WAR is Nonsense

[Hubs]

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.

 

Exactly. Which, because of WAR's positional adjustment would still quantify the LF Trout about one win less than the CF Trout. Even though they are exactly the same player, with the same defensive ability in the same number of chances, because in theory it is easier to replace a corner outfielder than a centerfielder.

 

The positional adjustment can be argued in the infield sure, but I don't get the 10 run switch between left or right and center. Center is tougher to play, but more offensive production has come from center than from the corners on the top end, which theoretically raises the replacement value, and should account for  a lower adjustment, or a higher adjustment for the corners. Despite the fact that 10 of the best 20 hitting OF in the Majors over the past two years played CF.

 

1. Trout (2012) 

2. Kemp (2011)

3. McCutchen (2012)

4. Braun (2011)

5. Stanton (2011)

6. Gordon (2012)

7. Hunter (2012)

8. Bautista (2011)

9. Jackson (2012)

10. Jones (2012)

11. Hamilton (2012)

12. Beltran (2011)

13. Ellsbury (2011)

14. Reddick (2012)

15. Span (2012) 

16. Heyward (2012)

17. Granderson (2011 or 2012)

18. Justin Upton (2011)

19. Pence (2011)

20. Holiday (2011)

21. Bourn (2012) / Pagan (2012) / Bourjos (2011) / Harper (2012) Take your pick here.

 

This doesn't even take into account a second season, just the best season for each player of the past two, and it's entirely a subjective order.

[AngelsLakersFan]

Exactly. Which, because of WAR's positional adjustment would still quantify the LF Trout about one win less than the CF Trout. Even though they are exactly the same player, with the same defensive ability in the same number of chances, because in theory it is easier to replace a corner outfielder than a centerfielder.

 

The positional adjustment can be argued in the infield sure, but I don't get the 10 run switch between left or right and center. Center is tougher to play, but more offensive production has come from center than from the corners on the top end, which theoretically raises the replacement value, and should account for  a lower adjustment, or a higher adjustment for the corners. Despite the fact that 10 of the best 20 hitting OF in the Majors over the past two years played CF.

 

This is incorrect. Trout would, in theory, have the same WAR total in LF as he would in CF if he got the same number of chances on defense. This is because while there is a 10 run difference in positional adjustment there is typically a 10 run difference on defense between an average CF and an average LF.

 

In other words Trout might save 20 runs over an average OF. An average CF would save 10 runs and an average LF would save 0. If he plays center his dWAR is +10, in LF it is +20. In LF he is hit with a -7.5 run positional adjustment, while the CF is credited with +2.5 - either way Trout finishes at +10 in the overall calculation.

[AZMike]

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.

 

The fact that Weaver's value is being disproportionately assigned to the defenders (and the defenders are arguably seeing their WAR increase) means that WAR by itself cannot be used as a one-stop-shop number to quantify that one player is a better overall player than another.

 

I personally prefer Win Probability Added (WPA), or context-netural win probability added (WPA/LI). These seem to agree with the "eyeball test" significantly better (particularly in my earlier example of of Weaver vs Kendrick).  I think stat is a much better starting point than WAR for fans who want to argue about who a team's MVP is.

 

 

Both figures are dependent on context and teammates etc. and both stats make arbitrary judgment calls when it comes to dividing up credit for plays, but in the case of win probability, the term "win" is actually appropriate. You can by definition sum up the win probability of all the players on the team and end up with the team's overall winning record. I agree with others who said that there might be less confusion if WAR were named something that didn't include the word "wins".

[Docwaukee]

I like WPA and WPA/LI  especially for relievers.  I think it's helpful is showing how important a good bullpen is relative to the number of innings they throw. 

[Lou]

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.

 

Just how does WAR measure these?

[Glen]

"This “replacement player” who constitutes the very linchpin of the entire premise is mythical. There is nothing measurable or precise about his existence. Yet supposedly intelligent people have signed off on this utterly bogus piece of baseball idiocy."

