The solution for me is to go back to something that Bill James said in the 1980s - don't combine everything into one big number.

This is a good Jeff Angus piece from several years ago on the subject.

-- MWE

It depends whether you'd rather have your value grounded in reality (in offense, e.g., Hamilton's offense) or inherently uncertain (in defense, e.g. Barney's defense).

Everyone here would choose the former, all else being equal. Or if you wouldn't, or wouldn't care, please go on record.

The solution for me is to go back to something that Bill James said in the 1980s - don't combine everything into one big number.

Does Win Shares still exist, and if so, how do Trout and Cabrera do?

edit > when you google "win shares leaderboards" most of the hits are fictional baseball leagues.

It is buried deep within the bowels of James' site. I don't even see an individual leaderboard. But it has it broken down by team.

Trout: 33.8 hitting, 4.1 fielding = 37.9
Cabrera: 26.3 hitting, 5.4 fielding = 31.7

That certainly is interesting.

The more uncertainty you have in a metric the more you regress it.

It is buried deep within the bowels of James' site. I don't even see an individual leaderboard. But it has it broken down by team.

Trout: 33.8 hitting, 4.1 fielding = 37.9
Cabrera: 26.3 hitting, 5.4 fielding = 31.7

That certainly is interesting.

Other than the fielding score, the total jives much better with my thinking than War.

Trout: 33.8 hitting, 4.1 fielding = 37.9
Cabrera: 26.3 hitting, 5.4 fielding = 31.7

That certainly is interesting.

Very much so.

Up until recently Bill James had Win Shares on his pay site. [I see he still does, I should have noted that "up until recently" meant "up until recently, when I stopped paying"]

Probably a big reason WAR dominates now is that it's freely available.

I remember just four or five years ago this time of the year was super fun time plugging numbers into a spreadsheet to spit out my own deeply flawed version of WAR for the season, inputing some version of linear weights and a witch's brew of various Hardball Times defensive stats. That was super fun, though now with freely available WAR my laziness takes over.

Man, remember when B-Ref didn't have minor league stats, or wasn't updated daily? Seems like twenty years ago to me, but I guess that's just my naturally awful memory talking.

Kind of hilarious that Cabrera leads him in fielding.

Other than the fielding score, the total jives much better with my thinking than War.

Jibes.

The solution for me is to go back to something that Bill James said in the 1980s - don't combine everything into one big number.

Even leaving aside James' later change of mind (Win Shares), this is a classic of the genre of statements that sound profound but are in fact ridiculous. The nice thing about WAR is that it's components are readily transparent, so you can, for example, substitute your own defensive assessment if you aren't a fan of DRS. You could even, I suppose, demonstrate your boldly independent thinking by refusing to add up the components! But since they are added up in real life -- real players do hit, run and field, and teams actually hire individual players to do all of these things in a single season, and even in single games -- I'm not sure what the objection could be to arriving at an overall assessment.

Probably a big reason WAR dominates now is that it's freely available

That, and because Win Shares was a total trainwreck.

Cano's Win Shares are equal to Cabrera, BTW. 21.7 hitting + 10.2 fielding = 31.9

And the Cubs look like:

Starlin: 14.7 + 7.5 = 22.1
A. Soriano: 15.8 + 2.9 = 18.6
DeJesus: 12 + 2.8 = 14.8
Barney: 7.7 + 6.7 = 14.4

I definitely don't get how this is working, as the 2B Runs Saved leaderboard is: Barney 28, A. Casilla 15, Cano 14

Win shares rules!

Even leaving aside James' later change of mind (Win Shares), this is a classic of the genre of statements that sound profound but are in fact ridiculous. The nice thing about WAR is that it's components are readily transparent, so you can, for example, substitute your own defensive assessment if you aren't a fan of DRS.

But you can't easily do this for each player in a list of 15 (e.g., if you're trying to determine a player's HOF case). Or if you are going to do it, you might as well start from a baseline where all the players in the list are average defensively. Because in a list ranking players by WAR, the needle is all over the bleeping map with defense. At least with VORP or oWAR, you understand that quality of defense isn't included, and it's FAR easier to eyeball the adjustments.

You could even, I suppose, demonstrate your boldly independent thinking by refusing to add up the components! But since they are added up in real life -- real players do hit, run and field, and teams actually hire individual players to do all of these things in a single season, and even in single games -- I'm not sure what the objection could be to arriving at an overall assessment.

The objection could be that a list where the players are ranked by WAR is more off base than a list where the players are ranked by VORP or oWAR.

Probably a big reason WAR dominates now is that it's freely available.

Win Shares is also freely available.

"The solution for me is to go back to something that Bill James said in the 1980s - don't combine everything into one big number."

