Here is a Problem I have with WAR

[olehippi]

WAR is based largely upon the subjective analysis of otherwise objective statistics. In other words, WAR is largely personal opinion and interpretation, not proven fact. IMO, that leaves the practical application of WAR wide open to attack.

[Fan4Life]

WAR is based largely upon the subjective analysis of otherwise objective statistics. In other words, WAR is largely personal opinion and interpretation, not proven fact. IMO, that leaves the practical application of WAR wide open to attack.

Why you lousy *#@#*&^*%$$*

I generally agree. However when it comes to measuring defensive abilities, it's an objective analysis of a subjective measurement.

[O'sFanThruJune]

This is coming from someone who thinks statistical analysis is a very good tool. I just think that these "invented" stats are too inconsistent to be useful. Give me basic stats (LD %, GB/FB, K/9, BB/9, etc.) that can't be influenced by how a particular service values various factors every time.

Down with WAR.

If we had the ability to quickly look at a player's basic stat line (PA, H, 2B, 3B, HR, BB, etc) and discern his value objectively, we wouldn't need any "invented" stats like batting average, slugging percentage, or wOBA. But somebody decided a long time ago, and pretty much everybody has agreed, that putting these simple numbers into a more complex calculation (SLG = [H + 2B + 2*3B + 3*HR]/[PA - BB - HBP - SF - SH]) makes it a lot easier to evaluate and compare players. Once we acknowledge that this type of tool is useful, it seems sensible to me to pay more attention to a formula using empirically derived values for each outcome than one that just arbitrarily counts a home run as four times better than a single.

[Mizerooskie]

If we had the ability to quickly look at a player's basic stat line (PA' date=' H, 2B, 3B, HR, BB, etc) and discern his value objectively, we wouldn't need any "invented" stats like batting average, slugging percentage, or wOBA. But somebody decided a long time ago, and pretty much everybody has agreed, that putting these simple numbers into a more complex calculation (SLG = [H + 2B + 2*3B + 3*HR']/[PA - BB - HBP - SF - SH]) makes it a lot easier to evaluate and compare players. Once we acknowledge that this type of tool is useful, it seems sensible to me to pay more attention to a formula using empirically derived values for each outcome than one that just arbitrarily counts a home run as four times better than a single.

Batting average and slugging percentage aren't "invented" statistics in the same sense. They don't place more value on certain measurements than on others. As evidenced by the the difference in WAR between FanGraphs and Baseball Reference, it's a stat completely fraught with opinion. Scratch that, it's a metric completely fraught with opinion. It's not a statistic at all.

I largely agree with what olehippi said.

And trying to compare two different players with a single statistic is foolish, IMO. You have to look at a broad range of stats to get a useful picture.

[Mizerooskie]

Yeah but...

All those micro-stats you cite are fine for what they measure, but each is so limited. So while they may be unquestionably accurate, we are left with the familiar question: What are they good for?

Some of us keep looking for the bigger picture - how to get a more holistic view of a pitcher's effectiveness, a player's value to the team, a contribution of the whole skill set to wins or losses. So we invent ways to combine the "facts on the ground" to create a virtual picture of the whole player and his whole value to the team; and ultimately (for that GM in all of us) a reliable way to compare one player to another - even a way to compare a starter to a reliever, or a pitcher to a position player. Granted, the result is also limited: virtual, not real.

Still, the quest continues...

My problem with WAR, RC, etc. is that they're not statistics at all. They're calculations of what one statistician believes equates to a players value.

The best way to get a holistic view of a player's effectiveness is to look at all the statistics available to you, not to try to use one or two catch-all metrics.

[O'sFanThruJune]

Batting average and slugging percentage aren't "invented" statistics in the same sense. They don't place more value on certain measurements than on others. As evidenced by the the difference in WAR between FanGraphs and Baseball Reference, it's a stat completely fraught with opinion. Scratch that, it's a metric completely fraught with opinion. It's not a statistic at all.

I largely agree with what olehippi said.

And trying to compare two different players with a single statistic is foolish, IMO. You have to look at a broad range of stats to get a useful picture.

Disagree completely. Here are two stat lines:

Player A: 10 PA, 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.

[Say O!]

Field f/x will hopefully assist fangraphs in calculating WAR for pitchers who tend to allow more contact but with successful results like Britton and Guthrie.

[JERSEYORIOLE]

Do you also discount the opinions of analysts and announcers?

