This is our old blog. It hasn't been active since 2011. Please see the link above for our current blog or click the logo above to see all of the great data and content on this site.

10 for 10: #4 Win Expectancy and Run Expectancy

Posted by Sean Forman on March 26, 2010

This is the fourth of ten features we are adding for our 10th anniversary.

The first place I saw Win Expectancy (WE) was when I was writing for the Big Baseball Annual back in the late 1990's. Doug Drinen, who has since moved on to become one of the leading NFL analysts and creator of, wrote a series of articles on reliever usage that built on the original Mills Brothers work on Win Expectancy (WE). Doug even included a WS Game WE Graph in one of the books. I always have thought it was an elegant measurement of what happened during the game and I'm excited to now have it on the site.

Tom Tango was very helpful in working out the details for implementation. His run environment adjusted numbers for Win Expectancy, Leverage, Base-Out Run Expectancy and Base-Out Leverage are the basis for the numbers you see on the site.

A quick rundown of what is on the site. WE and RE appear for seasons for which we have play-by-play, 1952-2009. The data is only 100% complete for 1974-present. For the earlier seasons, we are missing some games for each season. A full listing of our coverage is here.

I hope you are enjoying these improvements to the site as much as I'm enjoying putting them together. Be sure to three earlier features as well. Questions and comments are always welcome. I'm hoping update #5 will come next week, but given the onrushing advance of Opening Day it may occur after first pitch.

ERA+ update: yesterday, I mentioned that we erroneously changed the definition for ERA+. I am still considering a long-term change, but for the short-term have changed it back. I apologize for the confusion.

13 Responses to “10 for 10: #4 Win Expectancy and Run Expectancy”

  1. Charles Saeger Says:

    Oh, cool. Can I do something like get the LI of all saves in a given year?

  2. birtelcom Says:

    This is great, great stuff. Searching out these stats with PI makes them far more useful for historical research than they have ever been at any other site. Who would have realized that in the top 4 Mets pitchers in career WPA for the team, along with Seaver, Koosman and Gooden you would also find Armando Benitez?

    There will be some oddites along the way with the manner baserunning is treated for WPA purposes. Just looking at some random Mets boxscores, here's one: That's the Mets'second win in franchise history. The most valuable performer for the Mets in this game, as calculated by most WPA, is shown to be John DeMerit. DeMerit was in the game solely as a punch runner -- entered the game as a PR at first base, moved to second on a walk, moved to third on a ground out toward first base and scored on a wild pitch. Demerit gets .294 WPA for that sequence, more than any other Met netted in that game, which the Mets won 8-6. The crediting of wild pitches may need to be re-thought. Does the base runner really deserve WPA credit for a wild pitch? Maybe the batter, whose talents were perhaps sufficently impressive that the pitcher risked a wild pitch, deserves the credit more than the runner?

  3. birtelcom Says:

    Ah, typos. Pinch runner, of course, not punch runner. That's Pete Rose in this game:

  4. Chris J. Says:

    Is the batting runs/wins thing new to PI as well, or did I just not notice it before?


  5. Sean Says:

    It's new as well

  6. Mike Sandler Says:

    Sean, Thanks for putting batting runs in as a sortable stat., it's much appreciated. I'm thinking there's more information on Baseball Reference than the rest of the internet combined!! Fantastic site for us Baseball nuts.

  7. Dave Says:

    So which of these new stats is the most important in determining the player's worth to the club during a game and/or during a season?

  8. Johnny Twisto Says:

    Dave, I don't think there's an easy answer to that. They sort of measure different things, or they measure similar things in different ways. I don't think any is "most important" but you might look at different numbers depending on the question you're trying to answer, or look at all of them to get different perspectives on the player's performance/value.

  9. Bill Says:

    What are the differences between the WPA here compared to the numbers over at

  10. Sean Forman Says:


    Please read the about page.

    I am not 100% certain, but I'm fairly certain it relates to differing run environments used.

  11. Bill Says:

    Thanks for clearing that up Sean!

  12. Baseball-Reference Blog » Blog Archive » Card of the Week: 1967 Topps #122 George Brunet Says:

    [...] Now that WPA is widely available on, it's time to start using it. I feel it is perhaps the most useful stat around today. If you're not familiar with the stat, check out Sean's post announcing this addition. [...]

  13. Baseball-Reference Blog » Blog Archive » More on walking the leadoff hitter Says:

    [...] thought I'd go ahead and research this a bit more, using Win Expectancy [...]