Salvation from the Depths of Despair
Posted by Raphy on June 15, 2009
For the past few days , every time I turn on the radio it seems that invariably someone will ask "have you ever seen such a thing?" in reference to Friday night's Luis Castillo play. Well, let's take a look. Here are all the times since 1954 that a batter whose team was losing with 2 outs reached on an error to win the game.
Yr# G# Date Batter Tm Opp Pitcher Score Inn RoB Out Cnt Pit RBI Play Desc.
+-------+---+-------------+-----------------+---+----+-----------------+-----------+---+---+---+---+---+---+-------------------------+ 1 1 2009-06-12 Alex Rodriguez NYY NYM Francisco Rodrig down 7-8 b 9 12- 2 3-1 5 0 *ENDED GAME*:Reached on E4 (Pop Fly to Deep 2B); Jeter Scores/unER/No RBI; Teixeira Scores/unER/No RBI 1 1 2000-08-09 Juan Pierre COL PIT Mike Williams down 2-3 b 9 12- 2 1-0 2 0 *ENDED GAME*:Reached on E9 (Line Drive); Walker Scores/unER; Petrick Scores/unER; Pierre to 2B 1 1 1998-09-23 Geoff Jenkins MIL CHC Rod Beck down 5-7 b 9 123 2 2-2 5 0 *ENDED GAME*:Reached on E7 (Fly Ball); Loretta Scores/No RBI/unER; Cirillo Scores/No RBI/unER; Burnitz Scores/No RBI/unER; Jenkins to 2B 1 1 1997-06-15(2) Moises Alou FLA NYY Mariano Rivera down 4-5 b 9 123 1 0-2 3 1 *ENDED GAME*:Reached on E4 (Ground Ball); Renteria Scores; Floyd Scores/unER/No RBI; Sheffield to 2B 1 1 1992-09-19 Chad Kreuter DET BOS Greg Harris down 1-2 b 9 12- 2 1-2 4 0 *ENDED GAME*:Reached on E5 (throw) (Ground Ball to Weak 3B); Barnes Scores/No RBI/unER; Bergman Scores/No RBI/unER; Kreuter to 1B 1 1 1988-06-04 Cal Ripken BAL NYY Cecilio Guante down 4-6 b14 123 2 3-2 7 0 *ENDED GAME*:Reached on E5 (throw) (Ground Ball); Schu Scores/unER/No RBI; Gonzales Scores/unER/No RBI; Ripken Scores/unER/No RBI; Ripken to 2B 1 1 1984-04-02 Bob Boone CAL BOS Bob Stanley down 0-1 b 9 123 2 - 0 *ENDED GAME*:Reached on E6 (Ground Ball); Beniquez Scores/No RBI/unER; Lynn Scores/unER; Sconiers to 2B 1 1 1980-05-25 Larry Biittner CHC LAD Steve Howe down 0-1 b 9 -23 2 - 0 *ENDED GAME*:Reached on E6 (Ground Ball); Blackwell Scores/No RBI/unER; Randle Scores/No RBI/unER 1 1 1979-04-07 Willie Stargell PIT MON Elias Sosa down 5-6 b 9 1-3 2 - 0 *ENDED GAME*:Reached on E1 (Ground Ball); Alexander Scores/unER/No RBI; Parker Scores/Adv on E2/unER; Stargell to 1B 1 1 1967-07-07 Ron Hansen CHW MIN Al Worthington down 0-1 b 9 123 2 - 0 *ENDED GAME*:Reached on E6 (Ground Ball); McCraw Scores/No RBI/unER; Bradford Scores/unER; Horlen to 2B 1 1 1961-08-16 Jim King WSA LAA Jim Donohue down 1-2 b 9 12- 2 - 0 *ENDED GAME*:Reached on E5 (throw) (Ground Ball); Keough Scores/No RBI/unER; Zipfel Scores/Adv on E9 (throw)/No RBI/unER; King to 1B
A lot of ground balls, with a couple of flyballs mixed in. However, none of the others were on a popup. Some of the plays did involve multiple errors. I'll leave it up to you to decide if you want those to count.
To find these games use the Team Batting Event Finder. Include "all teams" and search for "reaches on error". Then filter your search for "behind", "as last play of game" "2 outs", and "put team in lead". Unfortunately, you need to search for each year individually.
