40 WAR

[RedsManRick]

This is another reason why WAR should not be a mainstream stat. It doesn't tell the whole story. You made a great argument below to Sea Ray which tells how it is distorted. Then you also add " distributional luck" as a ten game swing either way. WAR may have some value, but the way some preach it as gospel is bizzare.

It definitely depends on what you're trying to do with it. WAR is not the best at explaining historic team-level outcomes; that's what actual wins and loses are for. But if we're looking forward, the "distributional luck" piece isn't predictable. You don't plan for having a great record in 1-run games or having a disproportionate amount of clutch performance. You build a team with the highest baseline level of ability to produce and prevent runs. And we don't have a better way of assessing that than WAR, broadly speaking.

If your bar for whether a not a stat should be mainstream is whether or not it tells "the whole story", then we shouldn't look at any stats ever. But if you're trying get a handle on how "good" your team is and whether or not it's likely to be competitive, WAR is the best framework we have for doing that. As a starting point, at minimum, and likely still quite a bit better than whatever ad hoc process you'd attempt to put together absent WAR.

[Assembly Hall]

Where is Billy Beane when you need him?

[M2]

It definitely depends on what you're trying to do with it. WAR is not the best at explaining historic team-level outcomes; that's what actual wins and loses are for. But if we're looking forward, the "distributional luck" piece isn't predictable. You don't plan for having a great record in 1-run games or having a disproportionate amount of clutch performance. You build a team with the highest baseline level of ability to produce and prevent runs. And we don't have a better way of assessing that than WAR, broadly speaking.

If your bar for whether a not a stat should be mainstream is whether or not it tells "the whole story", then we shouldn't look at any stats ever. But if you're trying get a handle on how "good" your team is and whether or not it's likely to be competitive, WAR is the best framework we have for doing that. As a starting point, at minimum, and likely still quite a bit better than whatever ad hoc process you'd attempt to put together absent WAR.

I'm going to disagree with the notion of using WAR to look forward. The late 2000s Seattle Mariners conducted a failed experiment in trying to add WAR. They bought into WAR and tried to add like it was a Lego stack. The mistake there was not recognizing WAR reflects how a player's set of skills performed for a season, but that the player doesn't necessarily own that WAR moving forward. Tucker Barnhart is a good example. He was a 3.4 WAR guy in 2017, but a 0.9 WAR guy this year.

I think it's more a case that you add talent and, if it plays well, then you get higher WAR totals.

[BCubb2003]

I'm going to disagree with the notion of using WAR to look forward. The late 2000s Seattle Mariners conducted a failed experiment in trying to add WAR. They bought into WAR and tried to add like it was a Lego stack. The mistake there was not recognizing WAR reflects how a player's set of skills performed for a season, but that the player doesn't necessarily own that WAR moving forward. Tucker Barnhart is a good example. He was a 3.4 WAR guy in 2017, but a 0.9 WAR guy this year.

I think it's more a case that you add talent and, if it plays well, then you get higher WAR totals.

Interesting. Sounds like “clutch.” But isn’t there a lot of discussion about career WAR? And how guys like Trout and Votto compare to the greats? Probably can’t go astray measuring Trout or Votto, but still, if OBP and OPS aren’t random, to name a few, why would WAR be?

Is there a way to describe WAR so that someone like George Grande could understand? The player you don’t want to beat you? Just knows how to win? The straw that stirs the drink?

[kaldaniels]

Interesting. Sounds like “clutch.” But isn’t there a lot of discussion about career WAR? And how guys like Trout and Votto compare to the greats? Probably can’t go astray measuring Trout or Votto, but still, if OBP and OPS aren’t random, to name a few, why would WAR be?

Is there a way to describe WAR so that someone like George Grande could understand? The player you don’t want to beat you? Just knows how to win? The straw that stirs the drink?

None of the above on those.

I use it exactly as it states what it is. Let’s say you have a 4 win player. Over a season, that 4 win player did/did not do stuff on the diamond that caused his team to win 4 games more, than had that team just signed John Doe off the waiver wire. Now of course there are error bars on that, but it’s all done from a historical perspective as far as producing those 4 wins go.

And when you look at the WAR summed up for any given team, you will have a good idea of that teams W-L record and/or their playoff chances.

8 of the top 9 WAR teams made the postseason this year. The Braves were the outlier at 33 WAR. (Really - how they beat the Nats in the East, I don’t know)

But I agree with M2 - there is an art to building a team - you don’t just stack WAR.

[BCubb2003]

None of the above on those.

I use it exactly as it states what it is. Let’s say you have a o4 win player. Over a season, that 4 win player did/did not do stuff on the diamond that caused his team to win 4 games more, than had that team just signed John Doe off the waiver wire. Now of course there are error bars on that, but it’s all done from a historical perspective as far as producing those 4 wins go.

And when you look at the WAR summed up for any given team, you will have a good idea of that teams W-L record and/or their playoff chances.

8 of the top 9 WAR teams made the postseason this year. The Braves were the outlier at 33 WAR. (Really - how they beat the Nats in the East, I don’t know)

But I agree with M2 - there is an art to building a team - you don’t just stack WAR.

It sounds more like Pythag now.

If Votto played for a team that won more games, would his WAR be higher? If he went 3 for 4 with 2 home runs but his team lost, he didn’t cause the team to win. Or is it a theoretical “these are the actions that typically cause wins” even if the team blew it?

[M2]

Interesting. Sounds like “clutch.” But isn’t there a lot of discussion about career WAR? And how guys like Trout and Votto compare to the greats? Probably can’t go astray measuring Trout or Votto, but still, if OBP and OPS aren’t random, to name a few, why would WAR be?

Offense can be highly variable. Not so much for the super elite, but even in that crowd check out Mookie Betts and Christian Yelich, your likely MVPs, from last year to this year.

Add in defense, which can be all over the place, and base running and year-to-year differences in positional adjustments, and there's lots of moving parts to cause variance. So I wouldn't term WAR as random," but it can be volatile.