By In Stuff

Explaining Cabrera, Trout and WAR

Several Brilliant Readers have written in to ask — I think in a curious and serious way — how it is possible for Mike Trout to lead Miguel Cabrera in WAR. They readily concede that Trout is a faster baserunner and a better fielder. Still, they wonder (and I’m lumping seven or eight people together here, so I’m generalizing) how 30 more points in on-base percentage, 100 more points in slugging and 20 more home runs could not POSSIBLY make up that ground.

One BR sums up the question neatly: “Last year, I understood the difference,” he writes. “Trout had a better on-base percentage, was only 42 points behind in slugging, hit only 14 fewer homers — I could see how his defense and base-stealing might make him the better overall player. But this year, it makes absolutely no sense to me at all.”

I think it’s a fair question and so I went to the incomparable Tom Tango for some assistance in explaining the math here (and by this I mean explaining it to ME). And let me say up front that, I readily admit it’s possible you will come away certain that WAR is just wrong and Cabrera absolutely is having the better season. I know a few BRs have come to think I don’t like Cabrera — and I don’t mean to protest too much, but it’s really not true. I don’t like Bowie Kuhn very much. I don’t like the NCAA very much. I don’t like white chocolate at all. But Cabrera: Love the guy. I think he’s the best hitter on earth. I love watching him play. I also love Trout, love watching him play, and I’m fascinated by the question of which one is better.

We’re going to break down Baseball Reference WAR by runs — that is, runs gained and runs saved. So let’s start with the obvious stuff first, the stuff Trout excels in.

* * *

Base running

Trout is a better baserunner than Cabrera. I think reasonable people acknowledge this. Baseball Reference figures baserunning runs based on stolen bases/caught stealing and various other events like going first-to-third on a single, second-to-home on a single, first-to-home on a double, tagging up to advance a base, and so on.

Baseball Reference determines that Cabrera, though not fast, is an above-average baserunner. He has stolen three bases without getting caught and has shown effectiveness on the bases. B-Ref has Cabrera’s base running worth one run above average.

B-Ref determines that Trout’s baserunning is worth six runs. He has stolen 28 of 32 bases, scores from second pretty much every time and is effective going first-to-third.

So far, I think it makes sense — no?

Trout leads Cabrera by five runs.

* * *


Last year, fielding was the most controversial part of WAR, I think (although, as you will see, there might be something a bit more controversial this year). Last year, WAR estimated that Trout was worth an extra 21 runs for his defense while Cabrera cost his team four runs with his defense. That 25-run difference is GIGANTIC.

This year, the difference is much smaller. Basically, the defensive measure that WAR uses shows both Cabrera and Trout to be below average fielders. Again, Cabrera is viewed as the lesser fielder but it’s not that dramatic a difference.

The system has Cabrera being worth 15 runs below average as a third baseman and Trout being eight runs below average.

Again, you can argue with particulars but it’s pretty easy to follow so far.

Trout Leads Cabrera by 12 runs.

* * *

Double Play Groundouts

Here’s a small adjustment based on hitting into double plays. Cabrera has hit into 16 double plays this year — about 12% of the time he’s been in a double play situation. Trout has hit into seven double plays — about 7% of the time.

Basically, Trout gets one extra run because he’s hard to double up, and Cabrera has two runs taken away because he hits into more than his share of double plays.

Trout leads Cabrera by 15 runs.

* * *

Positional Adjustment

Another minor adjustment, this one takes into account the position of the player. This is so that a shortstop will get more credit than a left fielder. This adjustment goes Cabrera’s way because he has played third base exclusively, a core position, while Trout has played 88 games in center, 42 games in left and three games as a DH. Cabrera gets two runs and Trout gets zero.

Trout leads Cabrera by 13 runs.

* * *


Now we get to offense — and I know what you’re thinking: There is simply no way that Cabrera’s advantages in on-base percentage and power do not make up 13 runs. In a way you are right. Let’s go to our resident expert, Tom Tango, for some explanations.

Start where Trout has an advantage.

“Mike Trout has 9 more doubles and 7 more triples than Cabrera. That is worth roughly 14 runs to a typical team, if those events occurred in typical fashion. Trout also has 16 more walks plus hit batters (excluding intentional walks), which is worth roughly five runs. In these categories, Trout has a roughly 19-run advantage.”

That would give Trout a 32-run lead. But now, finally, we get to what makes Miggy Miggy.

“Cabrera has twenty more HR, which is already a 28-run advantage. He also has five more singles, which is worth just over two runs. Cabrera has made 24 fewer batting outs, which is worth nearly 7 runs. Put it all together, and in these categories, Cabrera is 37 runs ahead.”

And there you have it.

Defense, baserunning and other adjustments: Trout +13

Doubles, triples and walks: Trout +19

Homers, singles, fewer outs: Cabrera +37

Total it all up and Miguel Cabrera leads by five runs.

Now, you could argue that Cabrera should lead by MORE than five runs. You might want to include clutch performance. You might want to include high-leverage situations* . You might disagree with the defensive rankings.

