WAR is Nonsense

[AZMike]

Imagine where the Angels would have been if we had Darvish instead of Weaver CJ WIlson last year.  

 

That said, Weaver's only a 3 win player. Equivalent to Kendrick, and slightly less than half as valuable as Aybar. Clearly he sucks.

[Hubs]

I think it comes down to calling them WINS above Replacement. If they had just stuck with RUNS against replacement, we'd not have this argument as often. Wins do not = 10 runs created. 

[Wallerrrr]

WAR is bullsh*t. how can one players stats totally determine the won/loss of a team?

 

thats a tad harsh..lol

im not a big sabermetrics guy...i dont think WAR is the end all / be all but I do think it is a very interesting tool that helps teams/fans determine the value of a player.

it has it's place in the game just like old school scouting and "gut feelings" do. I think there is a happy medium somewhere in the middle and I think Joe Maddon is a coach that balances it out very nicely.

[Angelsjunky]

A "replacement player" is an abtraction for a player that is easy for any team to come by - a minor league journeyman. Think Robb Quinlan, Bobby Wilson, Paul McAnulty, etc...basically AAA filler.

 

The reason a replacement player is used rather than, say, an average player (a replacement player is below an average regular) is that average regulars aren't always easy to come by, while a replacement player is.

 

So WAR basically says "This is how good this player is relative to someone we could easily replace him with if we had to."

 

That said, I don't think you can say that if a player has a 5 WAR then he adds five more wins to the team than a replacement player would have. One thing it doesn't include is "collateral damage" - how much that player effects the play of other players. A player like Mike Trout makes the whole team better by being a sparkplug. Or a player like Miguel Cabrera isn't worth "only" 7 wins more than Paul McAnulty simply because of his presence at the plate, and the fear he instills in pitchers.

 

I agree with those that say that WAR, or any stat, cannot possibly be truly definitive. But I disagree with the implication that it is therefore meaningless or without utility. WAR is a very useful tool. It just isn't the end-all, be-all.

[YouthofToday]

To be fair their are a myriad different ways to calculation a WAResque stat.

Oakland's internal WAResque stat actually had Miggy more valuable than Trout -- FanGraphs and baseball-reference's clearly had Trout>>>Miggy.

No WAR calculation without incorporating CERA or Erstad grit characteristics is disrespecting Mathis and Michael "professional hitter" Young. Why Anger Mathis and The Professional Hitter?

[Lou]

it's not meant to be a true quantifiable number or an absolute. It is a relative number using common replacement factors rather than specific real replacement values for a team.

the problem is that a lot of people use it as a true quantifiable number. All you have to do is read the posts throughout the season and it'll be the first argument made when evaluating a player.

[GregAlso]

*sigh* This is becoming tiresome.

Hubs, you asked what a "replacement player" looks like. I posted an article in this thread that gives you 24 examples of what they look like. Please read it and your question will be answered. Also positional adjustment is important because Cano providing his offense at 2nd base is much greater than someone providing that at say first base. It is much easier, relatively, to replace Cano's offensive production at first than at second. This is a distinct advantage to him and his team. This is why players value is adjusted for position on offense. On defense it makes even more sense. A player providing exceptional defense at a corner OF spot is not adding as much to his team as someone providing exceptional defense at CF. this means he adds more value to his team at CF than RF, hence the adjustment. Finally, 10 runs equal 'about' 1 win based pythag theorem of determining expected wins from runs for/runs against over a season. The theory holds fairly accurately and the math can easily be repeated, even though its boring. In the end 10 runs do equal about 1 win. Sorry you disagree with the math but if you want to see the math then google it or search on FanGraphs. It'll be very clear.

YoT, you speak of Ersty's grit. WAR values Ersty greater than other players because it quatifies all that he did that we may call grit. That's why he had 3.4 WAR in 2002 & an 8.8 WAR in 2000. WAR sees his periphery abilities as very valuable. That is another reason why some people really rely on WAR, it shows a more rounded understanding of a player's value. The reasons differing groups have differing WAR values is generally how they value defense.