Not a solution for players and teams, as all of their contributions result in a single number - salary.

Trout 38 WS, probably 30 WS above bench. Divide by 3- 10 WAR.
Cabrera 32 WS, 23 above bench (more games than Trout). That's 7.7 WAR. Bbref has about a 1.5 win gap more than that. If you believe speedy superman is 15 runs more valuable on D than the big slow guy, there's your difference.

Probably a big reason WAR dominates now is that it's freely available

Yes, which actually happens to be similar to my objection to Play Index not allowing us to pull up a list of players ranked by oWAR.

I would use oWAR to get a first cut of ranked players much more than I use WAR, but I'm shunted into WAR instead.

(Actually, I would first use VORP, but Baseball Prospectus's stats page is bizarrely not easily workable. If any BP guys are lurking here, it's a big problem for your business, and the pay aspect compounds it.)

Win Shares is also freely available.

That's odd, those numbers don't match what's on James' site. Please don't tell me we have competing versions of Win Shares too...

But at least it does still agree with what we all know, that Trout is the more valuable offensive player and Cabrera the more valuable defensive player...

But you can't easily do this for each player in a list of 15 (e.g., if you're trying to determine a player's HOF case). Or if you are going to do it, you might as well start from a baseline where all the players in the list are average defensively. Because in a list ranking players by WAR, the needle is all over the bleeping map with defense.

That's not really true for a HOF assessment. At the career level, the main problem with defensive stats is that they are artificially regressed -- everyone is relatively close to zero -- not that they are "all over the map."

For a single year assessment, yes, evaluating defense is challenging and there is a case for heavily regressing the fielding component. But at the career level, the story is very different.

But you can't easily do this for each player in a list of 15 (e.g., if you're trying to determine a player's HOF case).

Well, you can if you put the time into it. [Edit: which is where "easily" comes in I imagine. But I think this is the crux of the debate. You want a version of WAR that does all the leg-work of triangulating a bunch of different defensive systems for you to increased certainty. Which would be great. If someone had a site where they did that I'd probably visit it often.]

Kind of hilarious that Cabrera leads him in fielding.

Not to me. 3b is quite possibly an inherently more valuable position to play than CF to a team. In other words, Cabrera hasn't been bad enough to detract from the real value that having a 3b represents.

I mean, what compels you to have a player standing at that place in the field? Why not have an inner outfielder instead?

Maybe I missed this, but Ray you can get batting WAR or runs anyway from BBREF play index. You would have to add in Baserunning/DP runs.

Hill's RF/9 is 0.24 higher than Barney's. Barney gets to fewer balls vs. league than his fellow middle infielder, Starlin Castro, and Hill gets to more than his (though the D-Back SS is a statue).

Yet, Hill has 0.2 dWAR and Barney has 3.6. That simply can't be right. Doesn't pass the laugh test

.

The issue obviously is the denominator. If we could use PO/A numbers to get reasonable defense numbers, we wouldn't need BIP data. Or field f/x.

That being said - Barneys TZ numbers are 15 runs higher than BIS and 20 runs higher than Fangraphs (BIS')

It depends whether you'd rather have your value grounded in reality (in offense, e.g., Hamilton's offense) or inherently uncertain (in defense, e.g. Barney's defense).

Everyone here would choose the former, all else being equal. Or if you wouldn't, or wouldn't care, please go on record.

But there is no choice. We are trying to measure the baseball players' entire value, as it happened in the real world. You do not get to choose your own reality.

But there is no choice. We are trying to measure the baseball players' entire value, as it happened in the real world. You do not get to choose your own reality.

? We're dealing with this reality. Whose WAR are you more comfortable with as an accurate representation of the player's value? Hamilton's, or Barney's?

The solution for me is to go back to something that Bill James said in the 1980s - don't combine everything into one big number.

This is a good Jeff Angus piece from several years ago on the subject.

-- MWE

A QMAX sighting!

That being said - Barneys TZ numbers are 15 runs higher than BIS and 20 runs higher than Fangraphs (BIS')

Hill is tied for third overall in the National League in Total Zone Runs, and leads Barney 17 to 15. It's DRS that's the massive outlier.

I was astonished to find BPro's FRAA has Beltre at the bottom of my list of fourteen AL 3bs with at least 600 innings, with a -6.9. Carera is ninth out of fourteen.

No other system I've checked has him lower than 8th (RZR).

The solution for me is to go back to something that Bill James said in the 1980s - don't combine everything into one big number.

Another option would be not to treat research that is still a work in progress as definitive.

Sean does a great job with the new WAR. Unfortunately, the underlying data and the nature of statistical analysis means that it's impossible to get an answer with pinpoint, definitive accuracy. (Yes, impossible.)