*The interesting but flawed statistic “Win Probability Added” — which adds up the value of every single play in a game — has Cabrera WAY ahead of Trout. Cabrera’s WPA is 6.6, Trout’s is 4.1. But before anyone gets too excited about the stat, Chris Davis actually has a 7.4 WPA and leads the American League.

Anyway, no matter how we got here, Miguel Cabrera should lead Trout in WAR.

Except, there’s another step …

Cabrera leads Trout by 5 runs.


Of the many, many ways that Bill James has contributed to the game, the most fascinating and controversial might be how he explains baseball as a contextual game. That was pretty subversive when he first started doing it. In many ways, Bill burst on the national scene because of what not seems only a mildly interesting prediction. He predicted that when Fred Lynn left Boston and Fenway Park his numbers would fall DRAMATICALLY. On May 25, 1981 — less than two years after Lynn hit .333/.423/.627 with 39 homers and 122 RBIs – James told Daniel Okrent at Sports Illustrated that Lynn would hit .285 and between 18 and 24 homers while with the Angels.

From 1982 to 1984, Fred Lynn hit .281 for the California Angels. His home runs: 21, 22, 23.

It was a brilliant prediction, but one that Bill didn’t think was a particularly strenuous. It was context. In 1979, Lynn hit .386/.470/.798 with 28 homers in 77 games at Fenway Park, which was then an extreme hitters ballpark. Away from home he hit .276/.371/.461 with 11 homers. Double those road numbers, add a little something for home field advantage (players do tend to hit better at home) and, voila, you have yourself a prediction for what Lynn likely would do in Anaheim. It seems semi-simple now.

But then? It was revolutionary. It was jaw-dropping. People generally expected Lynn to go to California and play exactly as he had played in Boston. It wasn’t that people were unaware that players played in vastly different environments, it’s just that few people too it to the next step, as a way to EVALUATE players. Everybody knew, for instance, that Fenway Park was a good hitters park and Baltimore had this amazing defense. But you didn’t hear people talk about the major role that played in, say, Fred Lynn’s MVP season in 1975 (he hit .368/.451/.609 at home, and .294/.347/.523 on the road) or Jim Rice’s MVP season in 1978 (He hit .361/.416/.690 at home with 28 homers; he hit .269/.325/.512 with 18 homers on road) or the fact that Orioles pitchers won SIX Cy Young Awards between 1969 and 1980. The context was known, but it did not interfere with the individual narratives.

Now, because of Bill and others, people think quite a lot more about context. Give you an example: In baseball history, only seven players have hit better than .360 with at least 42 home runs in a season — something Miggy is trying to do this year. Those seven are: Babe Ruth, Lou Gehrig, Jimmie Foxx, Rogers Hornsby, Barry Bonds, Todd Helton, Larry Walker.

Now do those seven seem alike to you in any way? No, you see the names and you are IMMEDIATELY struck the context. Ruth, Gehrig, Foxx and Hornsby all had their seasons many years ago, before the game was integrated, before air travel, before the prominence of the slider, before specialized pitching. Bonds, well, everybody immediately grasps the Barry Bonds context. And the Larry Walker and Todd Helton seasons are immediately viewed through the prism of Coors Field when that place was a circus act. I should say that both those guys were amazing hitters who hit brilliantly on the road — I actually think both have strong Hall of Fame cases — but my point is the context is something we just naturally understand now.

But, I would argue that this is only true to a point. It’s only true in extreme cases. In smaller doses, when it comes to minor adjustments, many people are not nearly as interested.

Which brings us back to Trout and Cabrera. Trout plays in a tough hitters park. And because he’s in the American League West, he plays 10 games in Texas (a good hitters park), Houston (a pretty good hitters park), Seattle (a pretty extreme pitchers park) and Oakland (a pretty extreme pitchers park). The three toughest hitters parks in the American League are probably Anaheim, Oakland and Seattle. Trout plays 101 games in those three parks.

Cabrera meanwhile plays his home games at Comerica Park, which long had the reputation as a pitcher’s park but the numbers no longer show that to be true. For three years now, Comerica has been a hitters park. Meanwhile, Cabrera plays his division games in Cleveland (more of a pitchers park), Kansas City (a good hitters park except for home runs), Minnesota (a neutral park leaning slightly toward hitters), and Chicago (a hitters park and very much a home run park).

There are three parks in the American League that have a ballpark factor of 105 and higher — 100 is neutral with every number above leaning toward the hitter. Cabrera will play 94 games in those three parks.

Now, you can take this for what it’s worth. You might not care at all. But WAR takes this into account. Tango explains:

“Baseball-Reference suggests that based in the parks they play in, you’d score ten percent more runs in all of Cabrera’s games than in all of Trout’s games. That ten percent is enormous. Baseball Reference has Cabrera as having created 145 runs, and Trout as 130 runs. A ten percent change will close that gap to almost nothing.”