To everyone who keeps talking about pitcher's WAR; it is generally done on what they think they know a pitcher can control (strikeouts, walks, & HRs.) The question remains open, can pitchers do things to control their BABIP or other stats? If so then how? There was a good article about controlling BABIP earlier today. If this is posdible then pitching WAR will need to be adjusted based on these facts.

Even though I defend WAR here it doesn't mean I'm in love with this stat. It is just useful and ridiculous to discount it without understanding it. If one player's WAR is similar to another's then it warrants looking into why and learning more about the value they provide to their team.

[Docwaukee]

If WAR is nonsense, then so is every other stat.  Clearly, that's not the case. 

 

What is the point of any stat? 

 

To me, it's an attempt to establish the value of a player so that we can understand the past and help predict the future.

 

The important considerations with any stat is to know it's strengths and weaknesses.  What it can tell you and what it doesn't.  What it considers, and what it doesn't. 

 

There will likely never be a stat that takes everything into consideration because of the vagaries of intangibles so any that claim a stat as the be all end all doesn't really understand what statistics are trying to accomplish.  

 

The so called 'mythical' replacement player, as some have said, is a point of reference.  But, it's a calculated point of reference that is no more vague than saying .300 is a 'good' batting average.  Yes there are players that hit .300 every year but we don't name that player as the reference point.  (although interestingly, we have associated a poor batting average with the mendoza line).  So in essence, all stats have a mythical point of reference that determines good, bad, or average.  WAR just gives their mythical point of reference a name - a replacement player. 

 

Saying WAR is nonsense, stupid or worthless is a 'the earth is flat' phenomena.  Metrics will continue to evolve and get better, more accurate, more important, more comprehensive etc.  The goal is to make them as meaningful as possible but they will never mean everything.  

[Lou]

GregAlso: " WAR sees his periphery abilities as very valuable."

 

What are these periphery abilities that were included? How is this measured as to incorporate it in his WAR calculation?

[Docwaukee]

GregAlso: " WAR sees his periphery abilities as very valuable."

 

What are these periphery abilities that were included? How is this measured as to incorporate it in his WAR calculation?

 

 

I think he means defense and baserunning to some degree. 

[Hubs]

*sigh* This is becoming tiresome.

Hubs, you asked what a "replacement player" looks like. I posted an article in this thread that gives you 24 examples of what they look like. Please read it and your question will be answered. Also positional adjustment is important because Cano providing his offense at 2nd base is much greater than someone providing that at say first base. It is much easier, relatively, to replace Cano's offensive production at first than at second. This is a distinct advantage to him and his team. This is why players value is adjusted for position on offense. On defense it makes even more sense. A player providing exceptional defense at a corner OF spot is not adding as much to his team as someone providing exceptional defense at CF. this means he adds more value to his team at CF than RF, hence the adjustment. Finally, 10 runs equal 'about' 1 win based pythag theorem of determining expected wins from runs for/runs against over a season. The theory holds fairly accurately and the math can easily be repeated, even though its boring. In the end 10 runs do equal about 1 win. Sorry you disagree with the math but if you want to see the math then google it or search on FanGraphs. It'll be very clear.

YoT, you speak of Ersty's grit. WAR values Ersty greater than other players because it quatifies all that he did that we may call grit. That's why he had 3.4 WAR in 2002 & an 8.8 WAR in 2000. WAR sees his periphery abilities as very valuable. That is another reason why some people really rely on WAR, it shows a more rounded understanding of a player's value. The reasons differing groups have differing WAR values is generally how they value defense.

To everyone who keeps talking about pitcher's WAR; it is generally done on what they think they know a pitcher can control (strikeouts, walks, & HRs.) The question remains open, can pitchers do things to control their BABIP or other stats? If so then how? There was a good article about controlling BABIP earlier today. If this is posdible then pitching WAR will need to be adjusted based on these facts.

