Game 100, by WPA
teh awes0mest play: Nyjer Morgan's 2-out, 2-run single in the 5th (+0.129)
teh suxx0rest play: Collin Balester gives up a game-tying one-run single in the 1st (-0.099)
MVP: Cap'n Morgan (+0.181), Battlestar (+0.128)
LVP: Alberto Gonzalez (-0.139)
This is a most unfamiliar situation--Nats out to an early lead, which they managed to regain and hang onto, mostly deciding the game early!
(Idea stolen from Amazin' Avenue. WPA adds up to +0.5 for a win, -0.5 for a loss.)
7 comments
|
0 recs |
Do you like this story?
Comments
I dont understand fangraphs
Sorry… I even googled a wikipedia type definition… why? what is it of? IS it of web hits? karma? good energy apportioned out by a the cosmos?
Why the hell cant someone appropriately label the X and Y axis of the damn thing!??
"On this team, the difference between Clippard and Julian Tavarez is like the difference between a 6-2 loss and a 9-2 loss." -- Chico Harlan
I believe it's all win probability, the effect each event has on altering the statistically predictable outcome of a game...
I have no idea what I’m talking aboot????
Label the axisesees…..
Vivian Jaffe: "Have you ever transcended space and time?"
Albert Markovski: "Yes. No. Uh, time, not space... No, I don't know what you're talking about."
by Patrick Reddington on Jul 29, 2009 9:35 AM EDT via mobile up reply actions
Here's the deal with WPA...
Baseball is game with a countable number of defined states that the game can in at any time: 8 combinations of men on base, 54 possibilities for total number of outs (starting at out #1 at top of the first, no out and ending at out #54, bottom of the 9th, two out), and the home team can be ahead or behind by some number of runs. If you look at a database of thousands of games, you can take every game where the home team was in each state and look at how many times they won. This gives you a “Win Percentage” for each state that the game is in. So you can build a table with things like, “When the home team is ahead by two runs with a runner on 2nd and one out in the bottom of the 5th, they win the game 58.1% of the time” (I’m making up the numbers). The fun part is when you graph the Win Percentage with each batter (which is what the graph is).
So if you take the example above, and the batter strikes out, now you go the “Home team ahead by two runs with a runner on 2nd and two outs in the bottom of the 5th” state, and maybe the WP drops to 56.1%. In this case the “Win Percentage Added” (WPA) would be .020 (.561.581). Broadly, whenever something good happens (run scores, runner advances, guy gets on base) the WP goes up; whenever something bad happens (make an out), the WP goes down.
The WP starts at .500, with each team having a 50% chance of winning. By the end of the game, the WP will have gone all the way to either 100% (home team wins) or 0% (visitor wins)—so the total WPA will be either +0.5 or -0.5. The default is that “good for home” is positive and “good for visitor” is negative, but I use the “good for my team is positive” standard when I call out plays and player totals.
And that’s the other fun part: you can look for the biggest WPA for each play, and call ‘em out. You can also add up the WPA for every play that each player is a part of and see how they add to the total. Usually the winning/losing pitcher has a big total WPA because they get credit/penalty for every out/hit that other side makes. Batters gets a big WPA for making a late-inning, go-ahead hit (like AK’s recent walkoff—getting the big hit in the critical spot shifted the odds more than Stammen’s entire outing!).
"It's not a secret, you don't need to be an expert on math to know that walks plus errors equals runs...." --Mannyger Acta
So its all about probability?
"On this team, the difference between Clippard and Julian Tavarez is like the difference between a 6-2 loss and a 9-2 loss." -- Chico Harlan
X-Axis = "Plays", Y-Axis = "Odds of home team winning"
The X-Axis is confusing, because each plays is a new tick, but it’s labelled by inning. An inning might have any number of plays from 6 (both teams get out in order) to OMG-WTH-ROUT!!! Keep in mind that a “play” is anything that results in change of the game “state”: anything that makes an out and/or changes where/how many runners are on base. Usually this will correspond to the outcome of a plate appearance, but a stolen base or caught stealing/pickoff counts too (since it changes where people are on base and/or makes an out).
"It's not a secret, you don't need to be an expert on math to know that walks plus errors equals runs...." --Mannyger Acta
Thank you
Sorry, im sure all this was taught in school for you guys… but i really wanted to understand.
"On this team, the difference between Clippard and Julian Tavarez is like the difference between a 6-2 loss and a 9-2 loss." -- Chico Harlan
So looking above the question has to be?
What happened just before the 4th inning…that swung the game from the Brewers back towards the Nats? Just before Dunns HR… THAT was the defining moment…
"On this team, the difference between Clippard and Julian Tavarez is like the difference between a 6-2 loss and a 9-2 loss." -- Chico Harlan

by 



