And that’s why Mike Trout leads in WAR. While Cabrera leads in the raw numbers Trout plays in a tougher run-scoring environment. After all the adjustments are made, Baseball Reference has Trout being 54 runs above average, and Cabrera being worth 50 runs above average. When you add in the replacement level runs (same for both) and convert it to Wins Above Replacement, this is what you end up with, at least for today:

Mike Trout: 7.7 WAR

Miguel Cabrera: 7.0 WAR

A Final Thought

I was trying to explain this whole thing to a friend of mine — trying to explain it to him while working it out in my own head — and he said something interesting. He said, “I know that to be perfectly fair, you have to judge context. But I don’t think baseball is perfectly fair. And it’s too much math for me.” I took his point to mean that all these contextual adjustments leave him cold. I find them fascinating because we are trying to level the playing field, trying to figure out who is REALLY the better player, not who is the player who looks the best because of ballpark or the weather or whatever. He yawned.

“The guy’s having a better year,” he said. “I’ll grant you that it’s fairly close because Trout can do so many things. But when you start taking away Cabrera’s lead because he has a better hitting ballpark, you’ve lost me.”

My friend thought, all in all, that Trout had a better year than Cabrera last year. He followed that logic. But this year, he thinks its Cabrera, and it’s not that close, and he is unimpressed by the math that shows Trout ahead. I have to say, I think most people probably agree with him.

Print Friendly, PDF & Email

137 Responses to Explaining Cabrera, Trout and WAR

  1. JV Batboys says:

    It’s a shame that so many people feel that advanced stats that provide context are boring, or take away from the enjoyment of the game. More knowledge is a magical thing. Even if I don’t fully grasp all the nuance of many stats, I still love to pick them apart, and it doesn’t diminish the game on the field at all. It only enhances it.

    • Nick Smith says:

      Exactly. And the beauty of it is that even if you don’t want to get at all sabermetric–not in the slightest–baseball still has plenty for you.

    • Matt Kendall says:

      One of the infinite reasons baseball is such a beautiful and endlessly interesting sport is that it can be as simple or complicated to you as you want it to be.

    • Why not just slightly modify every ball park so that to a 100 level. I’m sure with enough programming and math wizards we could figure it out. Give Anaheim a dehumidifier (opposite of Rockies Humidfier) to keep the baseballs in.

  2. Aaron Reese says:

    I recently went to the Baseball Prospectus event at Kauffman, in which the analytics department spoke. We all kind of know that baseball teams have WAY better statistics than we have. They have field f/x and hit f/x, which tracks the catcher’s glove to gauge command, and other cool things most of us chumps can only dream of.

    Anyway, despite having access to all these succulent treats, they said that park factors are a pretty good base stat to help compare players across context.

    I’ve always been pretty skeptical of park effects, because when the park has good hitters in it (like Detroit), all of the sudden, the park effects indicate that it’s a hitters park. Shocker. That being said, when the guys who have access to double-secret statistics say that park factors are pretty good for cross-comparison, well, then they’re probably pretty good for cross-comparison.

    • Atom says:

      Correct me if I’m wrong, but good hitters/bad hitters should have no bearing on park effects at all. Baseball reference is comparing hitting and pitching stats on both the road AND home. If a team routinely pitches worse on the road, but scores more runs at on road than at home…why, runs are obviously harder to come by in that ballpark. And, vice-versa, if a team scores much more at home and gives up more runs at home, they look like they have a hitters park. The talent level of the team shouldn’t matter at all.

      (and please, someone correct me if I’m mistaken on how park factors are calculated!)

    • You’re not wrong, Atom.

    • basebawful says:


      Park factors not only compare stats both at home and road for the home team; it also compares visitors stats at home and road.

      Park factors are also diferent depending on handness.

      Angel Stadium is almost neutral for LHB with a 97 rating on runs. So Josh Hamilton can’t complain that much about his horrible play due to moving to Anaheim (though the 87 rating on HR’s makes Hamilton’s approach of swinging always for the fences, a losing strategy).

      On the other hand, Angel Stadium is death to RHB with an 87 rating on runs and 82 rating on HR’s.

    • BobDD says:

      Those that believe that Cabrera’s hitting is part of what makes Detroit a hitting environment misunderstand Atom’s point, and need only look at Miggy’s splits:
      (for last three years as earlier mentioned)
      2013 – OPS+ is 22 better away
      2012 – OPS+ is 33 better at home
      2011 – OPS+ is 25 better away

      So for the last three years Miggy himself has hit for 14 OPS+ better on the road, though to be fair, he could have gone against that average because park effects are not individual stats. As you can see his results were so very different in 2012. It is less rocket science than most realize, and is freely available on the net.

      I usually use

    • wee 162 says:

      Yeah, park effects are something which I think needs a bit more looking into. I understand where it comes from, but there have been zero changes to Comerica Park other than adding a couple of scoreboards being added in the 13 years since it opened and I don’t buy that it’s became a hitters paradise in that time. It started off as a 96 (where 100 is average) lasted that way for 3 years, then was a 97/98 for another 3 years, before becoming average in 2006, then becoming a 101 to 102. (link to it is here –

      You look at every other team in baseball and they all have the same type of thing happening where there is some variance in ballparks which haven’t changed. Both Fenway & Wrigley have the same thing which has happened, the dimensions there haven’t changed for a long long time.

