Somebody please explain this to me. Offensive WAR. Josh Willingham's is higher than Ryan Braun's

[The stache]

2012 offensive WAR to date. I was browsing ESPN, and saw the baseball page has WAR leaders listed. I saw Willingham in the top 10. I didn't see Braun.

 

oWAR does not take into consideration defense at all, and Braun and Willingham both play left field.

 

Braun's oWAR is 3.4

Willingham's oWAR is 3.7

 

Huh?

 

Willingham is batting .271. Braun is batting .318

Willingham has 25 HR. Braun has 28.

Willingham has scored 57 runs, Braun has scored 65.

Willingham has 71 RBI. Braun has 69.

Willingham has 2 SB. Braun has 17.

Willingham has 52 BB and 88 K's. Braun has 40 BB and 76 K's.

Willingham's OBP is .384. Braun's is .400.

Willingham's OPS is .947. Braun's is 1.019.

 

Braun has 3 less RBI and 12 fewer walks. Braun is batting 47 points higher, has 3 more home runs, has 8 more runs scored, 15 more stolen bases, his OBP is 16 points higher, and his OPS is 72 points higher.

 

Take this a step further. Willingham is 14th in MLB in total bases. Braun is 4th. Willingham's isolated power is .292, 6th best in MLB. Braun? His isolated power is .301, third best in the majors. Runs created per 27 outs? Willingham's at 7.38, 21st in MLB. Braun? 6th in MLB at 8.78. Runs created? Willingham is 13th at 71. Braun? 2nd in MLB at 82.3, trailing only Andrew McCutchen at 87.7.

 

Could somebody PLEASE explain to me how Willingham could possibly have a higher offensive WAR than Braun? This is why I don't put a lot of faith in WAR.

[Baldkin]

They play in different leagues.

Braun plays in a better hitters parks.

Willingham faces better pitching and better defenses.

 

Those three things are all factors.

 

Edit: This probably all comes down to park factor, MP is one of the better hitters parks in the league, while Willingham plays in one that is in the middle of the pack.

[Raw Biz]

This probably all comes down to park factor, MP is one of the better hitters parks in the league, while Willingham plays in one that is in the middle of the pack.

 

That's immediately what comes to mind. And keep in mind, these stats aren't gospel. Sure there are flaws but they are still useful in general comparison terms.

[jjfanec]

They play in different leagues.

Braun plays in a better hitters parks.

Willingham faces better pitching and better defenses.

 

Those three things are all factors.

 

Edit: This probably all comes down to park factor, MP is one of the better hitters parks in the league, while Willingham plays in one that is in the middle of the pack.

 

Willingham faces better pitching? Team ERA ranking in all of baseball in the NL Central - Reds are 2, Pirates are 6, Cards are 10, and the Cubs and Astros are 24 and 28. The AL Central - Tigers are 14, White Sox are 17, Indians are 25 and Royals are 26. Now I know the AL has a DH but the AL Central is not a strong pitching division at all. As for fielding I know the Tigers are one of the worst fielding teams in baseball but I have no idea about the rest of the division.

[gregmag]

OPS+ also considers park factors, I believe. Willingham is at 158, and Braun is at 161. That still has Braun up, but it closes almost all of the differential in raw OPS.

 

Obviously stats like runs and RBI aren't going to figure into a WAR calculation, any more than they figure into OPS. If you believe runs and RBI provide highly significant information for evaluating a player's performance, then you have a good underlying reason not to put any stock in WAR. That's just a basic philosophical difference.

 

By most meaningful measures they're very close, as your original post shows. Their offensive WAR numbers are also very close, 3.4 and 3.7. I think most people who put substantial stock in WAR would tell you that a .3 WAR difference shouldn't end any discussion about which player is more valuable. I would certainly tell you that. With any stat, you need to know what assumptions it makes, what information it takes into account, and how it weighs different factors; but every stat will suffer from some degree of imprecision, even if you buy its premises.

 

For me, seeing Braun at 161 OPS+ and 3.4 WAR against Willingham at 158 OPS+ and 3.7 WAR suggests that both stats are telling roughly the same story. The variation just doesn't seem like a big deal.

