clock menu more-arrow no yes

Filed under:

Win Expectancy Added

New, 18 comments

At last, the moment five or six of you have been waiting for - the final season tallies. I'm not going to throw in much commentary; by this point, the numbers should really speak for themselves.

If you're unfamiliar with win expectancy, and its statistical derivatives, make sure you read this explanation before anything else.

The cumulative spreadsheet can be downloaded here, for your perusal.

Column Headings:

WEA: Cumulative Win Expectancy Added for the entire season.
WINS: Number of wins above or below what an average player (WEA of 0.000) could be expected to produce in a season. The Seattle Mariners, as a team, had a total WEA of -18.834, and finished 24 games below .500. Dividing -18.834 by 24 yielded a five-digit number, -0.78475, which is the total WEA necessary to be "worth" one loss. The absolute value (0.78475), then, is the total WEA necessary to be worth one win. WEA/0.78475 = WINS.
WEA AR: Win Expectancy Added Above Replacement, where the contribution made by a replacement-level player was estimated by using the worst contribution of any player on the team. This player, incidentally, was Adrian Beltre. So, WEA AR = WEA - (-3.017).
WINS AR: Wins Above Replacement. Same formula as WINS above, only showing a player's value above a replacement player, rather than an average player.

Your 2005 Seattle Mariners MVP? None other than Richie Sexson, which shouldn't come as much of a surprise to anyone. What's interesting is that he's followed by Felix Hernandez, who only made 12 starts, Randy Winn, who didn't have great raw numbers and was dealt at the deadline, and George Sherrill, a situational lefty who found his way into 19 games. And then...wait, what? Jamal friggin' Strong? You better believe it. This thing is vulnerable to small sample sizes, and a lot of Strong's contribution came from a run-saving throw from the outfield to catch Brian Roberts at the plate in an early September game. That play alone was worth 40% of a win.

And way at the bottom, below total disasters like Joel Pineiro, Matt Thornton, and Miguel Olivo? Our other big-money acquisition, Adrian Beltre, who rode a bunch of ugly strikeouts and groundouts with runners in scoring position to the lowest rating on the team, almost four wins below average. There's nowhere for this to go but up, people. There's no way Adrian Beltre is half a win worse than anyone else on the 2006 Mariners.

WEA per Plate Appearance:

Note: Because I didn't make hitting and fielding separate categories, defensive plays are mixed in there, skewing a few of the smaller sample size numbers. Like, say, Strong's. What's important is to look at the guys who actually got to swing the bat more than a few times during the season. Over a full year of 600 plate appearances, Scott Spiezio would have been worth almost 19 wins below average.

WEA per Game, Hitters:

WEA per Inning Pitched:

Have I told you how much I love George Sherrill?

WEA per Game, Pitchers:

It's important to remember that Win Expectancy Added has minimal predictive value going forward - it serves as a rating of past player performance, and never pretends to say anything about the future. Ideally, we could get these kinds of numbers tracked for every player in baseball, finally providing a definitive answer to the question of who was really the most valuable player in the league, but until then, we'll just have to make do with what we have. And that's that Richie Sexson had a spectacular 2005 season, while Adrian Beltre's was one to forget.

In 2006, I'll be using a new tool for game win expectancy charts and spreadsheets, which will help to eliminate the problems inherent in Christopher Shea's generator. Issues with sample size, ballpark, and general era will be fixed, and the final numbers will be more accurate as a result. I'm also planning on separating a player's defensive contributions from what he does at the plate. If you have any other suggestions, by all means, I'm listening. I have just as much to learn about this stuff as anyone else.