      What I do think has the biggest effect on changing the figures is not who hits there, but who defends there. The Tigers have been pretty good offensively over the last few years, but they have not been excellent when on defense. Now those same defensive plays aren’t been made at anywhere they play for 81 games a year away from home either, but they’re a much smaller sample in those stadiums so they’re spreading out and not having as much of an effect on other ballparks.

    • Unknown says:

      You have a point, but there’s more to park effects than stadium design. Weather, for instance, could have a big effect: a cold, rainy year in Detroit could make it a worse hitter’s park than it would be in a year with normal weather.

      Also, parks are relative. If newly-built parks are more pitcher-friendly than average, pre-existing parks will get relatively hitter-friendly over time.

    • @ wee 162; didn’t they dramatically move the fences in at Comerica?

    • Rob Smith says:

      The fences at Comerica were moved in as far as 25 feet in some areas. Somewhere between 2003 and 2005, I think.

    • Atom says:

      But wee, none of that matters. The only thing that matters is “are runs more scarce at this park than average. If yes, then a run has more value and a lower park factor. ” Those other things may be interesting reasons WHY the park factor changed, but has no bearing on the park factor itself.

      The fact is, more runs are scored at Comerica by both the Tigers and visitors, and this has been true for years. Therefore, a run if Tiger stadium is not worth as much as a run in Angel stadium.

    • wee 162 says:

      Thanks for the clarification re the fences moving in folks, it was something I did a quick check on and never seen anything, and I couldn’t remember anything about it. What I would say is that having checked it was in 2002, and that isn’t when Comerica became relatively higher scoring according to the Fangraphs stuff I linked to earlier…

      And I take the weather point as well, that was a variable that hadn’t occurred to me, and could offer some explanation. I also take the point that when other stadia change, or new ones are opened that changes the distribution. However, Angels Stadium has been substantially untouched in terms of anything which should make a difference since 1998. I’ve had another bit of a look through and Cleveland suddenly became a worse hitters park apparently in 2001. Now we can talk about it not being about the hitters, but that was the point that Manny Ramirez, and David Justice left the Indians… Comerica might be looking like a good hitters park because they’ve got a lot of good hitters, and not so great defense in my opinion.

      I’ve also had a quick look at the Angels this season, and they’re hitting substantially better at home than away (they were slightly worse last year). It could be that this year we’ll be saying based on park effects that Angels Stadium is actually a good stadium to be hitting in… Coincidentally the Tigers aren’t showing as big a split as they have been over the last few years between home and away so maybe Comerica heads back towards average as well…

      Anyway, the substantive point I think I’m trying to make is that WAR is partially calculated after working out park effects. If there’s any errors in that then it will show up in WAR but it won’t necessarily be obvious.

    • matt david says:

      Good points…If you took the best players at their positions and created some super team, the park they play in would instantly become a better hitter’s park.

      WAR is starting to look more like RBI or pitcher wins…it tells part of the story but not the whole story.

    • Gene Claude says:

      Also, park effects are relative. In other words, a park could stay exactly the same, but other parks change, making them more pitcher friendly. That would make the unchanged park “more” of a hitter’s park, relative to its peers.

    • Dinky says:

      Most teams hit better at home than on the road. They have a lot more games to get used to the backdrop of the pitcher’s release point. To a lesser extent that runners and coaches know better when a bad bounce off a wall might allow an extra base. At home players can sleep in their own beds, have routines that keep their body rhythms more stable, are not in a hotel at night with nothing to do but go out. That all is included in the park factor; it is the difference between the average team’s improvement at home and your team’s improvement at home.

      I despise the Triple Crown; it is outdated, team context sensitive, and does not easily compare a guy whose main job is to score runs (like Mike Trout last year) with a guy whose main job is to drive in run (like Cabrera). Last year, though, Cabrera was on a playoff team, and Trout came up late, and it’s an older player versus a younger player; if it’s close, the older player gets the MVP (see Yount versus Sierra, to Sierra’s detriment).

      This year Trout is being hampered by factors outside his control. The big factor is Peter Bourjos has been hitting well enough to move Trout from a very important defensive position (center) to a very unimportant one (left). Bourjos catches more balls than even Trout. So Trout has a lot fewer opportunities to improve his defensive WAR. And at this stage of their careers, Hamilton’s a better defender than Hunter, so even when Trout is in center, Trout gets fewer chances. Similarly, Shuck is better than whoever the Angels had in left last year (a converted first baseman a lot of the time). Last year was as good it gets for a center fielder wanting to look good. He’ll also steal fewer bases for a variety of reasons. One of those reasons is instead of Hunter batting behind him (high batting average, making being on second base worth more) he has Hamilton (more power, lower batting average). Pujols last year had a much better batting average as well. That creates an offensive context where the risk of stealing is not worth as much as being on first base for a two run homer.

      Bottom line, the Angels are not going to make the playoffs. Trout’s year is not quite as good. Cabrera’s year is every bit as good. I love Trout’s game, but I’d have to pick Cabrera for MVP for this intangible alone. If the Angels were making the playoffs, I might see my way clear to changing my mind. And the season is not over yet.

    • Atom says:

      I’m not explaining something very well. It’s about runs scored at home V away for both the home and and their opponents.