 

 

 

Just reading that made me feel smarter and younger than Bob Ryan.  Reading the rest of the aritcle confirmed both of those feelings.

[Lou]

i prefer WOB

 

(Wins Over Boogeyman)

[Tyler]

Just how does WAR measure these?

 

 

With a manometer.

[Lifetime]

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.

I was speakimg of the "above replacememt" numbers. 

[YouthofToday]

GregAlso, you must be new. The paragraph mentioning Erstad, grit, "The Professional Hitter", CERA and Mathis was total, 100% sarcasm.

[Lou]

GregAlso, you must be new. The paragraph mentioning Erstad, grit, "The Professional Hitter", CERA and Mathis was total, 100% sarcasm.

 

at least

[Docwaukee]

i prefer WOB

 

(Wins Over Boogeyman)

He said you ran away from him as if you were the boogetyman

Boogeyman

Boogey?

I'm quite sure.

[Docwaukee]

They could have named it widgets over bigfoot.  It's somewhat arbitrary.  They tried to give it a context that they thought would be helpful.  

[YouthofToday]

I've been one of the longest and most vocal supporters of using WAR as a player comparison in this community. That said, I was floored to learn Oakland's internal comparable stat had Miggy > Trout in 2012.

What is publicly available is like 3rd grade math compared to Math PhD folks running proprietary systems run on supercomputers used to simulate nuclear bombs and super novas MLB employs.

Yes, I exaggerated. Point stands. However, WAR is a heck of a lot better than RBIs, batting average, gold gloves, CERA or other preschool level counting stats.

[angelduck]

Really bad article. The replacement player makes perfect sense when you think of it as a player who is on a AAAA level. Vernon Wells was worse than the AAAA outfielder in 2011. WAR really does make perfect sense, and it is effective when evaluating a player's total game (Baserunning, hitting, fielding). I personally prefer wOBA and WRC+ for offense over WAR though. And WAR for pitchers cannot be compared to WAR for a positional player. So the whole argument that WAR sucks because it says Kendrick =Weaver is bogus. The whole point of the stat is that it gives us a guideline of how successful a player will be in the future. This is why RBI's and batting average suck so much. A player could have 100 RBI hitting 4th in the Tigers lineup, but the same player could have 30 RBI batting first in the Astro's order for example. That's the same player, only difference is that one of the player's is in a shit lineup and the other is in a good lineup. It's useless, and it's kind of a shock to me that the MLB still uses them every time a player steps to a plate.

[Lifetime]

stats are good when understood and used appropriately. stats are worthless when misused. and no stats should be used in a vacuum, including the new sabermetric stats.

[Lou]

He said you ran away from him as if you were the boogetyman

Boogeyman

Boogey?

I'm quite sure.

 

He won't share Desperado. Hey, what do you think of Oye Como Va?

[Tyler]

[Docwaukee]

The Japanese guys had sake in the
hot-tub! You gotta get 'em outta the drawers and get 'em down here, or I
don't have a focus group to sell the pilot to Japanese TV!

[GregAlso]

Sorry YoT, I missed the sarcasm. This whole discussions seems silly. I haven't been on the boards a long time. I came to AngelsWin via twitter.

Lou, As far as how they measure those things. UZR/150 or Defensive Runs Saved measures the defense based on each play. They assign values based on where they were when the out was made (UZR) and I forget(read as don't want to look up right now) how DRS is calculated. As far as baserunning; each decision on the base path, good or bad, has a value and when you add up those values you can see how many runs were added or subtracted from their total comtribution. This is based on linear weights relative to the year your looking at.

Yes, AngelDuck, wOBA and wRC+ are much better to look at in terms of just offensive production. I believe wOBA is used in calculating one of the offensive stats in WAR, not certain of this.

[YouthofToday]

The defensive stats MLB clubs have are multiple standard deviations better than what is publicly available. I think it was in 2011 or so the Giants were the first club to have a boatload of cameras set-up all over the stadium to monitor all actions players took on the field (reaction time and first step direction after ball hits the bat, etc). The NBA has a similar thing. So does the city of London to monitor its own citizens.