That doesn't mean it's not a task worth pursuing. It also doesn't mean that the numbers we have now don't have value and don't provide us useful information. It just means that the results shouldn't be taken, or used, as the end of discussion. Instead, they should be used as the beginning.

Instead, they should be used as the beginning.

That isn't really how encyclopedias work.

FRAA has Aaron Hill at plus-25, Darwin Barney at plus-10.

FRAA has Aaron Hill at plus-25, Darwin Barney at plus-10.

FRAA has Adrian Beltre as the worst third baseman in the American League.

Adrian Beltre is making far fewer plays than he typically does. But so is Elvis Andrus, so FRAA is probably missing something.

FRAA, Centerfielders:

Mike Trout -- minus-2
Josh Hamilton -- zero.

That doesn't mean it's not a task worth pursuing. It also doesn't mean that the numbers we have now don't have value and don't provide us useful information. It just means that the results shouldn't be taken, or used, as the end of discussion. Instead, they should be used as the beginning.

What I don't get is that it seems like EVERYONE on this site understands this ... and yet the same thing keeps happening with people lazily treating the WAR numbers like gospel, and the same debate springs back up.

Instead, they should be used as the beginning.

Isn't that what happened here? Darwin Barney's bb-ref dWAR was used as the starting point, examined, and found wanting. It's out of proportion by scale and magnitude to oWAR, not confirmed by plays made as compared with other NL 2Bs and Cub shortstops, and not confirmed by metrics that use play-by-play data with sabermetrically sensible adjustments, like FRAA.

What I don't get is that it seems like EVERYONE on this site understands this ... and yet the same thing keeps happening with people lazily pretending other people are treating the WAR numbers like gospel, and the same debate springs back up.

FTFY, etcetera

I thought BP admitted about half a decade ago that FRAA was crap. Although I imagine they made adjustments since then.

BP Glossary:

The biggest difference between Fielding Runs Above Average and similar defensive metrics comes in the data and philosophy used. Whereas other metrics use zone-based fielding data, Fielding Runs Above Average ignores that data due to the numerous biases present. Fielding Runs Above Average instead focuses on play-by-play data, taking a step back and focusing on the number of plays made compared to the average number of plays made by a player at said position. The pitcher's groundball tendencies, batter handedness, park, and base-out state all go into figuring out how many plays an average player at a position would make.

Here is an example of the Fielding Runs Above Average spectrum based upon the 2011 season-for the sake of consistency, the players featured below all play the same position (center field):

Excellent - Jacoby Ellsbury 11.6
Great - Nyjer Morgan 5.5
Average - Marlon Byrd 0.6
Poor - Roger Bernadina -5.2
Horrendous - Melky Cabrera -13.2

The philosophy makes sense. So does DPS's. Before trusting outliers, I'd probably want confirmation from both philosophies.

yet the same thing keeps happening with people lazily pretending other people are treating the WAR numbers like gospel,

FTFY, etcetera

Yes, you might be right about that.

I don't think FRAA is a play-by-play metric. It is more like Humphrys' Wizardry in that it is based on adjusting traditional counting stats of PO, A and E.

It was completely reworked in 2010, so anything admitted half a decade ago is no longer valid.

Honestly, I'm really curious how the whole argument got reduced to WAR vs. Triple Crown. Objectively, all that separates the two players offensively is about 10 HR by Cabrera, which are mitigated by a ton of stolen bases by Trout. Anybody who has seen the two guys play should have some idea that Trout more than makes up the difference with his fielding. The debate always should have been over whether or not the MVP voting should consider defense, but then there are a lot of writers who like to pretend it does while ignoring it so...

The debate always should have been over whether or not the MVP voting should consider defense, but then there are a lot of writers who like to pretend it does while ignoring it so...

Mostly agree, but I say it boils down to 20 extra games, defense, base running, positional adjustment, park factors and narrative.

Honestly, I'm really curious how the whole argument got reduced to WAR vs. Triple Crown. Objectively, all that separates the two players offensively is about 10 HR by Cabrera, which are mitigated by a ton of stolen bases by Trout.

How much is 10 home runs - it's actually 14 - "mitigated" by 48-4 on the bases? (And FWIW Cabrera is 4-1 on the bases.)

How much is 10 home runs - it's actually 14 - "mitigated" by 48-4 on the bases? (And FWIW Cabrera is 4-1 on the bases.)

Didn't say completely mitigated, and I figured around 10 HRs better because he also has 20 more games. That gives Trout 41 extra net steals. I don't mean to say Trout has been as good offensively as Miggy, but that the gap isn't so far that Trout's defense doesn't bridge it (or at the very least make it a legitimate question between the two of them).