      Our example will be the Angels. They are averaging 4.32 runs per game at home and 4.63 runs per game on the road (they have played 5 more games at home, hence they have scored more runs at home). Their pitchers have given up 4.76 runs at home and 4.70 runs at home. Add it together. An Angels game averages 9.08 runs at home and 9.33 runs on the road. That means an Angels game is likely to see 2.75% runs per game at home than on the road. To believe that park factors are related to hitters is to claim that A). These hitters only hit better than their home park and B). These pitchers pitch worse in their home park.

      Previous years for the Angels games are follows (combined total for Angels AND opponents)
      2012: 657 @ Angels Stadium, 809 away from Angels stadium
      2011: 592 @ Angels Stadium, 708 away from Angels stadium
      and so on…
      Simply put, more runs are scored when the Angels are playing away from Angels stadium. Therefore, a run is worth more in Angels stadium (IE, a run in a 2-1 game is more valuable than a run in a 10-5 game…it makes up a larger percentage of the score). Angels stadium is a pitcher’s park.

      And Comerica?
      2013: 630 at Comerica, 562 away
      2012: 722 at Comerica, 680 away
      2011: 771 at Comerica, 727 away

      Who is on the team doesn’t matter. The Tigers players play both at home and away, the opponents play both home and away. The only thing that matters: Are runs more plentiful in this environment then when the same people play AWAY from this environment? If yes, then it is a hitter’s park because a run does not make up as high of a percentage of the total number of runs scored.

  3. Atom says:

    I like WAR, and I think it’s easily the best single stat to look at when evaluating players. But it certainly doesn’t make it infallible and occasionally laughably off base. My favorite example of this:

    Guess which teams currently ranks 2nd in NL in Offensive WAR?
    Answer: The San Francisco Giants, at 20.4 oWAR…just 0.8 behind the 1st place Cardinals
    …a Giants team that currently ranks 14th out of 15 in runs scores.

    I know that SFG is a big time pitchers park, but this is down right ridiculous!

    Regardless, this is a big-time outlier. What WAR should provide is a number that helps you better put into context positional adjustments, baserunning, fielding, park factors, etc. It’s a starting point…one that’s far superior to OPS, or, god forbid, batting average. If someone has a 7.7 WAR and someone else a 7.0 WAR, that is nowhere near definitive proof for me that one player is having a better year than another. I’ll look to other numbers to see who I think is better. But if someone has an 11 WAR compared to a 7 WAR…I think the system is good enough to pretty well trust a gap that large (and smaller, less you think I demand a 4 win margin!)

  4. Peter Bass says:

    Just flipping around on…in the statistic of “Offensive WAR” Cabrera and Trout are neck-and-neck. Surely that must complicate matters? If Trout’s baserunning and defense are worth so much, how come they don’t put him miles ahead? I guess the replacement player’s defense and baserunning are better than his hitting. In the (assumedly) simpler stat “Runs Created” Cabrera leads Trout 145 to 130, which seems about right. But if Cabrera creates more runs, why doesn’t he create more wins? I guess I can see why people find this so confusing.

    • Dinky says:

      Cabrera creates more runs playing in ballparks where, on average, more runs score. Trout creates his runs playing in ballparks where, on average, fewer runs score. Still confused? Yaz won a triple crown with numbers that would not put him atop the leader board in any category in Cabrera’s triple crown year, because when Yaz did it, pitchers dominated and offense was harder to achieve. If you can understand league wide offense between then and now, then you can also understand ballpark offense differences between Detroit and the other Los Angeles.

  5. Peter Bass says:

    Just flipping around on…in the statistic of “Offensive WAR” Cabrera and Trout are neck-and-neck. Surely that must complicate matters? If Trout’s baserunning and defense are worth so much, how come they don’t put him miles ahead? I guess the replacement player’s defense and baserunning are better than his hitting. In the (assumedly) simpler stat “Runs Created” Cabrera leads Trout 145 to 130, which seems about right. But if Cabrera creates more runs, why doesn’t he create more wins? I guess I can see why people find this so confusing.

    • What would be helpful is if someone took the time to write an essay that tackles this, going step by step with the different components of WAR, comparing Trout and Cabrera. Wow, if someone did that, perhaps it wouldn’t be so confusing.

    • invitro says:

      I only know a little bit about WAR and its factors, so I may get this wrong.

      “But if Cabrera creates more runs, why doesn’t he create more wins?”

      Have those Runs Created numbers have been park-adjusted yet? If not, that might be the reason Cabrera is 15 ahead, and why that is the case while they have almost the same oWAR.

      “If Trout’s baserunning and defense are worth so much, how come they don’t put him miles ahead?”

      Can you quantify “so much” and “miles ahead”? It’s hard to answer that question without knowing if you’re asking “why isn’t he 0.3 ahead in WAR?” or “why isn’t he 3 ahead in WAR?” or “why isn’t he 10 ahead in WAR?”

    • BobDD says:

      I believe that people start with their own prejudice – in favor of one player or the other, or for or against Sabermetrics – and then based on that pick-and-choose which parts they will accept or not. Classic to accept the stats that favor the player you like better, but reject the stats that tout the other.