Even though I defend WAR here it doesn't mean I'm in love with this stat. It is just useful and ridiculous to discount it without understanding it. If one player's WAR is similar to another's then it warrants looking into why and learning more about the value they provide to their team.

 

 

I get the math. I know what a 0.1 or 0.2 or -0.1 or -0.2 player looks like, the problem with that article is it is still comparing actual performances of journeyman to a fictional player. That's what the OP's article discussed... Also these players all had positional adjustments, which in some cases took the guy from a 1.0 WAR to a negative WAR. 

 

The positional argument in this last post is flawed…And here's why...

 

Why is Cano's offense better at 2nd than at 1st? Speaking strictly from a run created point of view, can't the same player produce the same number of runs offensively whether he plays 1st, 2nd, 3rd, or Rover? Seriously what does the defensive alignment have to do with a players offense? NOTHING. It only matters when you look at the replacement player. Because yes, it's easier to replace a guy at 1st or DH or in the corner OF. Yet their offensive value does not change. 

 

What if you take a oft injured but great fielding guy out of the outfield and put him in the infield at 1st base. Does his offensive contribution change? No. Does his defensive contribution change? Absolutely. This is why positional adjustment should be factored in to defensive WAR only. Also, the positional adjustments do not change in either Fangraphs or Baseball-Reference's WAR, from season to season. How is this possible? 

 

In 2012, the top offensive WAR players: Trout (CF), McCutchen (CF), Cabrera (3B), Posey (Catcher), Cano (2B), Headley (3B), Braun (LF), Beltre (3B), Zobrist (UT), Jones (CF). The only player on the list who gets a negative positional adjustment is Braun. Everyone else would have received at least +2.5 runs in their Offensive WAR for every 600 PA in that position.

 

Being a 1st Baseman or a Corner OF or worse a DH, dramatically alters a players offensive contribution, because in theory they are easier to replace. 

 

Trumbo's offense is clearly there. He has had some rough stretches, but overall, the guy has been pretty good in his first two seasons. By WAR, putting him at 3rd or 2nd would increase his offensive value, even if he played horrifically on defense. How does that make sense? His WAR would go down, but only because of the fielding, not because of the offense value, which would increase…and it's significant as if he was a 3B versus a DH, it's a 20 run (2.0 WAR) switch. The guy could have a 1.000 OPS as a DH, but actually be less valuable to us as a DH by WAR than as our mediocre 3B. That's just silly. Defensively he'd likely cost us that many runs playing 3B versus a solid fielding 3rd base, but his offense wouldn't change. 

 

Another example is Mike Trout. Does his VALUE, what WAR is truly trying to determine, lessen by playing LF on the offensive side of the ball? NO. He may not make as much of an impact defensively in left, but then, doesn't the team get better by having excellent fielders in both spots? Assuming of course that his replacement makes as much of a difference in center defensively as he would have, and that his defense in left is an increase over whoever else would've played that position?

 

Mike Trout as a left fielder in 2012, would've finished with a 7.6 offensive WAR, instead of 8.6 as it's a 10 run switch. 

[Lou]

I think he means defense and baserunning to some degree. 

 

 

Not too sure about that.  He stated: "YoT, you speak of Ersty's grit. WAR values Ersty greater than other players because it quatifies all that he did that we may call grit."

[The God of Runs]

If the replacement player was Trout, that would mean that every player in the majors has a negative WAR

[Docwaukee]

I get the math. I know what a 0.1 or 0.2 or -0.1 or -0.2 player looks like, the problem with that article is it is still comparing actual performances of journeyman to a fictional player. That's what the OP's article discussed... Also these players all had positional adjustments, which in some cases took the guy from a 1.0 WAR to a negative WAR. 

 

The positional argument in this last post is flawed…And here's why...

 

Why is Cano's offense better at 2nd than at 1st? Speaking strictly from a run created point of view, can't the same player produce the same number of runs offensively whether he plays 1st, 2nd, 3rd, or Rover? Seriously what does the defensive alignment have to do with a players offense? NOTHING. It only matters when you look at the replacement player. Because yes, it's easier to replace a guy at 1st or DH or in the corner OF. Yet their offensive value does not change. 

