Win Probability Added Leaders Through the LDS
During the postseason, Win Expectancy charts become ubiquitous, because each play, misplay, decision and comeback is magnified in its importance in front of a national TV audience. While Win Expectancy (WE) and Win Probability Added (WPA) aren’t great stats to evaluate players, they are a tool to understand how the dynamics of how a game changes from the first pitch to the last out.
For those not all too familiar with Win Expectancy, our library has a good entry and the interpretation can be boiled down to
If a team is losing and has a 24% win expectancy, only 24% of teams in similar situations in the past have ever come back to win.
So using historical data and the current inning, score, outs and runners on base, WE tells you what percentage of teams have won given those circumstances. These numbers aren’t a prognostication, since anything can still happen, but they give an estimate of what you might expect from the situation.
Win Probability Added is derived from Win Expectancy — being the difference from one play to the next. For example, The batter/runner is given credit for a hit, while the pitcher on the mound will be debited an equal amount for that hit. Plays that dramatically swing the score late in the game with two outs in the inning generally have the highest WPA. WPA is written out like batting average (.000), but it should be interpreted in the same way as win expectancy (0.0%). A play with a .360 WPA increases the WE +36.0%.
Below is our standard WE chart combined with the signed* WPA chart. The WE chart is the running total of the WPA chart. The top chart shows the sum of all the plays until a certain point in the game, and the bottom chart shows the change in WE for each play, which is also the signed WPA.
Top Players
Batters
Now with the basics out of the way, we can make some WPA leaderboards for this postseason. First, batters through the end of the LDS.
This represents the sum of all the at-bats players had from the Wild Card games through the LDS round. Let’s look at Kyle Schwarber to understand why he is near the top of the leaderboard. He’s had 15 plate appearances and three of them have been home runs. Now, Colby Rasmus actually has four home runs this postseason. Even though Rasmus has more home runs, Scwarber’s three home runs affected the WE more at the time they were hit than Rasmus’ home runs did. In fact, Rasmus’ highest WPA play was actually a double. While home runs typically garner the most WPA, Justin Turner tops the chart and didn’t homer once in his 20 plate appearances in the NLDS.
Pitchers
Pitching WPA is a little bit of a different animal since it’s more constant and it depends when the pitcher has come into the game. For example, when a relief pitchers enters with runners on base and no outs then retires the side without giving up a run, his WPA is much larger than if a pitcher started and finished that inning. So there are a few ways to rack up WPA as a pitcher: pitch a lot of low-scoring innings, accomplish the same thing in fewer, but higher leverage situations or by finishing the game, thus removing any uncertainty of losing in a low-scoring game.
We can see those principles play out, as the list is mixed with a combination of starting and relief pitchers. Dallas Keuchel faced 58 batters compared to Jake Diekman, who faced just 19, and they both appear near the top the pitching leaderboard. Keuchel was used both as a starter and a reliever in a low-leverage situation. Since the Royals already had a two-run lead going into the eighth inning, the three-run home run Keuchel gave up to Kendrys Morales didn’t hurt his WPA, even though that home run cemented the Royals’ lead on the way to win the ALDS.
Top Plays
Batters
Looking at WPA on the scale of the individual play, we can find which plays affected the WE the most. For batters most of these are homers and late in the game.
| Game | Inning | Play Description | WPA | |||
| 1 | 10-9 TEX at TOR | 14 | Hanser Alberto singled to center (Fliner (Liner)). Rougned Odor scored. Chris Gimenez advanced to 3B. Hanser Alberto advanced to 2B. | .366 | ||
| 2 | 10-14 TEX at TOR | 7 | Jose Bautista homered (Fliner (Fly)). Ryan Goins scored. Josh Donaldson scored. | .346 | ||
| 3 | 10-10 NYN at LAN | 7 | Adrian Gonzalez doubled to right (Liner). Chase Utley scored. Howie Kendrick scored. Adrian Gonzalez advanced to 3B. | .327 | ||
| 4 | 10-12 LAN at NYN | 2 | Curtis Granderson doubled to center (Fliner (Fly)). Travis d’Arnaud scored. Wilmer Flores scored. Juan Lagares scored. | .323 | ||
| 5 | 10-13 SLN at CHN | 2 | Javier Baez homered (Fliner (Fly)). Kyle Schwarber scored. Jason Hammel scored. | .319 | ||
| 6 | 10-12 KCA at HOU | 8 | Kendrys Morales reached on fielder’s choice and error to shortstop (Grounder). Ben Zobrist scored on error. Lorenzo Cain scored on error. Eric Hosmer advanced to 3B on error. Error by Carlos Correa. | .311 | ||
| 7 | 10-12 SLN at CHN | 5 | Kris Bryant homered (Fliner (Fly)). Jorge Soler scored. | .252 | ||
| 8 | 10-9 TEX at TOR | 8 | Mike Napoli singled to right (Grounder). Delino Deshields scored. | .232 | ||
| 9 | 10-14 HOU at KCA | 2 | Luis Valbuena homered (Fly). Evan Gattis scored. | .214 | ||
| 10 | 10-13 SLN at CHN | 6 | Anthony Rizzo homered (Fliner (Fly)). | .208 | ||
Two plays in Rogers Centre top the list. Hanser Alberto’s single scored the go-ahead run in extra innings and with two outs, and then Bautista’s home run emphatically broke a seventh-inning tie.