The problem with our debate is we don't have enough data to truly rely on public defensive statistics. The data exists but is proprietary.

Gold gloves are about as worthless as Tony Reagins.

[xboom28x]

Before the WAR stat came to fruition, were what we used to evaluate players wrong? Are there guys in the Hall of Fame that don't belong because we overvalued them using stats like HR's, RBI's, BA, OBP, etc? Have we not been able to accurately determine how good a player was just by watching them play with our own eyes?  

 

WAR is fun. It's about as useful as the old GWRBI stat. Both have their flaws. That is just my opinion.

[YouthofToday]

Stats that were primarily used to evaluate players (average, RBIs, pitcher wins, etc) before Bill James and saber metrics are/were wrong. Jim Rice does not belong in the hall of fame. Jack Morris certainly does not belong in the hall of fame.

It's not just WAR... OBP matters a lot more than batting average... Woba is a lot better than OBP ... And so on.

If you don't properly use statistics you end up trading Napoli for Vernon Wells and missing out in Beltre, amongst other blunders (Gary Mathews Jr., etc.).

[xboom28x]

Missing my point YOT. I didn't need WAR to tell me the Wells deal was lopsided. I also don't need WAR to tell me that Rice and Morris are borderline HOFers. No one thought Matthews was a good signing. Again, WAR was not needed to confirm anyone's evaluation of that.

[Inside Pitch]

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

 

 

You seem to be getting hung up on the semantics, not saying that to be dismissive in any way, it just seems you just have issues with what is being defined as more than what it's actually trying to measure..  Would you like it more if they called it VAR?  Value over replacement?

 

WAR tries to assign a set $$ value to a win for the purpose of comparing value to a team.  Any player that does more to help you win (defense, playing time, offense), typically has more value.   People bring up that Harrison had a greater WAR than Weaver -- when you consider he pitched 24 more innings and may have done a better job of preventing runs after the park adjustments, it's really not as ridiculous as some are trying to make it.  Is there no benefit to a team getting more quality innings from a guy?  

WAR usez runs as the penalty/bonus simply because run prevention is just as important a factor as run scoring is.  Recent studies have shown that run prevention may actually have more of an impact on winning than offense.  The irony being that when it's all said and done all the math and new fangled formulas simply support the old notion that pitching and defense wins championships.  I understand your issue is with the values themselves but you cannot deny that position and defensive position has a massive impact on how valuable a player is.  A shortstop that will provide league average defense while hitting like a 1B is a massive boost to a team.  If you put a 1B at SS, how much would the pitching suffer?    If that SS is also a gold glover.  How much more valuable is he?   If that same team has a 1B above the league average for all players but below the league average for 1B, is it not dealing with a relative weakness at 1B?  Would that weakness be even more acute if he was also a poor defensive 1B?  That's the value WAR is trying to put a figure to.

If you were to consider the positional adjustment a BONUS to positions up the middle instead of a penalty for playing the corners does it somehow change things?

 

I think the biggest issues with WAR are tied into the defensive metrics in place.  They are still relatively new and much more volatile, as such there is criticism there. Even great defensive players can show some massive variances in the metrics from year to year.

 

Personally I don't see the mythical replacement player as a problem -- it's better to have that value be fluctuate based on league wide results than to have a fixed value that has little to no bearing on what the league tendencies are.   As an example, QUALITY STARTS.  If ever there was a stupid stat it's QS. it's defined by 6 innings and no more than three runs allowed -- a 4.50 ERA.   So, a guy goes 8 innings and allows 4 runs, still a 4.50 but no longer a QS.  A guy goes 9 innings allowed 4 runs, a lower ERA but not a QS.   If those values were based around the league average ERA -- then the QS data might be useful.

 

I think the people who have said that WAR is best used in conjunction with other stats have the best grasp on the situation.  It's a nice statistic to use for a quick and easy comparison, but it shouldn't be the end all be all.  

I don't think there will ever any one stat that can measure everything perfectly.