You answered your own question.  It's taken in account based on the ability to replace him.  not just his production.  If want an estimate of his value to the team as an offensive player, then just look at the number of runs he adds.  or look at his WPA or his wOBA or something else you think is helpful. 

What if you take a oft injured but great fielding guy out of the outfield and put him in the infield at 1st base. Does his offensive contribution change? No. Does his defensive contribution change? Absolutely. This is why positional adjustment should be factored in to defensive WAR only. Also, the positional adjustments do not change in either Fangraphs or Baseball-Reference's WAR, from season to season. How is this possible? 

Because it's not a real player that's actually doing the replacing.  It's a calculation based on league stats not who actually would do the replacing.  Again, it's a point of reference. 

In 2012, the top offensive WAR players: Trout (CF), McCutchen (CF), Cabrera (3B), Posey (Catcher), Cano (2B), Headley (3B), Braun (LF), Beltre (3B), Zobrist (UT), Jones (CF). The only player on the list who gets a negative positional adjustment is Braun. Everyone else would have received at least +2.5 runs in their Offensive WAR for every 600 PA in that position.

 

Being a 1st Baseman or a Corner OF or worse a DH, dramatically alters a players offensive contribution, because in theory they are easier to replace. 

it doesn't lessen their offensive contribution, it lessens their value relative to a reference point of a theoretical replacement

Trumbo's offense is clearly there. He has had some rough stretches, but overall, the guy has been pretty good in his first two seasons. By WAR, putting him at 3rd or 2nd would increase his offensive value, even if he played horrifically on defense. How does that make sense? His WAR would go down, but only because of the fielding, not because of the offense value, which would increase…and it's significant as if he was a 3B versus a DH, it's a 20 run (2.0 WAR) switch. The guy could have a 1.000 OPS as a DH, but actually be less valuable to us as a DH by WAR than as our mediocre 3B. That's just silly. Defensively he'd likely cost us that many runs playing 3B versus a solid fielding 3rd base, but his offense wouldn't change. 

it's relative value hubs.  not absolute

 

Another example is Mike Trout. Does his VALUE, what WAR is truly trying to determine, lessen by playing LF on the offensive side of the ball? NO. He may not make as much of an impact defensively in left, but then, doesn't the team get better by having excellent fielders in both spots? Assuming of course that his replacement makes as much of a difference in center defensively as he would have, and that his defense in left is an increase over whoever else would've played that position?

 

Mike Trout as a left fielder in 2012, would've finished with a 7.6 offensive WAR, instead of 8.6 as it's a 10 run switch. 

again, in theory, it's easier to replace his value in LF than CF. 

[steveuk]

What is it good for?

 

Absolutely nothing. Say it again.

 

I cant believe I read 4 pages of BS to make sure I didn't Craig this....

[Docwaukee]

Not too sure about that.  He stated: "YoT, you speak of Ersty's grit. WAR values Ersty greater than other players because it quatifies all that he did that we may call grit."

I think he knows that it doesn't measure actual grit.  Ersty's 'grit' is unquantifiable.  Things that don't usually get attributed value like baserunning and defense were generally considered intangibles in the past because they didn't have metrics for them built into any give stat.  They were 'of the things that could account for grit'.

[Hubs]

I like WAR, but I don't like the positional adjustments.

 

Which team is better:

 

CF Trout (10.7 WAR)

SS Aybar (4.0 WAR) 

2B Cano (8.2 WAR)

3B Cabrera (6.7 WAR)

1B Ike Davis (0.7 WAR)

LF Jason Kubel (0.8 WAR)

RF Nelson Cruz  (0.1 WAR)

DH Adam Dunn (0.9 WAR) 

C Buster Posey (7.2 WAR)

 

Total WAR = 39.3

 

Or 

 

CF Colby Rasmus (1.2 WAR) 

SS Johnny Peralta (0.9 WAR) 

2B Daniel Murphy (1.2 WAR)

3B Chris Johnson (0.7 WAR)

1B Albert Pujols (4.6 WAR)

LF Ryan Braun (6.8 WAR)

RF Giancarlo Stanton (5.2 WAR)

C Matt Weiters (3.2 WAR)

DH David Ortiz (2.9 WAR)

 

Total WAR = 26.7.