    • invitro says:

      I don’t think so. If someone accepts sabermetrics (which means accepting logic, mathematics, and statistics), then they are not allowed to pick-and-choose parts arbitrarily. Theory tells what things to accept, and to what extent to accept them. Now, sabermetric theory is not complete, though I understand that it’s essentially complete for hitting.

      In particular, a sabermetrician favors WAR and its major components over BA/HR/RBI not by their personal choice, but because that’s what sabermetric theory says they must do.

  6. Jake Bucsko says:

    I have a stat I like to pretend I made up (who knows, I may have) called Net Bases where you basically just try to add up all the bases a player contributes. TB + BB + SB + HBP – CS – GIDP = Net Bases. Anyway, as of the games tonight Miggy Cabrera leads Trout 395-386, which I bet is a lot closer than people would think given all the home runs.

    Like Joe mentioned, I love both these players and am fascinated by the constant debate over who is better. Miggy has the “clutch” label attached to him because he always seems to come through late in the game when it counts, those back to back home runs against Rivera were played nonstop, but here’s his slash line from the 7th inning on: .268/.373/.486, for an .859 OPS. Certainly nothing to be ashamed of, but WELL below his seasonal totals. Meanwhile, here’s Trout’s line: .329/.462/.527 for an OPS of .989 (leading the league in 7+ inning OPS: Eric Hosmer).

    Anyway, Joe always seems apologetic when he writes these Miggy/Trout pieces, but to be perfectly honest, I can’t get enough of them. Hopefully it’s a debate that lasts for years to come.

    • Grover Jones says:

      Do you include going first to third on a single? Tagging up on a flyball? Etc. Etc.? That would make it more precise, perhaps.

    • Dinky says:

      When I was playing Strat-O-Matic Baseball, I used to numbers to evaluate cards. The first was total chances to get on base (versus LHP or RHP) out of the 108. The second was based on 3 points for a walk or hit by pitch, 4 for a single, 6, 8, 10 for double/triple/homer. That was 30 years ago, give or take. These days I might tweak the numbers a little (a homer is probably worth more in an offensive context where my only average or below hitter versus LHP was Rich Gedman and RHP Ozzie Smith). Bill James published a number that IIRC used exactly your formula, also 30 years ago give or take.

    • James Martin says:

      I read about this stat somewhere recently. Basically every few years a journalist comes up with an idea like this and its pointed out to them that its a less useful version of linear weights.

      Note the article did not say it was a bad stat just that since there is already something better out there that you can look up why would you calculate it yourself. Pre-internet it may well have been worth the loss in precision for the gain in time.

  7. Jake Bucsko says:

    This comment has been removed by the author.

  8. “And it’s too much math for me.”

    That’s usually the culprit. People will accept, tolerate, and thus not dismiss math when it helps create a cool iPhone they play with, or helps optimize traffic signals to minimize your time sitting at red lights, or helps construct the ball park in which these games are played, or helps with virtually EVERYTHING in life.

    But they’ll be damned if they’re going to let these same mathematical principles tell them what’s what with baseball and baseball stats.

    If you don’t know the math, then study it, or ask about it, or have an open mind. But if you’re simply going to dismiss it because you can’t understand it, then STFU.

  9. Ross Holden says:

    I’m not yet a WAR expert, but it seems like a good stat, and as some have said, if I can have only one stat, I’d pick WAR.
    But has someone done backwards analysis to see if the combined WARs of the players equal lead to the wins a team gets that year? I think Joe or someone said that a team of replacement players should get 50-60 wins, the the combined WAR would be the win total greater than that. As an Astros fan, 55 wins seems about right for a team of replacement players.

    • Dinky says:

      Bill James tinkered with this for years. Total team WAR has decent predictive value. But it tends to break down when a team does crazy good or bed in tight ball games and blowouts. The Orioles last year were a good example.

    • Ross Holden says:

      Thanks. The Orioles last year came to mind as well. Makes sense, considering in Joe’s analysis above, it’s really just runs then eventually converted to wins.

  10. Ross Holden says:

    I would be curious to Joe’s and others thoughts on starting pitchers getting the MVP. We’ve all heard people say “I can’t give the MVP to someone who plays only once every 5 days”. But if their WAR is higher, they still add more cumulative value. I think starters don’t get enough consideration in MVP discussions.

    • Whenever, I hear somebody say that pitchers shouldn’t get MVP awards because they only play every five days, I point out that on days when they do play, they are involved in every play (at least every defensive play). Whereas a fielder is only involved in 1/9th of the at bats, some percentages of plays in the fields (depending on their position) and some percentage of base running plays. I think it kind of equals out.

