Would you trade Brandon Snyder to get Koji back at $4 mm for next year?

[DrungoHazewood]

I just thought that WAR is created to judge people by position right? It seems like it would work better if you could compare all closers to closers, set-up guys to set-up guys, 7th inning guys to 7th inning guys, but that's not how RP works, so you get different situations, and all kinds of variables. Then you can have a RP that has you know 45 IP and stellar numbers (think JJ), and it seems like they are getting getting similar WAR to guys pitching sub 4.0 ERA for 200IP.

But "7th inning guy" isn't a position. That just means he's probably the 3rd-best reliever on the team.

I think reliever WAR works because it's scaled to leverage index. If it was just scaled on the same scale as starters (who all have LI's near 1.00) then a 65-inning reliever would only rarely be worth more than 1/3rd of a 200-inning starter. But that's not true, since the late innings of close games have more impact on wins and losses, and you can control when good relievers pitch. So if Koji is pitching a disproportionate number of innings in key situations, WAR accounts for that.

[DrungoHazewood]

Yeah, with Type A's, it's a very bizarre and arbitrary formula using a whole bunch of stats. WAR is way different.

The Elias formula for free agent compensation was developed in 1981, in the negotiations that ended the strike. Bill James was only a couple years removed from publishing the Baseball Abstract on a mimeograph machine and hand-mailing it out. So the state-of-the-art was very primitive. It makes little sense for any position, but given the changes in reliever use over the past 30 years it's particularly ridiculous there.

WAR is much, much more rigorously constructed.

[allstar1579]

But "7th inning guy" isn't a position. That just means he's probably the 3rd-best reliever on the team.

I think reliever WAR works because it's scaled to leverage index. If it was just scaled on the same scale as starters (who all have LI's near 1.00) then a 65-inning reliever would only rarely be worth more than 1/3rd of a 200-inning starter. But that's not true, since the late innings of close games have more impact on wins and losses, and you can control when good relievers pitch. So if Koji is pitching a disproportionate number of innings in key situations, WAR accounts for that.

Right, that's kinda my point though, it can't differentiate between junk IP after a bad starter leaves in the 5th or 6th versus a guy that pitches only the 8th. In theory it would give them the same WAR if they had the same stats even though the 8th inning guy would be more practically important right? I'd feel better about it if it broke down things like pitching with a 1 run lead, in the 8th inning and gave that a higher weight than pitching down 5 runs in the 6th. Maybe it does, I dunno, I just always thought it didn't, kinda why I asked in here.

Kinda nitpicky but it's one of the problems I have with WAR, maybe because I could just never wrap my head around the mechanics of it.

[DrungoHazewood]

Right, that's kinda my point though, it can't differentiate between junk IP after a bad starter leaves in the 5th or 6th versus a guy that pitches only the 8th. In theory it would give them the same WAR if they had the same stats even though the 8th inning guy would be more practically important right? I'd feel better about it if it broke down things like pitching with a 1 run lead, in the 8th inning and gave that a higher weight than pitching down 5 runs in the 6th. Maybe it does, I dunno, I just always thought it didn't, kinda why I asked in here.

Kinda nitpicky but it's one of the problems I have with WAR, maybe because I could just never wrap my head around the mechanics of it.

Maybe I'm misunderstanding what you're saying, but I think you have it exactly backwards. WAR does take leverage into account. So if you're pitching in situations that are twice as valuable as an average inning you get twice as much credit in WAR. There's no way Jim Johnson gets 3 WAR this year and Guthrie gets less than that if leverage wasn't accounted for because Guthrie pitched almost three times as many innings.

[CA-ORIOLE]

Maybe I'm misunderstanding what you're saying, but I think you have it exactly backwards. WAR does take leverage into account. So if you're pitching in situations that are twice as valuable as an average inning you get twice as much credit in WAR. There's no way Jim Johnson gets 3 WAR this year and Guthrie gets less than that if leverage wasn't accounted for because Guthrie pitched almost three times as many innings.

One might think so, but the fangraphs page states fWAR for pitchers relies "totally on FIP". Not the case for rWAR. I'd assume the replacement level differences and IP contributions make up most of the differences in value. JJ's fWAR is 1.6. Guthrie 2.1. Not a big fan of Li (or rWAR in general). I lean more towards fWAR on this one.

[Frobby]

The Elias formula for free agent compensation was developed in 1981, in the negotiations that ended the strike. Bill James was only a couple years removed from publishing the Baseball Abstract on a mimeograph machine and hand-mailing it out. So the state-of-the-art was very primitive. It makes little sense for any position, but given the changes in reliever use over the past 30 years it's particularly ridiculous there.