[The stache]

Willingham plays in the American League with the designated hitter, whereas Braun plays in a league where the pitcher bats. Willingham gets more opportunities to drive in runs.

 

As far as the parks argument goes.

 

Willingham away from Target Field: 169 AB's, 10 HR, 32 RBI, .237 AVG

Braun away from Miller Park: 190 AB's, 12 HR, 27 RBI, .299 AVG

 

And as far as Braun playing in a more hitter friendly park, that's an argument that diminishes Braun's power (not saying you're making it, I know that is standard reasoning).

 

Based on ESPN's home run tracker, every one of Braun's 28 HR this year would have been out of Target Field:

 

http://img443.imageshack.us/img443/9926/overlay1343319629145714.jpg

 

If Braun were the type of hitter that could benefit from playing in a hitter's park, I could see this as part of the argument. But Braun doesn't get cheap home runs. This is a guy that hit two HR out of Target Field in one game. Willingham, playing half of his games there, has done that once. Braun hit three out of Petco. That's never been done before.

 

Braun is going to hit the ball out of any park in the majors, and I really don't think his power numbers are gaining any discernible benefit from playing there. There's more to it than pure dimensions, I know, but I just can't see Brauny going to Target Field, and seeing a massive drop in his power numbers.

[The stache]

Sorry if I'm a little punchy guys. I've been up all night, and vicodin for my back has me a bit doped up.

 

I just don't see the two of them as being close really. Braun is a much better hitter, and I cannot buy into the notion that Josh Willingham would somehow be worth more wins, or in this case, a third of a win more than Braun. Braun is in the discussion for NL MVP. Is Willingham in the discussion for AL MVP? No way. Yet this idiotic metric would have us believe that Willingham is somehow more worthwhile than Braun.

[Baldkin]

Go figure, the teams that play in the league where the pitchers hit have a way better ERA.

 

But, as I said, it is almost all park factor.

 

Offensively they are very similar, the difference is defensively Braun is average, Willingham is horrible.

 

You are overvaluing RBI and Runs, as those are a factor of the guys playing around the hitter as much as the hitter themselves and WAR does it's best to be context neutral.

[Ennder]

It is based on bad misleading useless park factors is where the biggest difference is coming. Target field kills left handed power hitters but doesn't really bother RH pull hitters so Willingham gets a significant adjustment to his OPS because his park hurts players that aren't like him. You can pretty much throw out the window any stat that says it is adjusted for park factors because they aren't any better than stats that aren't adjusted at all.

[thebruce44]

Why don't you guys just use Fangraphs for wOBA and WAR?

[sveumrules]

What thebruce44 said. Fangraphs has Braun 4th in MLB with 36.7 batting runs and Willingham 10th at 28.5. WAR is far from perfect as a whole but the batting component (as calculated by Fangraphs at least) is certainly not an issue.

[thebruce44]

What thebruce44 said. Fangraphs has Braun 4th in MLB with 36.7 batting runs and Willingham 10th at 28.5. WAR is far from perfect as a whole but the batting component (as calculated by Fangraphs at least) is certainly not an issue.

 

I feel like the biggest issue, or reason people don't like WAR beyond not understanding it, is that a bunch of other websites went and just started calculating it their own way without changing the metric's name. Their way seems to be more unreliable and lead to even more confusion. The obvious disclaimer is that WAR isn't perfect, but Fangraph's version is very good when used correctly.

[TooLiveBrew]

Yeah, completely agree on FG's WAR being the one to use. It just always seems to pass the 'common sense test', if that's a test that is a thing.

 

I go to BB-reference for almost everything, stats-wise. But not WAR.

[Funketown]

I don't know what's up with ESPN's WAR but that's pretty silly. fWAR has Braun at 5.4, Willingham 3.7, and Braun leads in oWAR by about 1 WAR.