June 15th, 2009 at 4:01 pm
Funny thing - I used this exact Event Finder today (I'm a diehard Met fan), and my friend made a post out of it.
http://theropolitans.com/2009/06/luis-castillo-not-first-wont-be-last.html
Beat you guys to the punch!
June 15th, 2009 at 6:01 pm
Nice post Andrew. Thanks for adding the link to our site.
June 15th, 2009 at 6:15 pm
The Brant Brown game was the first one I thought of. "NOOOOOO!!!"
June 15th, 2009 at 9:33 pm
There's something wrong with the Cal Ripken entry - under Play Description it says Ripken scores and then it says Ripken to 2B. Oh, wait, I figured it out - the Ripken who scored must have been Billy. Never mind.
June 16th, 2009 at 8:37 am
nice post raphy! this is but another example of why this is my favorite blog. superb!
i agree. when ever i hear any news or sports commentary say anything like "this has never happened before" (bob costas is one of the worst), it's a red flag - and i do my best to find an example of the event happening before. in this instance it was 10 times before! the proper headline the next day should have been "Yankees Vindicated!" twice before the Yanks were on the losing end of the same scenario. in fact, the 2nd time it was an error by Jeter, and only in his 2nd full season! i wish someone had known and asked Jeter after the Mets/Yankees game "what was it like for you?"
here's something interesting, in the 10 times: three times it was E5 (throw), and three times it was E6 (Ground Ball). only two of the ten times involved outfielders. if there's any "this has never happened before" involved, this was the first time it was a pop up. all others were: 8 grounders, 1 line drive, and 1 fly ball. and here are the names of the 10 fielders: Eddie Yost, Zoilo Versalles, Elias Sosa, Bill Russell, Jackie Gutierrez, Pags Pagliarulo, Tim Naehring, Derek Jeter, Brant Brown, and Alex Ramirez.
June 16th, 2009 at 9:09 am
this is a good time for me to bring up wWPA. ARod's wWPA for that play was 82%. Obviously this scenario is going to produce some mammoth wWPAs. Here the other ten from this list in ranking order:
*King, 83% (down 1-2, b 9, 12-)
*Kreuter, 83% (down 1-2, b 9, 12-)
*Ripken, 83% (down 4-6, b14, 123)
*Jenkins, 82% (down 5-7, b 9, 123)
*Pierre, 80% (down 2-3, b 9, 12-)
*Stargell, 80% (down 5-6, b 9, 1-3)
*Biittner, 76% (down 0-1, b 9, -23)
*Hansen, 75% (down 0-1, b 9, 123)
*Boone, 73% (down 0-1, b 9, 123)
*Alou, 47% (down 4-5, b 9, 123)
Why was Alou's substantially lower? I guess there's a much greater chance of the home team winning if the score is 4-5 in the 9th than other score possibilities. Is that why?
Some other questions I have about wWPA: Are wWPAs searchable? With the data available, what are the highest wWPAs in regular season history? What is the highest cumulative net wWPA for one player in one game? How about one player in one season? Is such information possible now? Could it ever been possible to create?
Last week I actually went through all the World Series games by hand and made a list of all wWPAs that were -40- or +40+. There were just under 60 such events. I'll post the top 20 shortly...give you all some time to guess which ones. Only two of them were 80+, and they are pretty obvious...and a third situation involved two different 40+s in one at-bat, making it a 80+ at bat, and that one is pretty obvious too.
Questions:
*If both Hansen and Boone were down 0-1, b9, 123, then why aren't their wWEs the same before their events?
*Why does Alou get an RBI for his AB? 2 outs, bases loaded. ground ball error. hmmm.
p.s. i just discovered that the columns in game play-by-plays are sortable! nice work!!
June 16th, 2009 at 11:10 am
On the Alou play there was only one out. I guess at that point in the game it makes that much difference. So Florida was actually considered more likely to win that game than lose it once Floyd doubled to put runners on 2nd and 3rd. They issued an IBB which *slightly* improved NY's chances, then Alou was up. (And since it was 1 out, that's why he gets the RBI.)
I believe that WPA varies according to the run environment. In 1967 it would have been a little tougher to score a run than in 1984, so Hansen's team had a slightly worse chance of winning that game than Boone's did prior to their respective ABs.