Identifying and explaining the outliers is an important part of the process. First, it can sometimes help to point out the flaws in the system. Second, after inspection, it allows us to appreciate special seasons from unexpected players. What it doesn't do is invalidate the whole system.

yet the same thing keeps happening with people lazily pretending other people are treating the WAR numbers like gospel,

FTFY, etcetera

There are those who do treat WAR numbers like gospel. They aren't a tremendous percentage of visitors to this site, but they are out there.

How much is 10 home runs - it's actually 14 - "mitigated" by 48-4 on the bases? (And FWIW Cabrera is 4-1 on the bases.)



Didn't say completely mitigated, and I figured around 10 HRs better because he also has 20 more games. That gives Trout 41 extra net steals. I don't mean to say Trout has been as good offensively as Miggy, but that the gap isn't so far that Trout's defense doesn't bridge it (or at the very least make it a legitimate question between the two of them).

The Miggy defenders and those attacking WAR (many of them the same people) are making a good case. But at the risk of going all SBB, I'll say this. If WAR is so flawed as to be off by ~40%, if Miggy really is better/more valuable than Trout despite a 40% lower total, then it is totally useless, even as a conversation starter, and I question how it ever came be accepted as a tool in the first place. IOW, those claiming Miggy is better, but WAR can be useful for some things remind me of the writer who in 2001 claimed Ichiro should be ineligible for ROY because of his prior experience as justification for voting for C.C first, and then put Ichiro second on his ballot.

Either Miggy is better and WAR, with at least a 40% error bar, is completely useless for anything, or there just might be something else going on.

Even leaving aside James' later change of mind (Win Shares)

He had a structure fairly similar to WAR in the 1983 and 1984 Abstract. And the intro to the 1984 Abstract player rankings section has a much more nuanced (and fairly lengthy) intro, talking about the pluses and minuses of a "great statistic".

#315 Somebody (Mr. High Standards IIRC) also has supplied a spreadsheet to calculate Win Shares yourself. Kind of neat in that if you want to recalculate them using (say) EQR for the offensive component you can.

There are those who do treat WAR numbers like gospel. They aren't a tremendous percentage of visitors to this site, but they are out there.

How can you tell? I think my web browser erroneously identifies me as a WAR gospel believer. Either that, or visiting the Baseball Prospectus site set a tracking cookie identifying me as such...

Also, WAR analysis can be interesting, but it's strange that it's come up in this debate at all. Trout was leading the batting title race for a long while and will likely finish a close second, hit 30 HR, maybe 50 steals at a phenomenal rate, 129 runs (way ahead of anyone else), despite missing the first month of the season, plus he's a Gold Glove contender in CF with several memorable highlight-reel plays. If anybody might be 40% better than Cabrera, it would be this guy.

If God has a perfect omniscient metric for judging on-field baseball value, I would be shocked if Trout's 2012 wasn't at least equal to Cabrera's, and in all likelihood it's probably better. Guess I won't know until the afterlife (unless God is watching this thread and cares to comment).

And for narrative, doing all this as a rookie in his age-20 season while reversing the in-season fortunes of his team ain't too shabby either. Frankly, Trout's narrative is a lot more interesting to me than Cabrera's triple crown, which is a neat bit of history/trivia but it essentially just boils down to simultaneously having a high average and being a slugger.

It's funny that people seem so interested in statistics here but they don't talk about them like a statistician. Of course there is a non-zero chance that Cabrera's season is better than Trout's. What we should be assessing is the p-value of Trout > Cabrera. In other words a WAR of 10 is really a probability distribution of values from 8-12 and a WAR of 8 is a distribution from 6-10.

There are various sources of error but defense is clearly the #1 source of error. When the defensive metrics agree one of two things could be happening: 1) The different approaches come close to measuring true talent 2) The defensive metrics all have the same inherent bias and are all wrong in the same direction. If option 1 is correct then averaging the defensive results (think Nate Silver and presidential polls) is the best approach.

Guess I won't know until the afterlife (unless God is watching this thread and cares to comment).

He's a pretty regular poster; He might show up.

If option 1 is correct then averaging the defensive results (think Nate Silver and presidential polls) is the best approach.

I cannot express my hatred of averaging results enough.

The example of BPro's FRAA and Beltre is a good reason to reconsider averaging as a method, at least a wholesale averaging. Clearly BPro is doing something very radical there, and I am intrigued to know what.

Fielding Win Shares Above Bench ranks Cabrera #2 (and Beltre #3) among AL 3bs, another anomaly, but one I think I understand.

And Lawrie's value (#6 in FWSAB, #1 in DRS and #5 in UZR) seems dependent on how one treats plays made in the shift.

Averaging seems to me more likely to taint any one system's results than improve on them.