 

The first team has the three best hitters in the AL by WAR in Trout, Cano, Cabrera plus Buster Posey who was the best in the NL. Adam Dunn hit 41 HR but would be as valuable as Scott Cousins as your DH? 

 

Is the second team really going to score 130 runs less? With a middle of the lineup of Pujols/Braun/Stanton/Ortiz/Weiters?

[Hubs]

Doc...you make some good points. 

 

I don't like the theoretical replacement values. I don't have an issue with positional adjustment, just the values that they use. 

 

Yes it's harder to replace Beltre or Headley at 3rd than replacing Mark Trumbo at DH, but is it really that much harder? Can't we just take this into account in the fielding replacement? As I just posted above, is a team with outstanding 1st baseman, corner outfielders, and DH that much worse than a team with an outstanding 2B, 3B, C and CF?

 

For those who don't know here are the positional adjustments:

 

C: +12.5 runs (or 1.25 WAR)

2B: +2.5 runs

3B: +2.5 runs 

 

SS: +2.5 runs 

CF: +2.5 runs 

LF: -7.5 runs

RF: -7.5 runs

1B: -12.5 runs

DH: -17.5 runs

 

 

 

[Docwaukee]

I like WAR, but I don't like the positional adjustments.

 

Which team is better:

 

CF Trout (10.7 WAR)

SS Aybar (4.0 WAR) 

2B Cano (8.2 WAR)

3B Cabrera (6.7 WAR)

1B Ike Davis (0.7 WAR)

LF Jason Kubel (0.8 WAR)

RF Nelson Cruz  (0.1 WAR)

DH Adam Dunn (0.9 WAR) 

C Buster Posey (7.2 WAR)

 

Total WAR = 39.3

 

Or 

 

CF Colby Rasmus (1.2 WAR) 

SS Johnny Peralta (0.9 WAR) 

2B Daniel Murphy (1.2 WAR)

3B Chris Johnson (0.7 WAR)

1B Albert Pujols (4.6 WAR)

LF Ryan Braun (6.8 WAR)

RF Giancarlo Stanton (5.2 WAR)

C Matt Weiters (3.2 WAR)

DH David Ortiz (2.9 WAR)

 

Total WAR = 26.7.

 

The first team has the three best hitters in the AL by WAR in Trout, Cano, Cabrera plus Buster Posey who was the best in the NL. Adam Dunn hit 41 HR but would be as valuable as Scott Cousins as your DH? 

 

Is the second team really going to score 130 runs less? With a middle of the lineup of Pujols/Braun/Stanton/Ortiz/Weiters?

not really understanding your point.

 

If scott cousins was a DH for an entire year and he put up a .250 avg with 10hrs, his WAR would be like -3

 

answer to second question - Who the frick knows, but I doubt it, because you put cano, aybar, trout and posey on that first team who get a lot of credit for runs created with baserunning and runs saved with defense.  Again, the positional adjustments are RELATIVE to the ability to replace that player with a calculated point of reference.

 

ex.  If Jason Kubel played CF instead of LF with similar defense, his value (not just offensive production) is likely more difficult to replace in CF than in LF.  WAR attempt to quantify that.  It's not perfect but having positional adjustments is more accurate for assessing value for that player relative to not. 

[Docwaukee]

Doc...you make some good points. 

 

I don't like the theoretical replacement values. I don't have an issue with positional adjustment, just the values that they use. 

 

Yes it's harder to replace Beltre or Headley at 3rd than replacing Mark Trumbo at DH, but is it really that much harder? Can't we just take this into account in the fielding replacement? As I just posted above, is a team with outstanding 1st baseman, corner outfielders, and DH that much worse than a team with an outstanding 2B, 3B, C and CF?