WAR is much, much more rigorously constructed.

It would be interesting to do a ranking based on WAR and then compare it to Elias and see how many players' rankings were affected.

[Can_of_corn]

The Elias formula for free agent compensation was developed in 1981, in the negotiations that ended the strike. Bill James was only a couple years removed from publishing the Baseball Abstract on a mimeograph machine and hand-mailing it out. So the state-of-the-art was very primitive. It makes little sense for any position, but given the changes in reliever use over the past 30 years it's particularly ridiculous there.

WAR is much, much more rigorously constructed.

Is it a surety that the formula has not been changed over the years?

[CA-ORIOLE]

Right, that's kinda my point though, it can't differentiate between junk IP after a bad starter leaves in the 5th or 6th versus a guy that pitches only the 8th. In theory it would give them the same WAR if they had the same stats even though the 8th inning guy would be more practically important right? I'd feel better about it if it broke down things like pitching with a 1 run lead, in the 8th inning and gave that a higher weight than pitching down 5 runs in the 6th. Maybe it does, I dunno, I just always thought it didn't, kinda why I asked in here.

Kinda nitpicky but it's one of the problems I have with WAR, maybe because I could just never wrap my head around the mechanics of it.

Well, rWAR does that. fWAR doesn't (or at least marginalizes it). Personally I think the leverage issue is nonsense. Applying the same logic you'd have to give offensive credit (good or bad ) to hitters each year based on their clutch stats. Sure a run is more important in the late inning of a close game, but it wouldn't have been important if the SP or middle relief didn't keep you in the game in the first place. Until someone can show me that RP's can pitch more effectively on a consistent basis in these situations than their regular performance would be, I don't buy it. I believe there are some studies showing this may be in some cases, but it was pretty unconvincing as I recall.

[SrMeowMeow]

One might think so, but the fangraphs page states fWAR for pitchers relies "totally on FIP". Not the case for rWAR. I'd assume the replacement level differences and IP contributions make up most of the differences in value. JJ's fWAR is 1.6. Guthrie 2.1. Not a big fan of Li (or rWAR in general). I lean more towards fWAR on this one.

Wait, I'll take another look, but I am 95% sure that Fangraphs WAR does include leverage.

[CA-ORIOLE]

Wait, I'll take another look, but I am 95% sure that Fangraphs WAR does include leverage.

Maybe, I just went off the quote that fWAR pitching relies "totally on FIP". The page I read compared rWAR to fWAR and at the bottom indicated that leverage was one of the factors for pitchers WAR (but didn't state which one). It's clear rWAR does include leverage. There is a pretty high differential between JJ's rWAR (3.2) and his fWAR (1.6).

[SrMeowMeow]

Maybe, I just went off the quote that pitching relies "totally on FIP". The page I read compared rWAR to fWAR and at the bottom indicated that leverage was one of the factors for pitchers WAR (but didn't state which one). It's clear rWAR does. In any event there is a pretty hgh differential bewtweed JJ's rWAR (3.2) and fWAR (1.6).

Here it is: WAR and Relievers.

The average Leverage Index of a closer is about 1.8, meaning that each plate appearance is about 80 percent more important than an average PA. We give the closer credit for half of that, based on the principle of chaining. Because relief pitchers are mostly fungible, and can move from one role to another if needed, replacing a closer is not the same process as replacing a starter or a position player. If a team’s closer gets hurt, they do not then call up a replacement level reliever from the minors and use him to close out games.

[CA-ORIOLE]

Here it is: WAR and Relievers.

Thanks. That's an amazingly interesting article, but if you read through the fine line and comments, they (fWAR) really doesn't use Li and that's what many are criticizing. What they apparently do is consider it in some sort of abstract way to calculate the replacement level of a RP.

[SrMeowMeow]

Thanks. That's an amazingly interesting article, but if you read through the fine line and comments, they (fWAR) really doesn't use Li and that's what many are criticizing. What they apparently do is consider it in some sort of abstract way to calculate the replacement level of a RP.

Where are you reading that? From what I can tell, they explicitly do use LI.

[CA-ORIOLE]

Where are you reading that? From what I can tell, they explicitly do use LI.

The way I read it is, they utilize an Li factor (half of 1.8 in this case) across the board for all RP's and apply it to calculating the replacement level for a relief pitcher. That is much, much different than rWAR's individualized approach.

[DrungoHazewood]

Is it a surety that the formula has not been changed over the years?

Within the last few years someone has done regression analysis of the Elias formulas and they were all pretty basic stuff derived from purely traditional metrics.