    • Dinky says:

      When we did the math back in my Strat-O-Matic days of the 1980s, we calculated that most players were involved in (say) 700 plate appearances (their own) if they played every inning for every game. Starting pitchers who threw 250 innings (rarer these days, I’ll admit) with an excellent WHIP of 1 would be involved in 1,000 plate appearances. A few hitters, a very few, would get over 1,000 from times on base (affecting running) and fielding plays, since pitchers also field, affect the running game on defense, and some get on base. We had zero difficulty determining that an ace starting pitcher, in a single season context, was probably deserving of the MVP, just because they influenced more games. These days, inning counts are a little lower, but Kershaw (for example) is already over 200, hits pretty well, has a great move to first and is an excellent fielder; I think it extremely likely that he will affect more plays than anyone else in the National League if not the majors this season. His big shortcoming, of course, is that the Dodgers just don’t support him; he lost his last game, for example, throwing 5.2 innings, 9 strikeouts, 1 earned run, slightly lowering his ERA. But the bullpen gave up the 3rd run before the Dodgers scored their second, so it became a loss; in the other two games of the series against the Cubs, the Dodgers scored 6 and 4, either of which would have been a win. That will make it harder for Kershaw to win MVP.

    • Atom says:

      I hate the “every 5 days” thing for starters because it ignores a simple fact:
      -the Starter is involved in every single play on defense.

      A position player bats four or five times, runs the bases once or twice, fields 3 or 4 plays every day
      A starting pitcher throws 100 pitches every five days.

      These things are not equal. In each of those five days, the pitcher has a larger involvement on that particular game than the position player. In the end, it seems to even out right? That pitcher going every five days has as much impact, for good and bad, as a position player every day.

    • Ross Holden says:

      I agree with all y’all. I’m glad I’m not the only one who feels pitchers are under-appreciated in MVP voting. I was looking at Pedro’s ’99 year. Amazing he got 0 first place votes.

  11. Scott says:

    I wish I could find someone to answer some of my WAR questions, I have so many. Let me try a few, see if anyone has some ideas:

    1)If WAR is obviously flawed but good for context why won’t more stat people admit RBI’s and ERA which are also flawed are just as valid for putting pitchers and hitters into a rough sense of context?

    2)If defensive stats need a 3 year spread to be valid why are they factored into single season WAR stats?

    3)Why does WAR credit defense at all? If it’s Wins Above REPLACEMENT shouldn’t the opposite be true? I mean, most minor leaguers are younger than 25, most major leaguers older, which means a minor leaguer will almost always be lighter, quicker, and have faster reflexes than the guy he replaces. Shouldn’t replacement imply better defense and less offense automatically?

    4)If speed and defense are so valuable, why do big league teams pay DH’s better than catchers, shortstops, and centerfielders? I wrote about this on my blog a couple of months ago. This isn’t the 70’s where GM’s are laughing off Bill James, these teams have advanced professional stat departments and pay big heavy sluggers who don’t run or field more than key defenders. Why is that?

    • Jovins says:

      1) WAR is less flawed. RBI and ERA are extremely limited, and look at only a surface reflection of a player. WAR is much less limited – it at least tries to look at the whole picture. The flaws in WAR are matters of degree. Both ERA and RBI are limited statistics because they give all the credit to an individual for the result of a team activity – in one case, scoring a run, and in the other, preventing runs. The best hitter would have low RBI totals playing on a terrible offensive team, and good pitchers would struggle if they had to play in front of a defense full of butchers behind them. WAR attempts to isolate each individual play, so that the contributions from the rest of the team don’t cloud our judgement.

      2) Defensive stats need a 3 year judgement to be considered stable. It’s similar to batting average – there are players who have a fluke year in batting average and hit really well or really poorly. Chris Johnson likely isn’t a true talent .331 hitter, but his hits still happened and are still calculated into WAR. Likewise, sometimes defensive players make a string of plays they didn’t usually make, or miss plays they normally would. These plays happen, so they’re calculated into WAR. However, that doesn’t mean that a player is necessarily a good or bad defender based on one years stats.

      3) Ignoring everything else, if Cabrera played defense like Machado he would be a better player. Defense is a part of the game. Wins above replacement doesn’t look at the wins above replacement for each component – it attempts to look at the total runs a player produces, and then removes the runs a replacement player would produce. Even if a replacement player would have better defense than two major leaguers, there can be a difference in defense between the two of them.
      Even if replacement is better defense and worse offense, that just means that bad defenders LOSE more runs on defense and GAIN more on offense. As long as replacement level is set to a run total, it doesn’t matter where the runs come from.

      4) There are a variety of reasons. DHs don’t rely on athleticism as much, so as they age there could conceivably be less of a dropoff. This ties into the peak age for these players – for example, second baseman anecdotally peak early. So more of their value comes before their big free agent contracts. Andrelton Simmons is playing ridiculous defense this year, but he probably is at his peak as a defender. A big slugger will peak later.

    • Unknown says:

      1. Adam Dunn is a good but obviously flawed baseball player. I’m just a flawed baseball player. He gets paid because his talents outweigh his flaws, I don’t because mine don’t. Same deal with stats: just because nothing’s perfect doesn’t mean they’re all equally useful.

      No matter what question we’re trying to answer, we should use the best stat we have to answer it. There are very few questions for which that stat is RBIs, mainly because it depends on two totally unrelated factors (how well a guy hits, and how often the guys in front of him are on base) and by itself, it tells us nothing about either one. It’s like an equation with two variables: impossible to solve without more information. ERA has similar issues with defense and park factors. WAR, whatever its limitations, tries to cut out all the influence of context so we’ve got nothing but the player’s performance.