Im not discounting WAR. I know nothing about it. If you read my post as confrontational, it wasn't. It was an informational question. I dont look at WAR, so im wondering if it is generally similar to the eye test.

[Can_of_corn]

Im not discounting WAR. I know nothing about it. If you read my post as confrontational, it wasn't. It was an informational question. I dont look at WAR, so im wondering if it is generally similar to the eye test.

I wasn't being confrontation either, I was curious since your approach is evidently much different from my own. I try and absorb as much quality information as I can to supplement what I see. You appear to take the opposite approach, I was just curious if you accepted input from other people's "eyes" or if you tuned them out and just went with what you see.

[JERSEYORIOLE]

I wasn't being confrontation either, I was curious since your approach is evidently much different from my own. I try and absorb as much quality information as I can to supplement what I see. You appear to take the opposite approach, I was just curious if you accepted input from other people's "eyes" or if you tuned them out and just went with what you see.

I was told to never answer a question with a question, however, I will answer yours none the less. I am not a scout or a coach. So I will gladly accept input from other peoples eyes that are qualified to see things I just can't. That leads back to my question. Does WAR and other metrics generally confirm what your eyes see? See here is my issue.....If people want to utilize these metrics as the bible, fine, go for it. I think they may have value, especially when you have not seen a player alot. Now these metrics have been around for I don't know how long, 30 years? Utilized by a vast number of teams for 20? Im guessing on these numbers and correct me if im wrong. Having said that. I have been a baseball fan for 42 years and I see just as many lopsided, bad trades and poor free agent signings as ever before. I'm sure these numbers have some value. I just find it hard to believe, that any qualified baseball "man" would get it wrong anymore then the metrics get it wrong.

[Can_of_corn]

I was told to never answer a question with a question, however, I will answer yours none the less. I am not a scout or a coach. So I will gladly accept input from other peoples eyes that are qualified to see things I just can't. That leads back to my question. Does WAR and other metrics generally confirm what your eyes see? See here is my issue.....If people want to utilize these metrics as the bible, fine, go for it. I think they may have value, especially when you have not seen a player alot. Now these metrics have been around for I don't know how long, 30 years? Utilized by a vast number of teams for 20? Im guessing on these numbers and correct me if im wrong. Having said that. I have been a baseball fan for 42 years and I see just as many lopsided, bad trades and poor free agent signings as ever before. I'm sure these numbers have some value. I just find it hard to believe, that any qualified baseball "man" would get it wrong anymore then the metrics get it wrong.

WAR does not necessarily match what you see. However what you see is not always what is happening.

Growing up I firmly believed that good pitchers could force weak contact from hitters. However Voros McCraken pretty conclusively proved that with a very few exceptions that is not the case (Guts may be one of those cases). That is a clear example of my eyes not being trustworthy. Honestly if all WAR did was reinforce what I already know I wouldn't bother with it.

I also know it is far from a perfect product and I draw from multiple sources to form my views.

[O'sFanThruJune]

WAR is based largely upon the subjective analysis of otherwise objective statistics. In other words, WAR is largely personal opinion and interpretation, not proven fact. IMO, that leaves the practical application of WAR wide open to attack.

There are so many things about this statement that are strange to me.

1. Even the basic statistics that I think you are referencing are not fully objective. The official scorer's opinion means the difference between a single and an error, or an earned run and an unearned run. Arbitrary rules declare that a batter can get an RBI on a groundout, but not a GIDP.

2. There is a different subjectivity inherent in every statistic beyond the basic counting stats (PA, H, BB, HR, etc) because the formulas for BA, OBP, ERA, etc, all assign value to different events. These values came out of someone's head subjectively. Someone decided how to define an unearned run and that they don't count against pitchers, and someone decided that walks and sacrifice flies are irrelevant to BA.

3. The absolute last thing I would say about WAR is that it is based on personal opinion. You can disagree with the methodology, but the truth is that batting runs above average, UZR, replacement level, and position factors are all mathematically derived. You can say they did it wrong, but I don't know why anyone would call it personal opinion.

4. The practical application of WAR is wide open to attack, just like the practical application of pitcher wins is open to attack. I don't think any knowledgeable SABR person ever claimed that WAR was a perfect statistic that could never be questioned. I happen to think it's quite useful, but I'm not going to report you to the SABR police if you tell me that the park adjustment factors don't look right to you.