Something else:

In the light of Ray's contrast of Cabrera's home runs against Trout's stolen bases, I looked at the different components of their batting lines, and considered how they are valued by wRAA. One reason Trout gets an advantage is a sizeable gap in IBBs. Cabrera has 17 to Trout's 4.

The problem, in using WAR in an MVP debate, is that WAR penalises the IBB recipient by (quite rightly) not awarding the full value of the walk. But I'm not sure that is being subjectively fair to Cabrera. It's not his fault that the bat was taken out of his hands; if anything, it represents TEH FEAR, which we like to laugh at, but suggests something about how the opponent values the two players. But potentially so does the reluctance to walk Trout.

One could interpret the low number of IBBs as indicating teams were afraid to put Trout on base, because that could just make matters worse; but they didn't even want to risk pitching to Cabrera, because that, too, could make matters worse. Have we got the balance of that right in WAR's coefficients?

Guess I won't know until the afterlife (unless God is watching this thread and cares to comment).
He's a pretty regular poster; He might show up.

I know that. I just don't know if he has time to read all these threads.

Averaging seems to me more likely to taint any one system's results than improve on them.

It's a different way of regressing to the mean. It will unfortunately compress defensive value to a level smaller than what I think the differences actually are. The error doesn't necessarily average out so you end up with a number that is regressed but still contains the same amount of error as your worst defensive system. In other words your system becomes more accurate but not more precise. For example suppose Trout's defensive contribution averages to +0.9 and Cabrera's to -0.5. Then your WAR numbers become 9.5 and 7. However, the range of the probability distribution is still (for example) 7.5-11.5 and 5-9 (assumed stdev, not calculated). Those distributions mean there is a 1% chance that Trout provided 6.5 wins more than Cabrera this year and a 1% chance Cabrera provided 1.5 wins more than Trout.

I'm with you that I wouldn't add any system I thought was more incorrect than correct. GIGO.

Have we got the balance of that right in WAR's coefficients?

I think you could lump that into error calculating baserunning wins.

I think you could lump that into error calculating baserunning wins.

I think the baserunning part is the most rock-solid component of WAR. Even more so than the batting component. With batting the main source of disagreement is the use of park factors. Is Anaheim really that extreme a pitcher's park? How many years should be used to calculate it? Should park factors be generic when we know parks have different effects on different players' strengths?

With baserunning park doesn't matter much. The opportunities definitely even out over the course of a season that we can be confidant that a player who goes 1st-3rd 80% of the time is better than one who makes it 40% of the time - for a series this may not be true, but for a season it works. If you have a normally good baserunner have a really bad year, you can probably explain it by looking at specific outs on base. Those can be fluky, but they represent real opportunities lost, and it is correct to include them as part of a single season metric.

Fielding is still the toughest to nail down.

Correct me if I'm wrong but I don't believe leverage is considered in baserunning wins. The breakeven point on stolen bases is situational, not an overall average. Slow baserunners with small totals typically are only taking the chance when a diminished success rate makes sense. The guys with large totals run all the time so I agree they probably regress to the mean. I agree that the error is small mostly because the % error is high on the players with low totals and low on the players with high totals. High%*Low is approximately equal to Low%*High

And for narrative, doing all this as a rookie in his age-20 season while reversing the in-season fortunes of his team ain't too shabby either. Frankly, Trout's narrative is a lot more interesting to me than Cabrera's triple crown, which is a neat bit of history/trivia but it essentially just boils down to simultaneously having a high average and being a slugger.

I don't disagreeing with anything you have said in comparing Trout and Cabrera. I am a Tiger's fan and while I would be happy for Cabrera to win the MVP, I think it should go to Trout.

That being said, I am genuinely surprised how Cabrera's narrative is "only" the triple crown stats. He willingly switched positions to enable the Tigers to sign a big bat at 1b (something even Capitan Jeter himself was unwilling to do). He clearly worked hard to be a less awful 3b, all the while keeping his batting stats up. His drinking problem seems under control, so he has the redemption angle. Not that this is a better story than a rookie putting up a season for the ages while seemingly turning around the Angels' season (as well as Pujols'), just that there are some nice story lines for Cabrera beyond the three stats.

I think you could lump that into error calculating baserunning wins.

It seems small to you, but given the 54-53 advantage to Cabrera in Rbat, it does have an impact on Cabrera's advantage in his strongest area relative to Trout. My interest is in interpreting the values displayed in WAR numbers.

Cabrera hits one home run per 15 PA, two XBH per 8 PA. For Trout that XBH figure is about 9 PA. 17 IBB possibly cost Cabrera a home run and a double. Or maybe he would have struck out 17 times. We don't know.

