch hit with two outs and the bases loaded. He then sends the first pitch he sees over the left field wall to take the lead 4-3.
In the bottom of the inning, Dave Smith gives up the lead on a wild pitch without even recording an out. And, after intentionally walking both Bonilla and Bonds with two outs, gives up a game-losing single to Jeff King. How great is that?
August 5th, 2011 at 4:59 pm
@19, I think hitting a HR with 2 outs in the bottom of the 9th (or later) with bases loaded and 3 down gives a bigger WPA, since the other team doesn't have a chance to come back.
The most recent such game was on 2006-06-30 by Adam Dunn, with WPA of .90.
August 5th, 2011 at 5:00 pm
@16
Whiz, regarding blown holds. I used that phrase a long time ago here and was gently reprimanded, by Johnny T., I think, for not calling them by their real name ....... losses. 🙂
Thanks for the list in #17. Strange that there are only two AL and one interleague game out of the sixteen on your negative WPA list.
August 5th, 2011 at 5:12 pm
@20, a list of all such games (in the play-by-play era, of course) is (with the player, inning it occurred and WPA):
2006-06-30, Adam Dunn, 9th, .90
2001-07-28(1), Brian Giles, 9th, .90
1996-05-17, Chris Hoiles, 9th, .90
1988-06-21, Alan Trammell, 9th, .90
1986-08-29, Dick Schofield, 9th, .90
1985-04-13, Phil Bradley, 9th, .90
1984-08-31, Buddy Bell, 9th, .90
1983-04-13, Bo Diaz, 9th, .91
1979-05-01, Roger Freed, 11th, .91
1973-04-22(1), Ron Lolich, 9th, .90
1970-08-11, Carl Taylor, 9th, .90
1963-08-31, Ellis Burton, 9th, .91
1955-09-11(1), Del Crandell, 9th, .91
A pretty nondescript list; I guess you had to be in the right place at the right time. Roger Freed is the only one to do it in extra innings. The WPAs weren't all identical, so maybe there was a slight difference due to park or era (or round-off error). My win expectancy tables give a WPA of .909 for this.
August 5th, 2011 at 5:20 pm
@21 Neil, nice pickup there -- I hadn't noticed that most were NL games. There must be a reason for that. Ignoring the inter-league game, the odds of that happening due to random chance are about 1 in 190. (Well, not quite that long since there have been more NL teams recently, and I assumed a 50-50 chance for NL versus AL.)
August 5th, 2011 at 5:36 pm
Eorns, @18
If you hadn't commented on the Cubs-Expos game I might not have bothered to look at it.
Wow, I've never seen a Win Probability Graph go vertical like at that Henry Rodriguez at bat. And then come crashing down on the other side of the peak as the game reversed itself on the Cubs.
Gotta be a little bit of advanced calculus we could apply to it to find the slope of the tangent there. 🙂
August 5th, 2011 at 5:51 pm
@22 Hard to say any such grand slam could be nondescript, but Schofield's deserves extra mention as it capped an 8-run bottom of the 9th rally.
August 5th, 2011 at 5:58 pm
@17
Whiz, again along the line of change in win expectancy during a half-inning or an at-bat, I surprised that what a Hardball Times article calls the greatest comeback of all time, the Clevleand-Seattle game from 2001hasn't entered any of our discussions.
http://www.baseball-reference.com/boxes/CLE/CLE200108050.shtml
I guess there weren't many lead changes during the game and the 12-run comeback was spread out over a number of batters and not concentrated in a plate appearance or two.
.
August 5th, 2011 at 6:06 pm
@26, Neil, yes, the first 8 runs or so of the comeback don't add much win expectancy. It's the repeated comebacks that really rack up the WPA.
August 5th, 2011 at 6:06 pm
@25, I meant that the players were nondescript, not the events 🙂
August 5th, 2011 at 6:33 pm
OK, I started wondering how often those situations (2 out, bottom of 9th, bases loaded, down by 3) come up. (I didn't do extra innings because that was more work.) Turns out it's about 6 times a year. There were a lot of these HR hit in the 1970s and especially the 1980s -- looking at the batters' slash stats in those PAs turned up this split:
1990-2011 .162/.228/.282
1970-1989 .308/.373/.533
What a difference!
August 5th, 2011 at 7:28 pm
I guess here's as good a place as any to post this.
According to his Web site, Billy Joel's album The Nylon Curtain was released on September 22, 1982. One of the songs from The Nylon Curtain is "Pressure," which contains the lyrics "But here you are in the ninth, / Two men out and three men on, / Nowhere to look but inside / Where we all respond to pressure."
Sez to myself, sez I, "These lyrics sound like the criteria for an Event Finder search." And that is what they became.
On September 22, 1982, Rafael Ramirez of the visiting Atlanta Braves faced Houston's Frank LaCorte with two outs in the top of the ninth inning, the bases loaded, and his team trailing 3-2. Pressure! Click here to see how the confrontation turned out.
August 5th, 2011 at 8:01 pm
There is a Non-Parametric Runs Test for Randomization of Results.
Fortunately, I have it on my computer. Send your data to MeadGuild@aol.com and I will send you the results.
To paraphrase Ben Franklin, Baseball is proof that-Reference.com, and everyone who has Baseball Research easier.
Dick
-----
Richard D. Adams
Ellicott City, MD 21042-3928
August 5th, 2011 at 11:54 pm
@29
Whiz, your list @22 shows which of the at-bats from #29 resulted in home runs, correct? There were three regular-inning, walk-off grand slams with two out, down by three in the seventies, five in the eighties and so on.
Any theories about why the pressure got to the batters a lot more in the last two decades than in the previous two? I guess it must be noise based on samll sample.
August 6th, 2011 at 1:20 am
@30
Kahuna, how do you come up with such obscure, but interesting, baseball stuff. Don't remember the Billy Joel song, but you picked a good ol' New York boy. He was a good lyricist who used his city for inspiration so why not a baseball allusion?
And I bit on your linked game after the Billy Joel teaser. Spoiler alert ....... Ramirez pulls a Carlos Beltran.
August 6th, 2011 at 2:08 am
@22,29,32. Might part of it not be that current "closer" usage leads to better results late in the 9th?
August 6th, 2011 at 8:13 am
In Thursday (Aug. 4) Toronto at TB game (12 innings, TB 7, Toronto 6), C Robinson Chirinos (yeah, THAT Robinson Chirinos) slingled in the bottom of the 11th to tie the game and then singled home the winning run in the bottom of the 12th to win the game. Chirinos was single-handedly responsible for both blown saves (Jon Rauch and Shawn Camp).
Chirinos, incidentally, had only been up about a week or so. Great way to break into the majors
August 6th, 2011 at 12:48 pm
@34
True, KDS, specialized closing started with Eckersley(?) so after his impact the batter in the bottom of the ninth, in these situations, is facing the opposition's ace reliever, as opposed to the previous decades.