Why WAR is the worst stat ever

[notoriousgod71]

Y! SPORTS

 

 

Punto > Kershaw

[Shane Wahl]

Can you explain what the issue is here? Punto is strangely having a RIDICULOUS season. Maybe OPS is horrible too since he has an OPS of .880???? How is WAR not measuring his performance right now?

[ThePuck]

That's easy...it's not.

[notoriousgod71]

55 innings of 1.62 ERA and .90 whip > 65 AB of .880 OPS

[snepp]

1. WAR uses FIP, not ERA (at least on Fragraphs)

 

2. Kershaw's WAR is higher than Punto's (on both FG and BRef)

 

 

I guess that kind of renders this entire thread moot, no?

[notoriousgod71]

1. WAR uses FIP, not ERA (at least on Fragraphs)

 

2. Kershaw's WAR is higher than Punto's (on both FG and BRef)

 

 

I guess that kind of renders this entire thread moot, no?

 

 

Thank God. Wonder what stupid calculation that article was referencing.

[snepp]

I didn't even read the article.

 

I just assumed that if it was on Yahoo that it was awful.

[snepp]

On this topic though, Kershaw's BRef WAR is the 3rd highest in the NL, 5th highest in the ML.

[Brad Swanson]

Wasn't the article focused on offensive players? He mentions Beckett, but not any other pitchers.

[kab21]

I love it when people completely misuse a stat to confirm their bias against it.

[iastfan112]

A .900 OPS from a infielder with good defense, why yes I think that would be valuable!

[Brock Beauchamp]

I love it when people completely misuse a stat to confirm their bias against it.

 

This. WAR isn't meant to be used in small sizes like that... So to "disprove" the metric, somebody uses it that way and screams "THIS SUCKS".

 

It's akin to buying a blender and then taking it back to the store, screaming that it didn't toast your bread.

[snepp]

It's akin to buying a blender and then taking it back to the store, screaming that it didn't toast your bread.

 

Hey, some of us are a little more demanding of our electronics than others.

[IdahoPilgrim]

I love it when people completely misuse a stat to confirm their bias against it.

 

I agree that is a misuse of the statistic.

 

Another misuse is to read more into a statistic than is actually there. Bottom line - WAR, like all other statistics, is a quantitative tool which, by itself, is incapable of truly showing a players value or contribution. It can be a part of the final analysis, but is never the final analysis itself.

 

Unfortunately, too many people use that particular metric as the final analysis, which is the point I think the original poster was trying to make.

[Shane Wahl]

"Why WAR is the worst stat ever" does NOT equate with "too many people use that particular metric as the final analysis" AT ALL, sbk!

 

 

Anyway, on BR, Kershaw's WAR is 2.3. Punto's 1.1.

 

Good god.

[notoriousgod71]

How about it's dumb that it's calculated in different ways since Punto is a 1.1 on BR and .9 on Fangraphs.

 

What other stat just spews out random numbers like that? You don't have one place listing his OPS as .868 and another at .831. So I guess it goes back to the use of defensive metrics to determine a player's value and how flawed that is.

[Shane Wahl]

Yes, I would recommend pretty much ignoring dWAR and looking only at oWAR.

[big dog]

WAR! Huh! Woo-oh-oh, what is it good for? Absolutely Nothing! Huh!

 

It was time for a little Edwin Starr in the thread.

[kab21]

How about it's dumb that it's calculated in different ways since Punto is a 1.1 on BR and .9 on Fangraphs.

 

What other stat just spews out random numbers like that? You don't have one place listing his OPS as .868 and another at .831. So I guess it goes back to the use of defensive metrics to determine a player's value and how flawed that is.

 

If you have a problem with WAR then perhaps you shouldn't use it. If you want to make a case against it then don't use it in a way that it would be highly flawed. It's also not that hard pick apart WAR even if it has been used correctly. It's not a perfect stat and I don't think anyone has declared it as such.

[cmathewson]

If you have a problem with WAR then perhaps you shouldn't use it. If you want to make a case against it then don't use it in a way that it would be highly flawed. It's also not that hard pick apart WAR even if it has been used correctly. It's not a perfect stat and I don't think anyone has declared it as such.

 

Some people use it as a proxy for value, as in, "the MVP should just be the guy with the highest WAR." But most posters here understand its limitations .

[Kobs]

Is it a stat?

[Mr. Brooks]

How about it's dumb that it's calculated in different ways since Punto is a 1.1 on BR and .9 on Fangraphs.

 

What other stat just spews out random numbers like that? You don't have one place listing his OPS as .868 and another at .831. So I guess it goes back to the use of defensive metrics to determine a player's value and how flawed that is.

 

WAR is not a stat, it's a metric. Slight difference.

Saying WAR should be completely tossed out because it is calculated slightly differently by different sites is akin to saying that real estate appraisal is completely useless because one firm appraises your home at 250k and another firm appraises it at 255k.

[IdahoPilgrim]

WAR is not a stat, it's a metric. Slight difference.

Saying WAR should be completely tossed out because it is calculated slightly differently by different sites is akin to saying that real estate appraisal is completely useless because one firm appraises your home at 250k and another firm appraises it at 255k.

 

The problem comes when the differences are significant, not small. Particularly in pitching, Fangraphs WAR and BR WAR can have a large difference. That's like having one person appraise your house at $300K and another appraise it at $150K - you haven't really learned anything.

[Mr. Brooks]

The problem comes when the differences are significant, not small. Particularly in pitching, Fangraphs WAR and BR WAR can have a large difference. That's like having one person appraise your house at $300K and another appraise it at $150K - you haven't really learned anything.

 

I'd think its safe to assume that different sites having the same player with WAR's that far out of whack over a substantial sample size is probably an extremely tiny percentage.

They are almost always very small differences by the end of the year.

[kab21]

The problem comes when the differences are significant, not small. Particularly in pitching, Fangraphs WAR and BR WAR can have a large difference. That's like having one person appraise your house at $300K and another appraise it at $150K - you haven't really learned anything.

 

Are you looking at monthly WAR, yearly WAR or multi-year WAR?

[snepp]

The problem comes when the differences are significant, not small. Particularly in pitching, Fangraphs WAR and BR WAR can have a large difference.

 

One is based off FIP, one is based off ERA, where is the problem?

[ALessKosherScott]

Someday, I'd like to reconfigure WAR as a measure for ranking WAR songs. My money is on Low Rider winning.

[stringer bell]

WAR is flawed I'm sure. Currently for the Twins, Mauer is their top rated O-WAR player (2.0) and second is ....................Trevor Plouffe (0.9). Oops, Hammer comes in at 1.0, so Plouffie is third.

[Mr. Brooks]

Someday, I'd like to reconfigure WAR as a measure for ranking WAR songs. My money is on Low Rider winning.

 

Low Rider should be a distant 3rd behind Cisco Kid and Spill the Wine.

[Mike Sixel]

Love Spill the Wine.....