He might show up

Or not. He moves in mysterious ways, like Michael Young trying to field a ground ball.

He willingly switched positions to enable the Tigers to sign a big bat at 1b

I don't get this. Maybe, I'm missing something, but is it really true that the Tigers would not have signed Fielder if Cabrera said "I'm not playing 3b?" Doesn't Leyland just put Fielder at DH?

That being said, I am genuinely surprised how Cabrera's narrative is "only" the triple crown stats.

I've seen several writers mention the Tigers making the playoffs and the Angels not as a point in Cabrera's favor. Which is absurd in that (a) the Angels have won more games than the Tigers, (b) the Angels failed to make the playoffs because they crapped the bed in April, while Trout was in the minors, and (c) the Angels record in games that Trout has played is one of the best in baseball and easily playoff-caliber. But I agree with you, Cabrera has a perfectly fine narrative: Triple Crown, switching positions, team comeback to win division.

BTW - are we spending all this energy talking about the AL MVP because the NL MVP is clearly Buster Posey?

I don't get this. Maybe, I'm missing something, but is it really true that the Tigers would not have signed Fielder if Cabrera said "I'm not playing 3b?" Doesn't Leyland just put Fielder at DH?

I understand that it was important to Fielder to play 1b. Everything I have read is that the Tigers went to Cabrera to make sure he was willing to make a move (regardless of whether this was true or even necessary). So even if it is false, it certainly makes a nice story about Cabrera and what he has done to help the team this year.

Also, Victor Martinez is back next year and expected to DH. That will leave the Tigers with three players for two positions.

I've seen several writers mention the Tigers making the playoffs and the Angels not as a point in Cabrera's favor.

Good point. I have seen this argument a couple of times, but it has been much less than I normally recall hearing or reading, so it slipped my mind. Although perhaps I have just dismissed it as a junk argument (in large part because even though the Angels did nto make the playoffs, Trout helped them fight until the last days of the season).

#365, I haven't looked in detail - and haven't looked at the pitchers - but I think McCutchen is right there with Posey. (And Molina, while a notch behind, has turned into a player that I simply didn't think he would turn into.)

----

By the way, I ask again whether people would, right now, take Trout or Bryce Harper going forward. I don't have a point - this doesn't relate to the WAR discussion - I'm just curious. Do you base your decision on the great season from the 20 year old, or the good season from the 19 year old? I'd have to still take Harper.

I was goading. The NL has about 5 right answers.

366 - So, I'm to assume that Fielder would have turned down what was the biggest contract (possibly by far considering the reaction to the deal) if he couldn't play 1B? And fine, put Cabrera at DH. Worry about Martinez when he can actually play again.

366 - So, I'm to assume that Fielder would have turned down what was the biggest contract (possibly by far considering the reaction to the deal) if he couldn't play 1B? And fine, put Cabrera at DH. Worry about Martinez when he can actually play again.

I don't know if this was the case, but stories along those lines have been hinted at. I recall one article reporting/speculating that if Cabrera had not agreed to go along, it is not clear that the Tigers would have even approached Fielder. I tend to doubt it, but it is out there - and frankly, that makes the good storyline.

Even if it is not true, spinning Cabrera as the selfless team player who changed to a position he had not played in 5 years (not counting the 14 games at 3b in his first year in Detroit) sounds great when arguing the he should be the MVP. Why more people have not latched on to it is what I am surprised about.

By the way, I ask again whether people would, right now, take Trout or Bryce Harper going forward. I don't have a point - this doesn't relate to the WAR discussion - I'm just curious. Do you base your decision on the great season from the 20 year old, or the good season from the 19 year old? I'd have to still take Harper.

I would chose Harper, but that it because (a) I have much more exposure to him since I live in the DC area and (b) I love his attitude. He seems to have a just the right edge to him that makes me enjoy watching him and his reactions (personally, I loved how he handled the Ozzie thing). Maybe Trout has that edge as well and I just don't know about it.

Of course, if those were the top two players available in a draft, I would be more than happy to stuck with the number two pick.

I like Trout's short, quick stroke more than Harper's swing and I think Trout's a better athlete. Trout.

371 - Once you make someone say more than "Triple Crown", I've heard that from just above every Cabrera-for-MVP supporter. A lot of people have latched on to it. And even if he didn't want to switch, and it was the only way to get Fielder on the roster, I'm still not sure how much meaning it should have. If he was as much of a disaster as many people predicted, then he's only hurting the team. If he makes it to below-average, then he still gets overrated by the people whose opinion doesn't go much past triple crown stats.

Shouldn't you take Trout since he's hooked on to BODYARMOR Superdrink™!?

Seriously, I'd take Trout. He may never perform as well as he's performed this year again, but he doesn't have to to be a tremendously valuable and entertaining player. Harper's got crazy upside, but less certainty of getting there I'd think.

