By Adam Gilfix
Recently, my fellow HSAC contributor and friend Carlos Pena-Lobel put up an article on the most pivotal plays in Super Bowl history, using the basis of win probability added. That got me thinking, and with pitchers and catchers reporting throughout this coming week, I figured why not look back and find the all-time most significant plays in the World Series? This marks the beginning of a multi-post series, and I start with enumerating the plays with the biggest change in win probability within a single game of the Fall Classic since 1905. Subsequent posts include ranking the biggest change in win expectancy (used interchangeably with probability) for an entire Series and ranking players in World Series history with the best win probability-added statistics on a game-level and on a series-level. Hope everyone else is as excited as I am for baseball to start soon, so this will hopefully hold you over until Spring Training games.
Having scraped every World Series game play-by-play from Baseball Reference’s Postseason archives since 1905, I had about 50,000 plays to work with. One of the only data-cleaning processes I had to go through was removing tie games from my spreadsheet and accordingly updating game numbers. There were 3 such games in World Series history: 1907 Game 1 (so Game 2 becomes Game 1, etc.), 1912 Game 2 (part of an 8-game series that I will henceforth refer to as 7 games), and 1922 Game 2. Furthermore, although this doesn’t matter as much for this post so much as the next, it is important to note that not all World Series followed the current 2-3-2 format. In fact, not all World Series have even been best-of-7 series (1903 wasn’t and 1904 didn’t have a Series, which is why I didn’t bother starting before 1905): the 1919, 1920, and 1921 were all best-of-9 series that went 8, 7, and 8 games long, respectively. Again, this will be more important in my subsequent article, but is worth taking note of in this contextual set-up.
Another data fix I dealt with was converting Baseball Reference’s winning team Win Expectancy (WE) and Win Probability Added (WPA) to the easier to work with, batting team WE and WPA. After applying some Excel formulae and using a pivot table, I was able to create new columns containing the Batting Team WE (before and after), as well as the difference between the two. This difference is the main focus of this post: the Batting Team WPA on each play, which simply reveals how much a batter (or baserunner) added to his team’s chance of winning the game on that play given the current inning, scoring margin, base-out scenario, and run environment. For more information on how Baseball Reference calculates the WE for each possible situation in a game, check out this helpful page.
Having created my Batting Team WE and WPA columns, all that was left to do was sort the plays by the greatest WPA in terms of absolute value. Green represents a walk-off, orange means the batting team eventually lost the game, and red denotes the batting team lost on that play. Without any further ado, here are the plays in World Series history that swung a team’s chances of winning that game (altered the WE) by at least 50%.
First of all, to put these plays in perspective, looking at all 48,500 plays in World Series history between 1905 and 2014 (including the ties and Game 8’s, which I will disregard in the next post), the average magnitude of the WPA was 3.35%. Compare that to the 61% average magnitude of WPA of these 17 plays. Just think: almost three-fourths of plays in the World Series have WPA magnitudes at or below 3% and only 3,000 plays saw WPA’s in double figures. Now add the fact that though these 17 at-bats represent a tiny sliver of all the plays, their combined WPA magnitude represents more than half a percent of all win probability fluctuation at the game-level in World Series history. OK, now that you have come to my conclusion that these plays had a truly gargantuan effect in their respective games, we can dive deeper into some of the most significant plays in the storied past of the pinnacle of postseason baseball.
Summarizing these plate appearances a bit, one can see that 16 of the 17 occurred with a man on 1st, 13 with a man on 2nd, 6 with a man on 3rd, all but 2 with RISP, and an average of just over 2 men on base. The home/away split was essentially even at 9 and 8, respectively, and these historic game-changers could be found spread out about evenly throughout a 7-game series. As expected, given the nature of adding at least 50% (in magnitude) to the batting team’s win expectancy, 14 of these top plays were extra-base hits (and half of those were home runs).
More interestingly, all but one of the plays occurred with the batting team down by 1 run, with one at-bat where the team was down 2. Combine this with the fact that 10 of the 17 occurred in the 9th inning or later (with the rest in the 8th except for one in the 7th) and the fact that 11 of the 17 occurred with 2 outs (the rest with 1), and one can see that high leverage situations make for big fluctuations in win expectancy. With the average runs scored on these top 17 pivotal plays being 2.17, 2.31 when not including the double play to end the game, the batting team clearly exhibited huge WPA shifts by taking the lead on all but 3 of the plays (2 plays tied up the game and the 1 batting team loss).