June 16th, 2009 at 11:19 am
Kingturtle --
As far as Alou goes, it might have something to do with not assuming a double play (bases were loaded). Only explanation I can think of.
Great job doing WS wWPA's, looking forward to the top 20.
My guesses for the 80+ are Gibson's HR in '88 and... hmmm tough call. Joe Carter? There was only one out and there were two on...
The 40/40 at bat has to be Mookie/Stanley/Buckner.
June 16th, 2009 at 11:26 am
kingturtle,
Pat Kelly made the error in the 1997 game, not Jeter.
June 16th, 2009 at 11:55 am
egads! thanks statboy! my brain short circuited there. sometimes my dyslexia does some astounding things. usually they are pretty fascinating. here's what i think happened...Jeter wears #2 for the Yankees. So did Sandy Alomar. Sandy Alomar played 2nd base for the Yankees, prior to Willie Randolph. So my brain short circuited and placed Jeter at 2nd base.
June 16th, 2009 at 12:50 pm
Here are the top wWPAs in World Series history....(note: i did this by hand, and may have missed one or two, but i think this legit)
1. 87%, 1988, Game 1, b9, down 3-4, 2 outs, -2-, Kirk Gibson's home run off Eckersley. "I don't believe what I just saw!" 87%!?!?! Was there ever a higher Major League wWPA??
2. 82%, 1947, Game 4, b9, down 1-2, 2 outs, 12-, Cookie Lavagetto's home run off Bevens that broke up the no-hitter. "Well, I'll be a suck-egg mule!"
3. 41%+40%=81%, 1986, Game 6, b13, down 4-5 and then tied 5-5, 2 outs, 1-3 and then -2-, wild pitch by Stanley brings in one run, then Buckner's error brings in the winner. "It gets by Buckner!"
4. 69%, 1941, Game 4, t9, 2 outs, 12-, down 3-4, Yankee Charlie Keller doubles in Henrich and DiMaggio
5. 69%, 1985, Game 2, t9, 2, outs, 123, down 1-2, Cardinal Terry Pendleton doubles in three runs
6. 67%, 1992, Game 2, t9, 1 out, 1--, down 3-4 Blue Jay Ed Sprague homers off of Jeff Reardon
7. 66%, 1993, Game 6,, b9, 1 out, 12- (Molitor and Rickey Henderson), down 5-6, Blue Jay Joe Carter hits a game-winning, World Series-winning home run off Mitch Williams. The unsung hero of the inning was Rickey Henderson. With over 1000 stolen bases to his name already in his career, he led off with a 4-pitch walk. He distracted Williams completely. Williams threw to first 5 times. Molitor was able to get a single. and the rest is history.
8. 64%, 1960, Game 7, b8, 2 outs, 1-3, down 6-7, Pirate Hal Smith hits a three run homer to take a 9-7 lead. Mantle and Berra each bat runs in the top of the 9th to tie it, and Mazeroski wins in with a dinger in the bottom of the ninth, it what may be the most exiting game seven ever. Interesting to note: Mazeroski's homer was only a 37% wWPA.
9. 58%, 2005, Game 2, b7, 2 outs, 123, down 2-4, White Sock Paul Konerko hits a grand slam.
10. -56%, 1972, Game 4, t8, 2outs, 1-3, down, Bobby Tolan doubles off Vida Blue scoring Concepcion and Morgan. But the A's won it in the bottom of the ninth with a rally that included a pinch hit single by Don Mincher with a 52% wWPA, a hit that will appear later in this list.
Numbers 11 to 20 I'll post later tonight.
June 16th, 2009 at 1:56 pm
11. 55%, 1933, Game 4, b11, 1 out, 123, down 1-2, Senator pinch hitter Cliff Bolton grounds into a 6-4-3 double play to end the game. Yes, even a GIDP can produce a massive wWPA. Let's set this one one...The Giant's Carl Hubbell and the Senator's Monte Weaver each pitched 10 strong innings. The score was 1-1 after 10 innings. In the top of the 11th, Weaver gave up a run on two hits and a sacrifice hit. After giving up another hit, this time to Hubbell himself, Weaver was taken out and Jack Russell got the last two outs. In the bottom of the 11th, two singles, a sacrifice and an intentional walk loaded the bases with just one out. One out, bases loaded, down 1-2. GIDP=55% wWPA. Astounding.