[e16bball]

The issue with the different WARs for FanGraphs and B-R comes down to FG choosing to represent a pitcher's value as "how he SHOULD have done" and B-R choosing to represent that value as "how he DID do." It's not a surprise that most are more comfortable with B-R's approach, because it accords with what they actually saw, in terms of hits falling and runs crossing the place. I think there are arguments to be made on each side. On one hand, it seems odd to use "should have" in a metric that is intended to demonstrate how a pitcher actually performed. But on the other hand, it isn't necessarily fair to penalize a pitcher for things out of his control (bad luck, bad defense, etc.).

The one thing that CAN'T be argued, though, is the legitimacy of the underlying BABIP assumption. I've seen a lot of comments indicating that it's absurd to assume that pitchers have no control over whether a ball put in play isn't a hit. Which is an inaccurate characterization of the theory, as everyone accepts that more LDs (for example) means a higher BA. The BABIP assumption, though, is that even though pitchers have some control over the batted ball, over time the vast majority of pitchers are going to have BABIPs in the .290-.300, regardless of how good they are. Good pitcher, bad pitcher, it doesn't really matter. With only a handful of exceptions (usually heavy GB pitchers --- like Hudson or potentially Britton --- or occasionally a FB pitcher with unusual late movement --- like Santana or Guthrie), it's going to regress to that .290-.300 range with enough innings.

To demonstrate, here are two groups of pitchers:

Group A:

Pitcher A: 2381 career innings, .292 career BABIP against

Pitcher B: 1592 career innings, .293 career BABIP against

Pitcher C: 1483 career innings, .297 career BABIP against

Pitcher D: 3055 career innings, .309 career BABIP against

Group B:

Pitcher A: 892 career innings, .293 career BABIP against

Pitcher B: 1198 career innings, .295 career BABIP against

Pitcher C: 1178 career innings, .298 career BABIP against

Pitcher D: 1760 career innings, .309 career BABIP against

Group A? Roy Halladay, Josh Beckett, Cliff Lee, and Andy Pettitte

Group B? Daniel Cabrera, Omar Daal, Adam Eaton, and Sidney Ponson

Four of the best pitchers of the past decade and four of the worst. No discernible difference in results when the ball is put into play. You can argue the wisdom of using "should have" considerations in WAR...but it's almost impossible to argue the assumption that BABIP will regress to the set range in all but a handful of cases. No matter how great the pitcher, it's going to eventually end up coming back to .290-.300. So anything significantly far from that range is luck --- that just isn't going to continue. Guthrie's pitched long enough that it's now reasonable to believe he's truly an exception to the rule. Britton? Not so much. There's some regression yet to come for him.

[TJ Wrangler]

Britton has been worth 1.0 fWAR this year.

Tillman has been worth 0.9 fWAR this year.

Guthrie has been worth 0.6 fWAR this year.

One of these things is not like the other.

I find it very ironic that you don't like WAR, WARoriole.

[skanar]

The issue with the different WARs for FanGraphs and B-R comes down to FG choosing to represent a pitcher's value as "how he SHOULD have done" and B-R choosing to represent that value as "how he DID do."

Exactly the way I see it, and an excellent statement of why I prefer B-R's version. We don't give hitters credit for a hit when a fielder makes a great play to rob them of one; in other words, we don't correct for luck. And B-R has their own defensive correction mathematics.

The one thing that CAN'T be argued, though, is the legitimacy of the underlying BABIP assumption.

Maybe not, but your argument for why is poor. After all, I could make the claim that "good pitcher, bad pitcher, all will eventually have a K/9 of about 5.5-7 and support it with the following pitchers:

Group 1

Pitcher A: 3507 innings, 5.9 K/9

Pitcher B: 3055 innings, 6.6 K/9

Pitcher C: 2381 innings, 6.8 K/9

Group 2

Pitcher A: 2501 innings, 5.7 K/9

Pitcher B: 1178 innings, 6.5 K/9

Pitcher C: 892 innings, 6.8 K/9

Group 1? Juan Marichal, Andy Pettitte, Roy Halladay

Group 2? Steve Trachsel, Adam Eaton, Daniel Cabrera

Not to mention the oft-quoted examples of Jim Palmer and Walter Johnson, who had career BABIPs of .249 and .264, respectively. Palmer played in front of perhaps the best defensive team of all time, as most of us know, but in addition to outperforming the league BABIP, he also outperformed other Orioles pitchers of the time, who had a cumulative BABIP of .261. That means that over the course of his career, he allowed 150 fewer hits than he ought to: it could just be luck.

I'm willing to believe that many pitchers can't control BABIP and that some of them are good and some of them are bad. But I'm not convinced that NO pitcher can control BABIP.