War

[Why_Knott?]

Can someone explain to me how WAR is measured between pitchers and offensive players?

I understand the basic premise, but is it measured differently for the two groups of players? Does this mean the team wins X more games with this player vs a replacement regardless of the position? Or does it mean that if I had a replacement pitcher who wins 9 games, Clayton Kershaw wins 16?

[sangar]

Can someone explain to me how WAR is measured between pitchers and offensive players?

I understand the basic premise, but is it measured differently for the two groups of players? Does this mean the team wins X more games with this player vs a replacement regardless of the position? Or does it mean that if I had a replacement pitcher who wins 9 games, Clayton Kershaw wins 16?

If I understand correctly, it is position specific and does not translate well to inter-positional comparison.

[Aglets]

I'm pretty sure the only way to correctly start off a thread called 'War' is:

"What is it good for?"

[Skeletor]

I believe fangraphs has a primer on it. I'd give you a link but I'm on my phone.

[scOtt]

I'm pretty sure the only way to correctly start off a thread called 'War' is:

"What is it good for?"

You forgot...

HUH!

Good God y'all!

[Why_Knott?]

What is it good for?

[Why_Knott?]

If I understand correctly, it is position specific and does not translate well to inter-positional comparison.

So if a pitcher's war is 7.0, does that mean he wins 7 games more than a replacement player, or does it mean that the team wins 7 more games with said pitcher than replacement player?

[scOtt]

So if a pitcher's war is 7.0, does that mean he wins 7 games more than a replacement player, or does it mean that the team wins 7 more games with said pitcher than replacement player?

That's sort of along the lines of what I wonder...

Justin Verlander was 7.0 WAR this year. Curtis Granderson was also 7.0. Are they equal? Since it's positionally adjusted... does this make the final numbers all equal?

[CA-ORIOLE]

So if a pitcher's war is 7.0, does that mean he wins 7 games more than a replacement player, or does it mean that the team wins 7 more games with said pitcher than replacement player?

Good question.

They are valued by their respective positions, so it's wins above their respective positional replacement level player. In the case of pitchers, a SP and a RP have different replacement level values but aren't broken out like position players are with adjustments. A RP will just not get enough innings/appearances to match an SP. I'm not sure a 7 WAR SP (or RP) would approximate the same value as a 7 WAR hitter/position player or not. Theoretically, I guess it should but the use of completely different variables and the divergence of rWar and fWar for pitchers makes me think it may not be accurate. Because WAR is a counting stat, it's harder for even the top tier SP's to attain enough appearances/innings to compete with top tier position players.

That said, total WAR (pitchers and hitters) tends to correlate pretty well with actual wins. I'm not confident enough to say if that's because the individual stats are reasonably accurate or if it's a team dynamic.

Or does it mean that if I had a replacement pitcher who wins 9 games, Clayton Kershaw wins 16?

It's the theoretical percentage of games the team would win in his starts, not the difference in wins by the respective pitcher(s).

[eb45]

So if a pitcher's war is 7.0, does that mean he wins 7 games more than a replacement player, or does it mean that the team wins 7 more games with said pitcher than replacement player?

It means that the player was "worth" seven more wins than a hypothetical player whose stats were replacement level. If everyone on the team was replacement level except a 7.0 WAR pitcher, the team could be assumed to win about 47 games, assuming a replacement level of 40.

[Frobby]

That's sort of along the lines of what I wonder...

Justin Verlander was 7.0 WAR this year. Curtis Granderson was also 7.0. Are they equal? Since it's positionally adjusted... does this make the final numbers all equal?

Yes, theoretically, if the Tigers had a replacement-level player in CF and the Yankees had one at Starting P, and the teams traded Granderson for Verlander and then replaced each with a replacement player at their position, it would be a wash. The whole point of WAR is to try to have an "apples to apples" comparison of value.

[scOtt]

Good question.

They are valued by their respective positions, so it's wins above their respective positional replacement level player. In the case of pitchers, a SP and a RP have different replacement level values but aren't broken out like position players are with adjustments. A RP will just not get enough innings/appearances to match an SP. I'm not sure a 7 WAR SP (or RP) would approximate the same value as a 7 WAR hitter/position player or not. Theoretically, I guess it should but the use of completely different variables and the divergence of rWar and fWar for pitchers makes me think it may not be accurate. Because WAR is a counting stat, it's harder for even the top tier SP's to attain enough appearances/innings to compete with top tier position players.

That said, total WAR (pitchers and hitters) tends to correlate pretty well with actual wins. I'm not confident enough to say if that's because the individual stats are reasonably accurate or if it's a team dynamic.

It's the theoretical percentage of games the team would win in his starts, not the difference in wins by the respective pitcher(s).

I'm far from a stat expert, but that's the way I'm leaning. The everyday player has to be more valuable.

[Why_Knott?]

Good question.

They are valued by their respective positions, so it's wins above their respective positional replacement level player. In the case of pitchers, a SP and a RP have different replacement level values but aren't broken out like position players are with adjustments. A RP will just not get enough innings/appearances to match an SP. I'm not sure a 7 WAR SP (or RP) would approximate the same value as a 7 WAR hitter/position player or not. Theoretically, I guess it should but the use of completely different variables and the divergence of rWar and fWar for pitchers makes me think it may not be accurate. Because WAR is a counting stat, it's harder for even the top tier SP's to attain enough appearances/innings to compete with top tier position players.

That said, total WAR (pitchers and hitters) tends to correlate pretty well with actual wins. I'm not confident enough to say if that's because the individual stats are reasonably accurate or if it's a team dynamic.

It's the theoretical percentage of games the team would win in his starts, not the difference in wins by the respective pitcher(s).

By this logic, that means that the difference between replacement pitchers win total and Clayton Kershaw or Just Verlanders is less than 7 wins for the pitcher. Also, does that mean if Verlander's WAR is 7, that a replacement level pitcher wins 17 games?

I understand the WAR logic for positional players but I still dont understand it for pitchers.

[Skeletor]

By this logic, that means that the difference between replacement pitchers win total and Clayton Kershaw or Just Verlanders is less than 7 wins for the pitcher. Also, does that mean if Verlander's WAR is 7, that a replacement level pitcher wins 17 games?

I understand the WAR logic for positional players but I still dont understand it for pitchers.

When Verlander is credited with a win, he's not given 1 WAR. Rather, it's an estimation of his contribution to that win. The hitters and the defense get some credit too.

WAR for pitchers (fangraphs way) is dependent on DIPS. The theory that a pitcher can only control Ks, BB, and HR allowed.

[Why_Knott?]

Is there a stat that shows how many wins (total) a player is responsible for?