 

For those who don't know here are the positional adjustments:

 

C: +12.5 runs (or 1.25 WAR)

2B: +2.5 runs

3B: +2.5 runs 

 

SS: +2.5 runs 

CF: +2.5 runs 

LF: -7.5 runs

RF: -7.5 runs

1B: -12.5 runs

DH: -17.5 runs

 

 

 

it's easier to give a set standard for a position than to try to accommodate such with adding or subtracting value based on defense alone.  mostly because defense metrics are the least well developed of the calc. 

 

Trying to pic apart the holes in WAR is actually pretty easy.  No one should argue with that.  It's a comprehensive counting stat that is what it is.  I can see with my own eyes that Matt Harrison is not 3 wins better than Jered Weaver.  It's up to the user to cross check what may seem like a better indicator of value. 

[AngelsLakersFan]

First, the article was terrible - something I rarely say. There is no real argument made against WAR within it.

 

Second - the problem with WAR is that it is being used by people who do not understand it. WAR is a tool designed to provide a deeper understanding of the sport, but the majority of people discussing WAR (pro or con) do not understand how the tool works. 

 

Third - people criticize 'replacement level' but this not an important part of the calculation because it has no effect on the results of the stat. It's akin to arguing whether we should start counting at 0 or 1, no matter what 3 is still greater than 2.

 

Fourth - WAR is a mathematical calculation and the individual components do not always make sense without considering the mathematical processes used in other parts of the algorithm. Someone complained about the use of positional adjustments in the stat, saying that offense is equally valuable no matter where it comes from. I don't entirely disagree with this but the stat is using positional adjustments in one area to offset the fact that offense and defensive productions are compared to LEAGUE AVERAGE when added to the calculation. Being 10 runs above average on defense at SS is not the same as being 10 runs above average at 1B.

[gotbeer]

 

 

Wouldn't this be a better comparison?  

 

The Angels starting team last year, with the Giants.  Just on oWAR, (yeah too lazy to look it up other than on espn) Angels 31.4/Giants with no DH 25.3.  Morales was a 2.1.

 

Giants starting pitching 5 WAR 5.0/Angels 2.8 (Angels combined Greinke with Williams)  

 

Giants 94 wins and a minor trophy.  Angels 89 Wins.  

 

 

[AngelsLakersFan]

Another example is Mike Trout. Does his VALUE, what WAR is truly trying to determine, lessen by playing LF on the offensive side of the ball? NO. He may not make as much of an impact defensively in left, but then, doesn't the team get better by having excellent fielders in both spots? Assuming of course that his replacement makes as much of a difference in center defensively as he would have, and that his defense in left is an increase over whoever else would've played that position?

 

Mike Trout as a left fielder in 2012, would've finished with a 7.6 offensive WAR, instead of 8.6 as it's a 10 run switch. 

 

If Trout is in LF his actual defensive performance would increase compared to the average. Theoretically his value will be decreased only by the fact that there are fewer chances in LF than CF. If he were to get the same number of chances he should theoretically perform at the same rate.

[AngelsLakersFan]

Also the Weaver / Harrison pitcher WAR thing.

 

Harrison might have a higher WAR than Weaver but the stat is NOT SAYING that switching the two players would result in a net increase in wins for the Angels. WAR is handing out value to the pitchers that they can directly account for. Jered Weaver has excess value that is likely being attributed to Mike Trout and other Angels defenders as well as to Angels stadium. Swapping the two pitchers would result in Angels pitchers having higher pitcher WAR but Angels defenders receiving less dWAR.

 

This is an accounting issue.

[GregAlso]

GregAlso: " WAR sees his periphery abilities as very valuable."

What are these periphery abilities that were included? How is this measured as to incorporate it in his WAR calculation?

Yes, I mean his gritty defense, his smart baserunning, going first to third, etc. Those things not generally measured in traditional stats that defensive metrics and linear weights can sometimes see that we have called grit.