Question about WAR

[bp4baseball]

Hi,
I really enjoy the statistical side to baseball and had two questions about WAR (wins above replacement) that I thought maybe someone could answer.

1. How many games is a players WAR based off. I'm pretty sure it's 1 whole season, but wanted to make sure. If you're good for 2 wins above a replacement level, is that in 1 month, 1 year, etc.

2. When people discuss a teams WAR is that just a sum of each players WAR? And if so, is that the 25 man roster or 40 man roster?

Thanks.

[AC]

Re: Question about WAR

WAR is based off the current season, the same way HR and RBI would be.

It is based off of the combined WAR of every player that has played a game for the team the current season.

[bp4baseball]

Re: Question about WAR

WAR is based off the current season, the same way HR and RBI would be.

It is based off of the combined WAR of every player that has played a game for the team the current season.

Got it...Thanks.

[jvalverde88]

Re: Question about WAR

Please tell me I'm not the only who was hoping the question was "WAR what is it good for?"

[55]

Re: Question about WAR

Absolutely NOTHIN'!

[Redacted01]

Re: Question about WAR

Absolutely NOTHIN'!

I guess you can stick to your batting average and RBI.

[AC]

Re: Question about WAR

Absolutely NOTHIN'!

PITCHER WINZZZ!!!

[snepp]

Re: Question about WAR

I guess you can stick to your batting average and RBI.

*whoosh*</i>

[lonewolf371]

Re: Question about WAR

WAR is based off the current season, the same way HR and RBI would be.

It is based off of the combined WAR of every player that has played a game for the team the current season.

It is kind of like HR and RBI, but I wouldn't say it's based on the current season. You simply accumulate WAR over your career by making certain types of plays. Based on linear regression, you say a home run is worth a certain number of runs, let's say 1.7 on average. It's greater than 1 because home runs with players on base usually knock in a bunch of other runs. They get this number using a technique called linear regression.

You take all the runs in the league and find out how many runs a replacement player contributes. Replacement level is generally thought to be players in the 30th-40th percentile, i.e. 60-70% of players in MLB baseball are better than replacement level.

Once you find what a replacement player contributes, you subtract that from what the player you're interested in does. That's how many runs that player contributes above replacement.

To get to wins, you divide by ten (a "win" is thought to be worth ten "runs").

So basically, as you pile up singles, doubles, HRs, and stolen bases your WAR will get bigger. You can combine multiple seasons and see who had the best career in history (as measured by WAR). Right now Fangraphs says that Babe Ruth is the best player ever by WAR, contributing 163.8 WAR. Barry Bonds is #2 with 163.6, and Ty Cobb is #3 with 150.0. These numbers are adjusted for the position the player played, the year he played in, and the league he played in (National or American).