Adventures in Playoff Leverage and Win Probability Added

Playoff baseball is interesting as a concept. After a regular season of 162 games to determine the game’s best teams, the sport’s champion is then determined by a few best-of-five and best-of-seven series. It’s not unlike asking the top 10 finishers of a marathon to run a 5K in order to decide who should receive first place. The sprint-like nature of the postseason is baseball’s Theatre of the Absurd (especially where small sample sizes are concerned): entertaining and a bit preposterous at the same time.

One of the areas where the effect is most pronounced is in the realm of Win Probability Added (WPA) and Leverage Index (LI). Championships are on the line and the lens of the postseason only serves to magnify what would be tense moments even on a quiet night in July. A big WPA day turns a player into a legend, while going the opposite direction turns a player into the goat. But not every intriguing event with a high WPA or LI is a starring turn. With that in mind, let’s look at a few of the stranger WPA- and LI-related things we’ve seen during the League Championship Series.

Caleb Ferguson and Playoff Stress

Caleb Ferguson was a 38th-round pick out of high school for the Dodgers in 2014. A starter through his whole minor-league career — he recorded only three relief appearances in the minors prior to this year — he found a home in the Dodgers’ bullpen this year. While he doesn’t have an incredible arsenal — Eric Longenhagen and Kiley McDaniel graded his fastball as a 50, curveball as a 45, and changeup as a 45 — he put up solid numbers as a reliever, striking out over 30% of batters and produced a 2.55 xFIP. After that solid rookie season, Ferguson joined the playoff roster as one of three lefties — the other two being Alex Wood and Julio Urias — in the Los Angeles bullpen.

Generally speaking, he didn’t pitch in high-leverage situations this season. With an average leverage index of 1.08 (Overall average is 1), he ranked 123rd in baseball for relievers with at least 30 innings pitched. In the League Championship Series, however, things have been a little different.

2018 LCS Leverage Index Leaders
Player pLI WPA/TBF
Kenley Jansen 2.48 0.031
Caleb Ferguson 2.14 0.017
Jeremy Jeffress 2.09 -0.026
Junior Guerra 1.87 0.004
Ryan Brasier 1.66 0.018

Ferguson appeared in Games Two, Four, and Five for the Dodgers, all of which the team eventually won. In each, he was called in to get the left-handed batter at the plate out; in all three, he left his team in a better position to win. His lowest leverage moment came on Wednesday night, when he faced off against potential 2018 NL MVP Christian Yelich in the ninth inning. His highest-leverage moment saw him facing Mike Moustakas and Travis Shaw with his team leading 4-3 in the eighth inning.

While not dominant in any of his performances — he walked one and was 10 feet from giving up a go-ahead home run to Curtis Granderson — Ferguson has competently gotten the job done. Even if the highlight reels might desire something more spectacular, you can’t ask for anything more from a result perspective from the nickname-less rookie.

The Most Consequential Hit-by-Pitch Ever?

On Tuesday night, with the bases loaded, two outs, and his team up 3-2 in the top of the eighth inning, pinch-hitter Mitch Moreland stepped to the plate. Roberto Osuna was on the mound for the Astros and had been a little knocked around. The batter immediately before Moreland had reached after being hit by a pitch, and the game was seemingly on the line. After falling behind 1-2, Moreland saw a 96 mph fastball headed toward him, and he smartly decided to wear the pitch.

Not the best outcome for the Astros. Jackie Bradley Jr. followed immediately with a grand slam to end all chance of a comeback, and the Red Sox took back home-field advantage with an 8-2 victory.

While the Red Sox were already leading and Bradley’s home run would grab more headlines, Moreland’s hit-by-pitch added nearly an identical amount of win probability to the Red Sox’ cause (.108 versus .109 for the Bradley grand slam). In fact, it was one of the most valuable hit batters in playoff history.

