Here is a Problem I have with WAR

[Skeletor]

Hopefully, field f/x will make a lot of this moot in a couple years.

[RVAbird]

Hopefully, field f/x will make a lot of this moot in a couple years.

Yes, I'm really excited to see what hit f/x and field f/x will teach us. It's probably not healthy to be fixated on all this trivium, but I find everything about baseball endlessly fascinating.

[SrMeowMeow]

There's a good chance Field f/x will be proprietary.

[SrMeowMeow]

Of course you can't use WPA as a component in WAR. But for this particular example, it illustrates the point I was making.

And I get why Fangraphs uses FIP in WAR. I'm saying that they shouldn't. FIP is designed to measure how well a pitcher performed, not how much value he added. WAR is designed to measure how much value he added, not how well he performed.

If you want to subtract defense from a pitcher's contributions, then do just that, as B-R does. Don't remove a big chunk of contribution because of defense if you're seeking to measure contributions.

If you read the article on why they chose FIP over B-R's solution, I think you'll find it compelling. For the last time, they chose FIP as a least of many evils, not because they thought it was a flawless solution. They did a lot of hard work on WAR. I think they deserve to at least have their reasoning understood as a prerequisite for criticism.

[Skeletor]

There's a good chance Field f/x will be proprietary.

What I've read is just that it won't be made available to individuals. That's fine. We don't need the hard data anyway.

[SteveA]

Let's put this in context, ok? The argument againt fWAR is similar to the argument against Democracy - which is the worst form of government except for all the others.

Really, all joking aside, fWAR makes some assumptions that have a lot of grounding in reality and rates a player. And does that, arguably (and I believe definitely) better than the traditional numbers used to rate pitchers. If you're going to eviscerate a stat for not being perfect, or taking a particular point of view and quantifying it, please be fair and rip into the other metrics, too.

I mean, we don't have long, involved threads where the entire community tears apart ERA. ERA is fantastically absurd on so many levels. It tells me that the very same pitcher isn't as valuable if his starting shortstop took the day off and a lesser fielder took his place. It tells me that a pitcher is totally absolved of all runs that score after a questionable scoring decision awards an error to one of his fielders with two outs. It tells me that a pitcher who pitches in PETCO with a 4.00 ERA is better than a pitcher who pitches to a 4.05 in Coors.

C'mon. WAR is a solid metric whose results are well matched with qualitative observations a big majority of the time, with an underlying set of assumptions that may or may not be what you're looking for. Most of the stuff commonly used tocay are halfway decent metrics with underlying assumptions that range from ridiculous to decent.

This thread reminds me a bit of the Markakis or Roberts criticisms. They have flaws and sometimes produce results that aren't what we expect, so let's talk mostly about how disappointing and overrated they are.

But Skanar had a point: fWAR is a stat that is better for prediction...because as many have pointed out here, a pitcher's K rate, HR rate, an d BB rate stay more consistent year after year. So if you want to decide which pitcher to trade for, you are better off going with fWAR. But when comparing results, you want a stat that describes what they did, not seomthing that normalizes for what you think they should have done. A descriptive stat, not a predictive one. I think that is a very reasonable objection.

But if you want to look at what someone actually did, it seems to me that using hte fielding independent stat is not the way to go. At this point, you are measuring what someone has actually done. I mean ... why not do something similar with hitters, where they are judged on the groundballs, fly balls, and line drives they hit, rather than whether those balls fell in for hits or not?

[SteveA]

Actually, it DOES mean that. And that is one of the times where "eyes" come in: observation lets you know that the batter, in that case, was extremely lucky and unlikely to continue at that pace. And if you can't watch the games, you'll be able to see the same thing by looking at certain stats.

That's why the ideal thing is to use both. Eyes are not perfect, and stats are like lenses. Some are corrective, to help you see better; others may be polarized to strain out light rays bad for you; others may be colored, rose or orange or very dark tints, and prevent you from seeing certain things. They all have their uses.