[rluzinski]

It is based on bad misleading useless park factors is where the biggest difference is coming. Target field kills left handed power hitters but doesn't really bother RH pull hitters so Willingham gets a significant adjustment to his OPS because his park hurts players that aren't like him. You can pretty much throw out the window any stat that says it is adjusted for park factors because they aren't any better than stats that aren't adjusted at all.

Are you sure they dont calculate separate pf for lefties and righties?

 

There are a lot of problems with park factors but I'm not convinced that ignoring park effects completely is a better approach. Can you provide any evidence?

[nate82]

It is based on bad misleading useless park factors is where the biggest difference is coming. Target field kills left handed power hitters but doesn't really bother RH pull hitters so Willingham gets a significant adjustment to his OPS because his park hurts players that aren't like him. You can pretty much throw out the window any stat that says it is adjusted for park factors because they aren't any better than stats that aren't adjusted at all.

Are you sure they dont calculate separate pf for lefties and righties?

 

There are a lot of problems with park factors but I'm not convinced that ignoring park effects completely is a better approach.

 

Shouldn't the park factors cancel themselves out at some point?

[gregmag]

I didn't realize ESPN calculated their own numbers. Yeah, use Fangraphs WAR. I learned that the hard way, relying on Baseball Reference WAR for a pitcher comparison last year.

 

I'm not buying Ennder's blanket criticism of park factors without a lot more evidence, though I certainly agree that using more and better information to make those calculations is always preferable. But that's the same reason I believe in park factors to begin with: context matters, and unless the context in question completely eludes capture, you should take it into account.

 

That said, 'stache, I really like the way you made your argument above. You're doing exactly what I think all statheads at their best try to do: gathering relevant information, sifting through it, and trying to find meaningful facts. You're also right to add the qualification that dimensions aren't everything when looking at parks.

[rluzinski]

I just don't see the two of them as being close really. Braun is a much better hitter, and I cannot buy into the notion that Josh Willingham would somehow be worth more wins, or in this case, a third of a win more than Braun. Braun is in the discussion for NL MVP. Is Willingham in the discussion for AL MVP? No way. Yet this idiotic metric would have us believe that Willingham is somehow more worthwhile than Braun.

 

You are going to thing WAR is idiotic if you think it's trying to estimate who is the better hitter, who is more "worthwhile" or who's production has been worth more to their respective teams. It's not trying to answer any of those questions. It's just the context neutral value of a player's production over a certain period of time. Basically, it takes a linear weights approach, where all batting events are given a run value based on their average values. I suggest your read Fangraph's primer on WAR:

 

http://www.fangraphs.com/library/index.php/misc/war/

 

As I've written many times before, there is plenty of room to criticize certain aspects of WAR's construction but the tendency is for people to simply cherry pick two players and use it as the foundation for their criticism. The comparison is usually between an established superstar who is predictably doing well and some league average batter playing well over his head. Is the expectation that WAR should take that into account?

 

As for park factors, at best they adjust for the batter's handedness. Otherwise, they are for the "average player". If Willingham's HR's barely made it over the fence in a small park and Braun's HR's would have been a HR in any park, park factors are going to unfairly hurt Braun in that comparison.

 

And finally, the AL clearly has more talent. There are a number of ways to support that claim but the most easy way is to simply see how poorly the NL has performed against the AL since the inception of interleague play. As a fan of an NL team, it's kind of embarrassing.

[Funketown]

But in all seriousness what the heck is ESPN possibly screwing up with their WAR calculation that says Willingham is worth more oWAR than Braun?

 

Leave it up to ESPN to introduce a new statistic, calculate it incorrectly, and use it incorrectly.

 

Colin Cowherd was on the radio trying to use WAR to say that there is no way Justin Verlander is worth $20million per season. I'm not kidding, it went something like this:

 

- Verlander won 21 games last season. WAR tells us a replacement player probably would have won 13. But, Verlander would have probably been replaced by the number 2 guy in the rotation who would probably win 17 games, so he's really only worth 4 wins per season. Clearly, there's no way another 4 wins is worth $20 million dollars.

 

Needless to say, he does not fully grasp baseball or WAR.

[TURBO]

Who cares really? It only matters if you believe in WAR as an actual stat.