Most Consequential Playoff HBPs
Series Hit Batter WPA LI WPA/LI
2002 NLCS Jeff Kent 0.156 6.82 0.023
2018 ALDS Neil Walker 0.129 7.37 0.017
2003 ALDS Chris Singleton 0.126 7.16 0.018
2009 ALCS Carlos Ruiz 0.126 7.15 0.018
2011 ALCS Nelson Cruz 0.115 2.38 0.048
2004 NLCS Morgan Ensberg 0.115 4.61 0.025
2011 ALDS Adrian Beltre 0.109 2.85 0.038
2018 ALCS Mitch Moreland 0.108 1.42 0.076
2013 WS Shane Victorino 0.108 3.97 0.027
2007 NLCS Chris Young 0.103 4.91 0.021

Based just on pure win probability, Moreland added the eighth-highest fraction of a win for a hit by pitch since 2002. If you strip out the context of leverage index and look at context neutral win probability added (WPA/LI), it is the most important hit batter in baseball history. While it likely won’t live in history like Cleon Jones and the shoe polish ball in the 1969 World Series, it played a large role in sending the Red Sox past the Astros.

The Game Worth One-and-a-Half Wins

One of the odd things about WPA — at first glance at least — is that the win probabilities don’t add up to any number in particular. They don’t add to 1, they don’t add to 0; the only guarantee is that the difference between the winning and losing team’s pitching and batting WPAs will be 0.5. Teams’ combined pitching WPA could theoretically range from negative infinity to positive infinity, given a long enough game.

Usually, you see those large WPA values (both positive and negative) in extra-inning games. Games that involve several instances of a team scoring in the top half of an extra inning and the home team tying it in the bottom half will result in a game with a low WPA on both accounts. (See Game Six of the 2011 World Series for an example.) On the other hand, games with no scoring and many extra innings will have a high WPA for both teams. (Game Two of the 2014 NLDS between the Nationals and Giants is the prime example here.)

If you combine both team’s pitching WPA, you can generally get a good idea about how exciting the game was. Combined WPAs near 0 tend to be more generic games without too many swings or close calls. The farther away from 0 you get, the more exciting the game. For reference sake, the average for the absolute combined WPA (to account for the possible negative WPAs) since 2002 falls around 0.3.

Highest Combined WPA Since 2002
Series Loser Loser
WPA
Winner Winner
WPA
Total
WPA
2014 NLDS Nationals 1.032 Giants 1.532 2.563
2011 WS Rangers -1.405 Cardinals -0.905 -2.311
2017 WS Dodgers -1.320 Astros -0.820 -2.140
2012 ALDS Yankees 0.659 Orioles 1.159 1.818
2005 NLDS Braves 0.546 Astros 1.046 1.592
2014 AL WC Athletics -1.020 Royals -0.520 -1.539
2018 NL WC Cubs 0.514 Rockies 1.014 1.529
2018 NLCS Brewers 0.496 Dodgers 0.996 1.492
2003 ALDS Athletics 0.391 Red Sox 0.891 1.283
2013 NLCS Dodgers 0.378 Cardinals 0.878 1.256

Included here are many of the most exciting games of the 21st century, the games that will grace ESPN Classic one day. Game Four of the current Dodgers-Brewers series is one such example. That game’s combined WPA is eighth-highest in history, placing it among the most thrilling ever. Almost makes it a shame that they followed it up with the comparatively dull Game Four, with its combined WPA of 0.047 — 0.273 for the Dodgers and -0.227 for the Brewers.

We hoped you liked reading Adventures in Playoff Leverage and Win Probability Added by Stephen Loftus!

Please support FanGraphs by becoming a member. We publish thousands of articles a year, host multiple podcasts, and have an ever growing database of baseball stats.

FanGraphs does not have a paywall. With your membership, we can continue to offer the content you've come to rely on and add to our unique baseball coverage.

Support FanGraphs




Stephen Loftus is a Visiting Assistant Professor in Mathematical Sciences at Sweet Briar College in Virginia. In his spare time he usually can be found playing the pipe organ or working on his rambling sabermetric thoughts.

newest oldest most voted
cartographer
Member
Member
cartographer

Shouldn’t the last table have a column for the game number? These are individual games and not series, right?