12. 54%, 1923, Game 6, t8, 2 outs, 123, down 3-4, Yankee Bob Meusel singles two in, and a third scores on a throwing error by the center fielder. These are the last three of a five run inning, and the Yankees win the game and clinch the World Series.
13. 52%, 1928, Game 2, t8, 2 outs, 1--, down 2-3, Yankee Frank Crosetti homers off of Dizzy Dean. Man would I have loved to see that.
14. 52%, 1972, Game 4, b9, 1 out, 12-, down 1-2, Athletic Mincher (as already mentioned above) singles off Clay Carroll. Two +50 wWPAs in one game? Wow. Mincher'a single ties the game, and the next batter, pinch hitter Angel Mangual, wins it with a single.
15. 52%, 1998, Game 3, t8, 1 out, 12-, down 2-3, Yankee Scott Brosius hits a three-run dinger off Trevor Hoffman. Remember it? I sure do.
16. 50%, 1993, Game 4, t8, 2 outs, 12-, down 13-14, Blue Jay Devon White hits a two-run triple off Mitch Williams. Williams is on this list twice, and for two incidents in different games of the same World Series. He was not popular in Philadelphia after that, as we know.
17. 50%, 2001, Game 7, b9, 1 out, 12-, down 1-2, Diamondback Tony Womack doubles in the tying run off Mariano Rivera. Counsell is then hit by a pitch, and Luis Gonzalez hits his game-winning, Series-clinching single (which was only a 16%wWPA). As is the case with Mazeroski's 1960 homer, it was the play that set up the climactic hit that gets the big wWPA, not the climactic hit itself.
18. 49%, 1922, Game 5, b8, 2 outs, 123, down 2-3, Giant George Kelly singles in two, solidifying the win that clinches the WS.
19. 49%, 2001, Game 4, b9, 2 outs, 1--, down 1-3, Yankee Tino Martinez hits a game-tying homer off Byung-Hyun Kim. The Yankees win in the b10 with a homer by Jeter (a 46%). We'll hear from Kim again later.
20. 49%, 2001, Game 5, b9, 2 outs, -2-, down 0-2, Yankee Scott Brosius (yes, he is on this top-20 list twice!) hits a game-tying home run off of Byung-Hyun Kim (I told you we'd hear from Kim again; poor guy.)
21. -49%, 1964, Game 5, b9, 2 outs, -2-, Yankee Tom Tresh hits a game-tying home run off Bob Gibson! Damn Yankees! Mantle was the guy on 2nd. But the Cardinals got three runs in the t10 with a dinger by Tim McCarver, and the Cards won the game, as Gibson shut the Yanks down in the b10.
June 16th, 2009 at 2:52 pm
That was yeoman's work, Kingturtle. What could be even more interesting is somehow converting those into series WPA...a Game 1 hit would not change the overall chances of winning the Series as much as a Game 7.
Also, of course win probability can never be absolutely accurate. It's based on averages. I am 95% sure the fact that Eckersley is pitching instead of Stan Javier makes no difference in the calculated Win Prob, though obviously it would make a huge difference in the actual Win Prob. So if Oakland was calculated to have an 87% chance of winning when Gibson came up, when you consider it was Eck on the mound and an injured player at the bat, perhaps their chances were really 95% or something.
June 16th, 2009 at 4:31 pm
My apologies. The Alou play does not belong because there was only 1 out.
As always nice work KingTurtle.
June 16th, 2009 at 5:45 pm
"So if Oakland was calculated to have an 87% chance of winning when Gibson came up, when you consider it was Eck on the mound and an injured player at the bat, perhaps their chances were really 95% or something."
That's a great point, and one that emphasizes why storytelling isn't dead - sometimes the numbers only give you so much!
June 18th, 2009 at 10:03 am
On June 16 in the Blue Jay/Phillie game, Scott Rolen had two consecutive at bats which produced 28%+ wWPAs. In the top of the ninth, none out, a runner on first, and down 2-3, Rolen singled and got to second base on the throw to third. That was a +33% wWPA. No run scored. But it really changed the Blue Jay's potential. Then in the top of the tenth, one out, runners on first and second, and tied 3-3, Rolen singled, scoring Hill, and second Wells to second. +28% wWPA.