 

I myself think WAR is unreliable and inaccurate, therefore, I pretty much ignore it's findings.

[The stache]

Thanks guys, I feel much better. I'd heard that WAR was calculated differently depending on where you went, but this just struck me as odd. I've never really delved into it before. I'm a stock broker and deal with numbers a lot, and things like this perplex me, and set off my ADD. It's amazing to me that there can be such wide variance in calculations from one source to another, but it's still universally called WAR.

 

Thanks for the compliment, gregmag. I'm a longtime baseball nut, and started memorizing player stats when I was a little kid, long before metrics like babip, WAR, etc came into the baseball vernacular.

 

FYI, when I went to ESPN's home run tracker, and looked up Josh Willingham, overlaying Miller Park on his tracked home runs, one of Willingham's home runs would have been an out....to Ryan Braun, of all people:

 

http://img838.imageshack.us/img838/4243/overlay1343349452210136.jpg

[The stache]

But in all seriousness what the heck is ESPN possibly screwing up with their WAR calculation that says Willingham is worth more oWAR than Braun?

 

Remember, it's baseball-reference providing their data. I don't really like ESPN (though I strangely still go there. I guess ESPN is like eating peas as a kid. You never really learn to like them, but you still eat them because you've grown used to them), but in this one area, they are not really in the wrong.

[Funketown]

But in all seriousness what the heck is ESPN possibly screwing up with their WAR calculation that says Willingham is worth more oWAR than Braun?

 

Remember, it's baseball-reference providing their data. I don't really like ESPN (though I strangely still go there. I guess ESPN is like eating peas as a kid. You never really learn to like them, but you still eat them because you've grown used to them), but in this one area, they are not really in the wrong.

 

Well then, what the heck does bWAR do that makes their oWAR calculation so goofy? I imagine their weights are pretty similar.

[rluzinski]

Who cares really? It only matters if you believe in WAR as an actual stat.

 

I myself think WAR is unreliable and inaccurate, therefore, I pretty much ignore it's findings.

 

I suspect that you would say that about any baseball stat your are unfamiliar with. Like it or now, you are going to have to ignore the majority of stats in 20 years, then. That's bound to happen when a sport uses the same metrics for 100 years. Time to play some catch up!

[Thurston Fluff]

I suspect that you would say that about any baseball stat your are unfamiliar with. Like it or now, you are going to have to ignore the majority of stats in 20 years, then. That's bound to happen when a sport uses the same metrics for 100 years. Time to play some catch up!

 

Are you suggesting if you don't know about one stat you have to be ignorant of every stat? I also think it a bit off base to suggest the only people who don't think WAR is useful are ones who don't understand them. Blanket statements like that make it seem like only people who agree with you actually understand what it is. I am familiar with WAR. Frankly I've been so against WAR I've spent more time trying to understand them than any other metric. I know how they work and what is in them. I know the formula and how it's calculated. It is based on that knowledge why I don't accept it as useful. I think it is so far off of accurate in it's portrayal of a players performance that it is virtually worthless. In fact given how often it is misused it's less than worthless. It misleads people into thinking players are something they are not.

 

Most of the things I disagree with I have stated many times. Poorly weighted metrics put into poorly thought out equations. Yes, the numbers work when you put them in. Yes, it does create a baseline to compare players. But if that baseline is created by poor stats like single season UZR it isn't going to create accurate numbers on any given player. Thus the comparison baseline isn't useful. As the saying goes garbage in garbage out.

If they would get rid of defense and just rate them offensively I think it would help a lot. They wouldn't evaluate the total player but at least it would be accurate in what they do measure. I also would prefer they change the baseline from replacement level to a comparison against average. It would help compare what the player actually does contribute because it is possible to evaluate league average to a much greater degree of accuracy than some mythical AAAA player. I think comparisons to actual production is always better than hypothetical production. So change WAR to only measure accurately measured offensive metrics in proper quantities and base the number off league average and then lets talk. Until then I think there are far better stats to use to evaluate a player. Ones that use accurate numbers to calculate value are preferable for me.