Tuesday, February 22, 2011

WAR Explained

What is WAR?

Throughout the decades, baseball fans have sought to quantify a players value. Batting average, runs batted in, home runs, wins, losses... all metrics that have throughout the years been used to measure a players performance. As the years have gone on, new metrics like runs, earned run average, and stolen bases have been added to the lexicon.

Over the past few decades however, the desire to properly assign value to a players performance has been driven forward by leaps and bounds. Today, one of the easiest ways to value any player, fielder or pitcher, is a metric called Wins Above Replacement, or WAR.

Ok, give me more.

The basic premise is to determine how many wins a player adds to his team over a given season when compared to a hypothetical 'replacement level player.'  As a general rule of thumb, an 'average' full-time position player or starting pitcher would be worth around 2 WAR, less for bench players or relief pitchers (they take fewer at-bats and throw fewer innings).

That of course, is a VERY general rule of thumb.

A players value in the case of WAR is a by product of their runs created, and runs prevented. A third baseman for example, effects a teams ability to win a game both with his bat, and with his glove, so focusing solely on his OPS or batting average leaves a huge part of the 'vale' equation up in the air. We can utilize defensive metrics like UZR to help fill in those blanks, but it still leaves us with a somewhat incomplete picture of a players value.

Furthermore, WAR values players based on their specific position. We know that historically speaking the catchers position doesn't provide as much offense as the first base position. Therefor a catcher who provides a .800 OPS is far more valuable than a first baseman who does the same.

So how does it work?

WAR works by using runs created formulas. In turn, runs created formulas are made by calculating the sum total of a pitcher, hitter, or fielders contributions. The basic formulas can be a bit lengthy, but the actual math involved is no more difficult than basic addition, multiplication, and division. However, there are more advanced formulas that get more specific, and ultimately, those are the formulas used to derive the WAR values you'll find at Fangraphs or Baseball-Reference.

Fangrpahs, for example, utilizes something called wRAA, or weighted runs above average to compute their runs created for offense and UZR for defense. But don't be too thrown off, both are essentially just different methods of runs created that utilize a league average instead of the simple baseline that simple RC formulas use.

If all that's too technical, don't worry about it too much. I admit, it can be a bit overwhelming trying to determine exactly how much value should be given to any one event on a baseball field. Thankfully, we have enough data compiled over the past hundred and twenty years that we can calculate that. Furthermore, there are also enough mega-nerds who are intelligent enough to figure it all out for us.

In the end, this is what you really need to know.

For every runs above, or below replacement a player creates, he is credited with 0.1 WAR (or -0.1 WAR). That's because we know historically every ten runs created is worth a win. That seems like a ton, after all, games are frequently won 4-2 right?

Well, yes. And no. Let's look at the five AL Central teams from 2010, shall we?

Twins: +110 runs (11 WAR), actually 13 games above.
White Sox: +48 runs (4.8 WAR), actually 7 games above.
Tigers: +8 runs (0.8 WAR), actually even.
Indians: -106 runs (-10.6 WAR), actually 12 games below.
Royals: -169 runs (-16.9 WAR), actually 14 games below.

It doesn't line up PERFECTLY, but we can see why runs created formulas are good predictors of a teams success, even when the seemingly high 10 run per win amount is given. Of course, we also know that using a five team sample in a world where 30 teams play constitutes a mall sample size. Should we have looked at the league as a whole, you'd see that RC is even more accurate than our tiny, but still illustrative sample size shows.

How should I use it?

Well, one of the many great things about WAR is that you can utilize it in all kinds of ways. You can use it to make a projection (how many games will a given team win?) provided you have a good projection system in place. You can use it to determine whether or not a certain trade makes sense. Or you can use it to simply answer the age old question, "who was better?"

Of course, like any catch-all metric, it shouldn't be observed and used in a vacuum. Essentially any metric, no matter how good should be viewed through a carefully focused lens. Like a hitters OPS, or a pitchers ERA, can be inflated by a fluky BABIP mark, so too can WAR.

Still, when it comes down to answering the basic question, "how much was player x worth in 2010" there really isn't anything better.

Corey Ettinger is a proud contributor to both 612Sports.net, 312Sports.com, and 313sports.com. He also provides extensive analysis of the American League Central Division at his own blog, AL Central In Focus. Be sure to follow me on Twitter @Coreyettinger for the latest updates, random thoughts and general tomfoolery.

No comments:

Post a Comment