Understanding the Wins Above Replacement (WAR) baseball statistic

There has long been an idea in theoretical physics of a single “theory of everything”, a single formula that contains the ultimate answer to the universe, describing all phenomena that has ever been or that ever will be.

Like an ephemeral alchemy, this formula has been sought fruitlessly for over a century and there are two trains of thought. The first says that it is out there and we simply haven’t found it yet. And the other says that it is a chimera, non-existent, and he who searches for it runs a fool’s errand.

In the world of baseball, statisticians are on a similarly quixotic quest, and while it is acknowledged to be lacking perfection, the current doyen around MLB circles is WAR.

What is WAR? How is it determined?

WAR stands for “wins above replacement” and it attempts to measure a player’s overall value to a team by positing how many wins he is worth when compared to a replacement-level player.

It is a complex system which is calculated differently for each defensive position. For example, for a shortstop and a first baseman with the same offensive numbers, the WAR would be higher for the shortstop because a replacement-level shortstop would have a lower level of production than his compatriot at first.

Position players

(batting, base running and fielding + adjustment for position + adjustment for league + the number of runs provided by a replacement-level player) / runs per win

Pitchers

Different WAR computations use either runs allowed per nine innings pitched or fielding independent pitching. Those numbers are adjusted for league and ballpark. Then, using league averages, the resulting numbers and his innings pitched total.

Unlike all other stats, which are either offensive or defensive stats, WAR attempts to posit a single number which can measure a player in the round.

This is the principal feature of WAR which the statisticians who evangelize about its virtues cite. It is also the main weakness of the statistic for its detractors.

Pure hitting or fielding stats focus on outcomes that have already happened; a strikeout, a base hit, a home run, a fielding error, a throwing error, etc. WAR, on the other hand, tries to synthesize other statistics to predict what may happen in the future. Also in attempting to find an adjustment figure that can level the field between various positions, WAR is forced to rely on a hypothetical replacement-level player, making the baseline effectively arbitrary.

WAR’s proponents will counter with the statement that no stat can be seen in a vacuum, that multiple areas must be taken into account, and since WAR does this, it is still the most reliable number that we have to measure a player’s overall contribution.

One thing is certain: for all of its virtues, WAR is far from a universal “theory of everything” style statistic. But that will not stop the true believers from searching.