Pitchers
The biggest WPA swings for pitchers occur when the inning ends, the score is close and there are runners on base.
| Game | Inning | Pitcher | Play Description | WPA | |||
| 1 | 10-9 TEX at TOR | 11 | Aaron Loup | Josh Hamilton flied out to center (Fliner (Fly)). | .137 | ||
| 2 | 10-7 CHN at PIT | 6 | Jake Arrieta | Starling Marte grounded into a double play to shortstop (Grounder). Andrew McCutchen out at second. | .130 | ||
| 3 | 10-14 TEX at TOR | 7 | Cole Hamels | Ben Revere reached on fielder’s choice to first (Grounder). Dalton Pompey out at home. Kevin Pillar advanced to 3B. Ryan Goins advanced to 2B. | .128 | ||
| 4 | 10-9 HOU at KCA | 9 | Wade Davis | Carlos Gomez picked off. | .127 | ||
| 5 | 10-9 TEX at TOR | 12 | Shawn Tolleson | Ryan Goins grounded out to second (Grounder). | .127 | ||
| 6 | 10-12 KCA at HOU | 8 | Tony Sipp | Mike Moustakas struck out swinging. | .122 | ||
| 7 | 10-11 TOR at TEX | 4 | Martin Perez | Dioner Navarro grounded into a double play to second (Grounder). Troy Tulowitzki out at second. | .112 | ||
| 8 | 10-9 HOU at KCA | 6 | Josh Fields | Alex Gordon struck out swinging. | .109 | ||
| 9 | 10-15 NYN at LAN | 3 | Jacob deGrom | Kiké Hernandez grounded into a double play to pitcher (Grounder). Yasmani Grandal out at second. | .105 | ||
| 10 | 10-11 TOR at TEX | 6 | Chi Chi Gonzalez | Chris Colabello grounded into a double play to first (Grounder). Josh Donaldson out at home. Jose Bautista advanced to 3B. Edwin Encarnacion advanced to 2B. | .104 | ||
Jake Arrieta had the bases loaded with one out in the Wild Card game. The Cubs had a four-run lead, so the WE was only 17.8%, but the difference in the run-scoring potential before Starling Marte’s at-bat and at the end of the double play brought the Pirates WE down to 4.8%. Aaron Loup getting Josh Hamilton to fly out also produced a high WPA, and that’s because it occurred during a tie game in extra innings.
*I use signed WPA because WE charts use a set orientation with regards to teams, and WPA is from the player’s perspective. Signed WPA is WPA multiplied by -1 for events in the top of the inning.
I build things here.
Ok, thanks.
No mention of fielding? I thought I saw a commenter point out that Andrus’ fielding in Game 5 alone was like -.22.
“The batter/runner is given credit for a hit, while the pitcher on the mound will be debited an equal amount for that hit”
The classic issue of weighting pitching and defense in run prevention; I guess for a stat such as WPA where all we’re looking at is the black and white outcome of an on-field event, does it make sense to debit all win probability of a “hit” to the pitcher? Similarly if a pitcher gets an out with the help of a fielder, does he get all the credited WPA?
Yea I would really be interested to see fielding, at least errors, considered. As an Astros fan, I loved Correa’s two HRs in game 4, and would love to know whether or not his big error in the 8th washes that out in this model.
You’d have to redo how FanGraphs processes WPA to add in fielding, since we don’t track the fielder in our database. You could look at just errors or specific plays and then aggregate them. But you hit upon the big issue throughout baseball analytics, the weighting. Stats like DRA from BP use advanced mixed modeling to tease that out. Unfortunately, no one has approached WPA like that.
Great, now do leverage (and WPA/LI leaders)? The Dodgers-Mets Game 5 leverage graph looks like just what it felt like, a continuous 9-inning-long heart attack.
That game was killer on my insides. Worst heartburn I’ve had since I gave up liquor
That’s pretty much how it felt, too.
Same type of gibberish as when you here “only 39% of teams who are down 2 games, heads on the road, on a Tuesday, score over 4 runs during their next game”.
That, and this, are all horseshit.
No need to be rude. If you don’t like WPA, just skip this article and read the next free article.
Hay Painfully, I know horseshit, I’m full of horseshit, my stall keeps getting a fresh supply of horseshit.
Baseball stats are not horseshit.
I didn’t say baseball stats were useless, now did I ******* But with win probability, where’s the value? If historically, 24% of teams win while in this particular circumstance, so what? What did all these teams do that was in common to win? Nothing. One bad throw, and a percentage jumps up 40%.
The whole win probability phase of baseball stats is goin wayyyyy overboard.
there is definitely something to be said about WPA based on how quickly it can experience a huge swing, which does suggest that we are likely relying too strongly on something which is actually relatively flimsy.
Thanks so much JABO.
Your example suffers from small sample size, WPA uses years of games played across the league. WPA tells you how likely an outcome is in the context that game compared to a volume of other games. People do this intuitively all the time. WPA is only different because it can capture more games and puts some mathematical rigor behind it.
Yeah, you’re the dick who made the Greinke comment. We know not to listen to you.
Your attitude and writing skills are horseshit.
Good stuff, Sean.
Thanks!
does baserunning count? Murph’s sb had to have moved the needle.
It made my needle move.
Alternatively you could calculate, in the spirit of context and WPA, series-WPA by adjusting games by their leverage in winning the series. I.e. game 5 has a value of 1 because winning that game gives you a 100% chance of winning the series. Game 3 of a 2-0 series only has a 1/4 game-to-series leverage, since if the series continues there is still 3/4 chance (using the usual wpa that games start 50-50) of being won by the team ahead 2-1. So the wpa in that game would be multiplied by 25% in the leaderboard. This would give extra credit to particularly meaningful plays, in terms of determining the series outcome, since winning the series is what matters.
Does the batter always, never or sometimes get credit if the defence makes an error, like on the Morales play?
Batter +0.xxx, pitcher -0.xxx. The way WPA is currently done, only the batter and pitcher are credited or blamed.