371 - Once you make someone say more than "Triple Crown", I've heard that from just above every Cabrera-for-MVP supporter. A lot of people have latched on to it.

I'd love to see some quotes, because I have not seen it or heard it except for one mention here on a different thread. I don't pretend to read a ton on this topic, but it seems like an easy one to make.

And even if he didn't want to switch, and it was the only way to get Fielder on the roster, I'm still not sure how much meaning it should have. If he was as much of a disaster as many people predicted, then he's only hurting the team. If he makes it to below-average, then he still gets overrated by the people whose opinion doesn't go much past triple crown stats.

I think you are way too down in the weeds on this. The story is that he took one for the team by making a move that he perhaps did not want to but that seemed, on its face, to be good for the team. To someone focused on the Triple Crown stats and who hates new fangled ideas like WAR and the like, this sounds like an intangible that supports voting for Cabrera. Add in that he was not historically awful at the position (and probably looked decent at times) and you have the non-thoughout argument that he is helping the team. Jeter would not do it, yet Cabrera did thus demonstrating leadership and character. I don't believe this, yet am surprised I just have not seen that position advocated very often. Then again, maybe it has been I have just not seen it.

Jeter would not do it, yet Cabrera did thus demonstrating leadership and character. I don't believe this,

I do.

377 - To be clearer, I do believe it was Cabrera being a team player and showing leadership. I don't believe it should move him into getting the MVP over Trout.

"377 - To be clearer, I do believe it was Cabrera being a team player and showing leadership. I don't believe it should move him into getting the MVP over Trout."

I agree. Cabrera deserves a lot of credit for being willing to switch positions, and handling 3B better than anyone thought he would. He had any easy out too - remember when he took a ground ball to the face in his first game there? It would have been hard to blame him if he didn't want to go back there. His position switch though was not necessary to sign Fielder. The two could have split 1b/DH. I don't think the Tigers back away from that move if you tell them it means Delmon Young is out of a job.

I agree. Cabrera deserves a lot of credit for being willing to switch positions, and handling 3B better than anyone thought he would. He had any easy out too - remember when he took a ground ball to the face in his first game there? It would have been hard to blame him if he didn't want to go back there.

When that happened, I figured the experiment was over. I was very surprised he continued with it.

His position switch though was not necessary to sign Fielder. The two could have split 1b/DH. I don't think the Tigers back away from that move if you tell them it means Delmon Young is out of a job.

I can't imagine it either. I recall reading an article about the Tigers consulting with Cabrera before the signing to make sure he was on board. The article hinted that if he had vetoed such a move, the Tigers would not have made the offer, in part because of Victor Martinez returning in 2013. I also remember laughing and thinking that they would be happy to deal with that issue then.

To return to the odd rankings of AL 3bs, I think I've figured out why Miguel Cabrera ranks highly under Fielding Win Shares Above Bench.

Here's a ranked list of AL 3bs, having at least 600 innings at their position:

Moustakas
Cabrera
Beltre
Lawrie
Seager
Inge
Hannahan
Donaldson
Callaspo
Youkilis
A-Rod
Middlebrooks
Plouffe
Betemit

And here is a second list:

Moustakas
Cabrera
Beltre
Lawrie
Seager
Youkilis
Callaspo
Hannahan
Inge
Donaldson
Plouffe
Middlebrooks
A-Rod
Betemit

Fairly close, don't you think?

The first list is ranked by Fielding Win Shares Above Bench.

The second list is ranked by DPs turned.

I don't quite understand how, but for this season ranking AL 3bs fielding by Win Shares seems to track their ability to turn the double play.

fp, interesting (although of course it'd still only explain half of the equation, the evaluation of Trout's defense being the other.) I put the Trout/Cabrera Offensive/Defensive Win Share comparison question to James in his reader mailbag. He seems to have "pass"ed on it ;-)

It's funny that people seem so interested in statistics here but they don't talk about them like a statistician. Of course there is a non-zero chance that Cabrera's season is better than Trout's. What we should be assessing is the p-value of Trout > Cabrera. In other words a WAR of 10 is really a probability distribution of values from 8-12 and a WAR of 8 is a distribution from 6-10.

Colin Wyers did something like this with hitting stats. I totally forgot about that article until your post jogged my memory. Thanks. It is one of my favorite sabermetric pieces of the online era.

I think it is interesting how the evaluation of defense has changed over the years not just due to different metrics available over time, but also due to evolutionary pressure towards the mean.