Agree 100%. But very often on this board, and I think moreso with WAR than any other stat, someone will make a post which at least gives the impression that they believe WAR is an end-all be-all stat to describe a player's value (even if the people doing this sometimes use different WARs!). There will be a Hall of Fame discussion on the MLB board and someone will come in and post the WARs of the players involved, and perhaps other HOFers, and basically the thrust of their post is not that they have just contributed something more to the discussion, but that they have provided the facts that settle it, for good. End of discussion. I do realize that sometimes reading tone into an internet posting is fraught with peril, but I certainly get that sense from a lot of posts.

And similarly, on here, someone will post some WAR stats in reference to a past or future trade and basically say or imply that the discussion has been settled.

WAR gets tossed around here too much as an argument-settling, no-further-discussion-is-needed, because the Oracle has Spoken, type of stat. In my opinion. Not just as a stat that falls under your "they all have their uses" statement.

[SteveA]

I mean, the question of whether Fangraphs or BB-ref made the better choice of defense-independent pitching statistics is legitimate, but it's a lesser of two evils. In my opinion, BB-ref's has way more serious problems. But what there is no question about is this: you must use a pitching metric that factors out the contributions of the defense. If you want to argue that BB-ref's is better, you're more than welcome to, but you have an uphill battle ahead of you. And that argument is a sideshow to your original point, which was that WAR is flawed because it is predictive rather than descriptive. That is incorrect, and the decision to use FIP for the defense-independent pitching metric was a conscious design choice to keep WAR as descriptive as possible.

Is such a metric really possible? What of the effect that a catcher can have by calling a good game? What about windblown home runs? Is it "fair" if a guy pitched more games with the wind blowing out than otherwise? What about things that aren't measured by any of the currently used defensive stats, like a first baseman's ability to scoop bad throws and prevent the batter from getting to 2nd or even getting an out?

And since UZR is the basis for most of the defensive stats that make up WAR, and it is pretty well known that UZR needs a multiseason sample size to drown out the noise and provide a realistic picture, shouldn't every WAR stat for a position player that is based on just one season be taken with a huge grain of salt?

I think that in the zeal to create the perfect stat that eliminates all varying factors so that comparisons of players at different positions, or pitchers and hitters can be made, the stat guys have come up with a decent stat, but it is still flawed in many, many ways. But because it's goal is such a huge and noble goal, a one-stat-fits-all that eliminates all random varying factors, people have started thinking it is the penicillin of baseball stats and use it as a cure-all argument settler, rather than a rough approximation, that can occasionally inform us with soemthing about a player's value that we didn't know. But it can also occasionally mislead. I don't think it's penicillin yet, just a promising looking bread mold. But too often it is used as the single killer stat that it aspires to be, rather than the work in progress that it is right now.

[SrMeowMeow]

Instead of choice a) have a stroke, I'm going with choice b) bow out of the thread.

[Mizerooskie]

Disagree completely. Here are two stat lines:

Player A: 10 PA' date=' 1 BB, 3 hits

Player B: 10 PA, 3 BB, 2 hits

How good was each player? Different statistics will, in your words, "place more value on certain measurements than on others". Batting average says that the first guy was better (.333 for Player A, .286 for Player B). On-Base Percentage says the second guy is better (.500 for Player B, .400 for Player A). These two stats are just like every other "invented" stat. They declare a certain value for each outcome and arrive at different conclusions based on those assumptions.

You can look at as broad a range of stats as you like, but at some point it starts to look like the computer screens in The Matrix if you can't use tools to distill it into one or a few numbers.[/quote']This analysis doesn't make any sense at all.

The BA and OBP of the two examples you cited don't tell you who the better player is at all. They tell you who had a higher percentage of hits, and who had a higher percentage of getting on base. Neither of those statistics attempt to draw any conclusion. They're simply statistics. They are available for any sort of interpretation a person wishes to use. That's a stark difference from WAR et al. They essentially try to quantify a player's value with one number, which is always going to be prone to error, due to the game of baseball being such an uncontrolled environment.

As for a broad range of stats looking like a computer screen: oh well. Again, it gives you the best picture. Using one or two numbers is just lazy statistical analysis.