It would be awesome if someday we could use Play Index to search wWPAs. What are all the instances of a batter getting two 28% wWPAs in a game? From that list we could see who did it in two consecutive at bats.
June 18th, 2009 at 12:05 pm
Dusty Rhodes has died. He didn't make Kingturtle's lists above but had a number of big hits in the '54 Series. The biggest was his 3-run homer to win the game in the bottom of the 10th in Game 1. The Giants already had a 71% chance of winning at that point, so his hit was +29. In the next two games, he had three more hits which were +14, +10, and +13.
The odd thing which I never realized (or had long forgotten) was that he was pinch hitting for Monte Irvin, a very fine player albeit past his prime by then. In Game 3, he pinch hit for Irvin in the top of the 3rd inning. He did have the platoon advantage, but Cleveland's starter (Garcia) was still in the game. Why didn't Rhodes just start?
June 18th, 2009 at 1:48 pm
Kingturtle --
One game immediately comes to mind, in 2 consecutive at bats. Mets @ Philly, 6/2/04. Down 3-0 in the 8th, Todd Zeile hits a game-tying 3-run homer (wWPA - 33%) then in the top of the 10th hits a go-ahead 2-run homer (wWPA - 44%). Not bad for the cagey veteran!
June 18th, 2009 at 1:56 pm
Also, "87%!?!?! Was there ever a higher Major League wWPA??"
I just did the research - since Gibson homered with the bases -2-, what would the wWPA if, still with two outs, the bases were 1--? Answer - 90%. (Last time it was done - David DeJesus, 7/12/08.) Same goes for the "ultimate" walkoff: bases loaded, two outs, down 3 runs, walkoff grand slam - wWPA = 90%. (Last time it was done - Adam Dunn, 6/30/06.) But for some reason, down 2, 2 outs, 12- gives a wWPA of 91%. (Last time it was done - Marco Scutaro [off Rivera!], 4/15/07.)
June 18th, 2009 at 2:28 pm
I remember the Scutaro homer. I think it hit the LF foul pole, or maybe just wrapped around it. Of course it probably set off another batch of WWWMR articles. Scutaro had an inordinate number of GW hits/homers as a utility player for Oakland. Now in his first chance as a regular he's having an excellent season for Toronto.
June 18th, 2009 at 2:35 pm
Look for low-scoring seasons to find higher WPA plays. I just checked 1968. Tony Horton's HR off Pat Dobson with 2 outs in the bottom of the 13th on 6/21/68 is +92%, and that's with only one man on base, down one run.
June 18th, 2009 at 4:35 pm
Very interesting. But if it varies by year, how do you decide what the likelihood when a situation hasn't happened all year? For example, let's say in April of this year being up 3 runs in the 9th has resulted in that team winning 100% of the time, when usually it's more like 90%. Does that make a grand slam walkoff a 100% wWPA, since that year it's worth 100?
June 18th, 2009 at 8:38 pm
No. It's not based so much on the exact history of specific situations, but on run expectancy for every base/out situation. Given a certain run-scoring level, it is known that at the beginning of an inning an average team has (say) a 5% chance of scoring 3 or more runs in an inning. Once they load the bases, if there's still no one out, maybe it goes up to 25%. Start here: http://www.baseball-reference.com/about/wpa.shtml
June 18th, 2009 at 10:05 pm
Right, that I understand. But once we establish there's a year difference (the same play in 1968 produced a 92% wWPA as opposed to 90% 30 years later), how do we establish when those changes occur?
June 19th, 2009 at 12:52 am
It would be nice if Sean could chime in and/or make his glossary more explicit. I believe he gets his WPA and Leverage data from Tango Tiger (insidethebook.com). But in skimming through some of Tango's stuff, it's not clear to me whether he is generating his run/win expectancy charts based on actual data, based on Markov chains, or what. I do tend to "trust" the numbers on the site, but I'd like it to be a little more obvious how some of these things are calculated. I use the site a lot and have a good understanding of sabermetrics, but if I can't answer Student's question better, I imagine most B-R users are looking at these numbers with either skepticism or faith, rather than understanding.