Evaluation/Evolution:

When "The Hidden Game of Baseball" came out in '84, I was fascinated by the first rudimentary attempts shown at using Linear Weights to evaluate offensive and defensive contributions. Offensively LW has held up fairly well over the years but defensively, not so much. At any rate -at the time one of the things that intrigued me was Rabbit Maranville. According to LW, he should have been the NL MVP in 1914 - the "Miracle Braves" season - because even with a negative bat, his defense was saving so many runs above average (almost 60)that his overall ranking was the best in the league. At first my reaction was "No way could somebody be 60 runs better than average in the field" - but then I began to think about Stephen Jay Gould's comments about evolution as applied to baseball. Looking back at it now - 1914 was almost a century ago now - I have no doubt that the "average" player then was significantly worse than the "average" player today. This would mean it could be entirely feasible that one truly outstanding player might really be 60 runs better than average, and there may also have been players 50 or 60 runs below average. I think with the level of play at the time it would be entirely possible to have such a wide range of variance. It would also maybe explain how Maranville was actually elected to the HOF by the BBWAA and not the Veterans Committee - he really was viewed during his playing days as an unbelievably exceptional fielder and +60 runs in a season would certainly indicate that.

Fast forward 50 years to the 1960s - I would think the level of play would have improved a great deal for a variety of reasons - racial integration & better fielding equipment for starters. Maybe at that time the range of variance between top notch fielders and the poorest with the glove would be about 60 runs total: +30 on the upside, -30 on the downside? This is in keeping with Gould's thoughts about evolutionary trends usually narrowing the variances. This is an important distinction to make, because according to LW, Bill Mazeroski was still having some seasons in the late '50s and throughout the '60s where he was +50 or more runs above average, with several others in the +40 range. This had a lot to do with building the narrative for Maz getting into the HOF. I think it is probably more realistic that he was at the +20 or +30 level which, when mixed with his slightly negative batting results (accounting for position), might still result in an All-Star player, but not necessarily HOF caliber.

Fast forward another 50 years to today's game - is the variance in fielding down to about 40 runs now? +20 on the upside and only -20 on the downside? I think many here at BBTF probably view it as such and some may even think we are at a point where the range is even narrower than that. How about only +10 at the top and -10 at the bottom? This becomes critical to the whole Trout/Cabrera comparison. If Trout is really +25 or +30 runs defensively and Cabrera is -10 or something like that at his position - that's a 35 or 40 run difference that could certainly overcome any edge Cabrera might have offensively in determining who actually helped his team more during the season. But if the gap between great and terrible fielders is only at most 20 runs these days - and if Trout is, say, only really at +5 or +10 and Cabrera at -5; then I can see how some would think that's not enough of a difference to sway the MVP vote in Trout's favor.

Furthermore, if the gap between good and lousy fielders has narrowed this much - shouldn't this become a whole new paradigm in lineup selection? Shouldn't the days of Omar Vizquel-type players be numbered? Shouldn't teams pretty much always go with the "good hit, no field" player over the "good field, no hit" player (if they are stuck with having to make such a decision) because there is a much larger gap offensively than defensively? Should Seattle bid adieu to Brendan Ryan?

I'm really curious as to where we go from here. Your thoughts?

The biggest conceptual problem with WAR is that risk-free replacement-level production is not obtainable in the marketplace. There are some good things about WAR, but that problem damages the model pretty badly. While the "old guard" may not have said exactly this, they've certainly noted things like the replacement player being fictitious.

If this is the biggest problem with WAR, then it is actually a damned good system. If you want to adjust the replacement level that is actually a fairly easy adjustment to make.

I think there is value in consistency and durability that is not captured in measurements like single season WAR, but it's also not captured by how the voters typically vote for the major awards, either. Otherwise you wouldn't have results like Terry Pendleton over Barry Bonds, etc.

Otherwise you wouldn't have results like Terry Pendleton over Barry Bonds, etc.

Well, IIRC, there were a couple of factors back in the 1991 MVP balloting:

1)The Braves surprised everyone by having such a great season when they were lousy the year before - at the time people didn't realize it was the pitching staff turning things around and maturing that made the biggest difference. Years of continued success later people could look back and realize this. But at the time a lot of people thought Pendleton had joined the team, had a great season, the team improved dramatically, therefore he must be MVP.

2)The uglier part of the voting process was revealed when Peter Gammons mentioned he had interviewed several writers who had voted for Pendleton instead of Bonds. According to Gammons, almost across the board the Pendleton voters he talked to said they recognized that Bonds had done much more to help the Pirates than Pendleton had done to help the Braves, but Bonds had won his first MVP the year before and the writers hated Barry's guts and so were looking for any reasons to NOT vote for him and Pendleton gave them an excuse.

It's not how a vote should be carried out, but it wasn't the first (or the last) time writers would exhibit this kind of behaviour.