Besides that game-ending double play for the Washington Senators in 1933, only one other play in this list of the biggest swings in single-game Win Expectancy featured a batting team that went on to lose that game. In Game 4 of the 1972 Fall Classic, the Reds were down 0-1 with 2 on and 2 out and took a 2-1 lead in the top of the 8th on Bobby Tolan’s clutch double (WPA of 56%), only to have Oakland’s Don Mincher (WPA of 52%) and Angel Mangual’s (WPA of 19%) back-to-back RBI singles help the Athletics walk off with the win that helped them on their way to an eventual Game 7 celebration. Like Tolan’s, Mincher’s play can be found in the top 17 above, but Mangual’s walk-off came in a situation where his team was actually pretty likely to win the game (tied in the bottom of the 9th, 1 out, and men on the corners), so despite the heroics, it is nowhere near the biggest magnitude change in Win Expectancy in a game in World Series history.
However, the list above does include 3 walk-off hits (2 homers and a double): Brooklyn’s Cookie Lavagetto’s pinch-hit double (WPA of 82%, 2nd highest ever) with 2 on and 2 out in the bottom of the 9th of Game 4 (the “Cookie Game”) of the 1947 World Series (Jackie Robsin’s Dodgers would eventually fall to the Yankees in Game 7); Kirk Gibson’s 1988 Game 1 iconic blast for the Dodgers (WPA of 87%, highest ever) with 2 outs and a full count off one of the best relievers ever; and, of course, Joe Carter’s 1993 Game 6 dinger to win the Series (WPA of 66%) for the Blue Jays, who won their second straight banner.
Two plays in the list occurred in pivotal, late Game 7 situations…both at the expense of the Yankees. One of those is Arizona Diamondback Tony Womack’s 1-out line-drive double down the right field line (WPA of 50%) to tie the game at 2 against the Yanks in the bottom of the 9th inning in a 3-3 series in 2001; it doesn’t get much more clutch than that. Oh, did I forget to mention that hit came off the greatest closer in Major League Baseball, perhaps one of the best players ever, playing for (though I begrudge giving them the title, but at least they lost the series) the winningest franchise in MLB history? Of course, everyone knows what happened 2 plays later, but that World Series-winning walk-off can’t quite cut it with a 16% WPA for Luis Gonzalez.
The other Game 7 phenomenon of WPA was Pittsburgh Pirate Hal Smith’s 3-run blast to left against New York in 1960 in the bottom of the 8th with 2 outs to give his team a 2-run lead that helped them on their way to an historic victory over the Yankees (more on this in the next installment of this multi-post collection). This home run represented a 64% increase in WE for the game, and due to its highly significant situation and timing, is the most pivotal play in an entire series (in terms of Championship Probability Added). In fact, my next post addresses the issue of ranking the all-time biggest magnitude changes in Championship (World Series) Win Expectancy at a series level and will include a corresponding list as well (look out for Hal Smith).
For those of you wondering, and those of you who are Mets fans (or hate the Red Sox), there is a very significant play in World Series history that exhibited a drastic swing in Win Expectancy. Although it makes me (although I wasn’t even born) and millions of other members of Red Sox nation cringe, I will share a very bizarre play that is one of the all-team highest in WPA for a single World Series game.
If your mind went immediately to the infamous ball between Bill Buckner’s legs…you’re wrong. Well, not completely. Before that notorious play, the Red Sox were actually still winning in the bottom of the 10th in Game 6 of the 1986 World Series in Shea Stadium. Having taken a 5-3 lead in the top half of that extraordinary inning, the Red Sox got two outs before surrendering a few hits and it was 5-4. What happened next when Mookie Wilson stepped up to the plate against Bob Stanley is forever imprinted on the minds of Sox fans, but we often only remember the end. Something happened in the middle however: on the 7th pitch of the at-bat, after Wilson fouled off a few pitches to stay alive, with the Sox one strike away from ending the Curse of the Bambino, Stanley delivered a pitch too far inside that went to the backstop and allowed Mitchell to score from the third base and knot the game at 5 apiece. This unlikely (and unfortunate) event happens to be the largest game-level WPA for a play not involving the batter in World Series history with a 41% swing in WE (ranking tied for 23rd of all-time WPA in a World Series game). Just 3 pitches later, a 40% change in WPA rolled through the ankles of poor old Bill Buckner when Knight came around from second base.
I can’t end on that momentously tragic note. So I leave with you a happy reminder that at some point, I will put up my next post on the World Series plays that had the largest magnitude change in Championship Win Expectancy.