      2. I think the issue is that you’re assume a black and white split- valid or invalid- when really it’s more of a continuum in which the more data you have, the better job defensive stats (or any stats) will do in predicting future outcomes. If someone goes 5-for-5 in a single game, that’s obviously a great game, even if it doesn’t necessarily mean he’s a great hitter. If a player has a season of +20 defensive runs, maybe he had a great year in the field but that doesn’t necessarily mean he’s a great fielder.

      3. You’re right that average is much closer to replacement level for defense than it is for offense, but great fielders are still way better than your random triple-A callup. And of course, some major-league players are [i]below[/i] replacement level for defense, but stick around because they hit well enough to make up for it.

      Also, defense is more than raw physical ability. Technique, positioning, etc. all take time to learn. Lots of players get better at it in their mid-to-late 20s.

      4. Well, I don’t think anybody’s denying that hitting is still the most important part of a position player’s value, so it’s a lot easier to make up for bad defense with good hitting than vice-versa. The more interesting question is how much to weigh each part of a player’s game. 10 years ago, traditionalists and sabermetricians seemed to agree that hitting was way, way bigger than everything else combined- maybe 90% of a position player’s value, to arbitrarily pick a number based on what I remember reading back then. In a way, they were right back then, because we didn’t have even a half-decent way to measure defense, so even if it was important there was no way to know who was good and who wasn’t. Now, I’d say there’s been enough of a revision that hitting’s more like 65%- still the biggest factor, but the balance has totally changed.

      Also, I haven’t conducted a survey, but I’d guess that this change is reflected in salaries, too. Look at Carl Crawford’s monster contract, or Dustin Pedroia’s.

    • Unknown says:

      Grr, Jovins stole my thunder by finishing his reply while I was still typing! What impudence! What arrogance!

      Anyway, I’d add another detail to his very good answer to question #4: a lot of DHs are veterans at the end of monster contracts who’ve declined to the point where they can’t field anymore, but still get paid like the stars they used to be. In your blog post on the subject, you note that Adam Dunn makes more than the top 4 defensive shortstops combined, but that’s kind of misleading because Adam Dunn is vastly overpaid. I think if we went through it on a case-by-case basis, we’d find DHs are more overpaid on average (and when ARod starts DHing full-time in a year or two, it’ll get ridiculous).

    • Dinky says:

      In 1985 Don Mattingly won an MVP when he wasn’t the MVP on his own team. He led the league in doubles, total bases, sacrifice flies, and RBI. Rickey Henderson had a season for the ages, scoring 146 runs and leading the league with 80 stolen bases. Mattingly’s OPS was .939, OPS+ 156; Henderson’s .934 and 157. Meanwhile, George Brett led the league in slugging, OPS, OPS+ (1.022, 179) and IBB (31) but had Willie Wilson batting leadoff for him, whose OBP was more than .100 worse than Rickey and scored 59 fewer runs. In WAR, they ranked Rickey (9.9, an amazing total), Brett (8.2), Mattingly (6.4), and Wilson (2.0) (Boggs was second between Rickey and Brett). Mattingly also won a gold glove with a DWAR of -0.9 at first base; Rickey did not win a gold glove with a DWAR of +1.5 playing mostly center field; how much did that GG for his hitting affect his MVP vote, we’ll never know. Brett also won a GG at 3rd base with a DWAR of +0.4, also probably a mistake but less of one. I don’t think WAR even existed yet, though. The advantage of WAR is that it eliminates the chaff that confuses a decent year (9th in WAR) like Mattingly’s and an all time great year like Rickey’s. Cabrera’s 2012 was a great year; so was Trout’s. Either deserved the MVP a lot more than did Mattingly in 1985.

  12. Herb Smith says:

    This comment has been removed by the author.

    • BobDD says:

      Perhaps you mean that it would be tough for the casual fan who started with the same prejudices you have – none of those WAR outcomes are surprising to me . . . but I’ve been a stathead forever (slide rule in the 2nd grade worn out in three months), so finding out these things were exciting for me. The stats were and are much more fun for me than pitcher wins or RBI ever were. So I’m not bragging about anything about myself here, just acknowledging that my mathematics freak set me up to excitedly be open to advanced stats.

      Both your examples use pitcher wins and unadjusted ERA, but taking into effect the offense of the teams pitched for and the park effects changes things to what I would call a predictable state. And that is why the WAR is easy for me to digest. Downright exciting. Better than DVD Special Features!

    • Herb Smith says:

      You’re right. I was simply trying to point out why the “casual fan” doesn’t get it (or like it).
      For example, despite Koufax’s brilliance, WAR doesn’t include a lot of what many fans think are important.
      WAR is oblivious to:
      -an unadjusted ERA
      -ANY post-seaon heroics
      -ANY awards, including the league MVP, the World Series MVP, the Cy, being a starter in the All-Star game, etc.
      -whether your team won the pennant, or was even in the race at all

      I was trying to point that out. Not very well, I guess.

  13. Herb Smith says:

    I actually like WAR a lot, for all the reasons stated above. I think Bill James’s focus on CONTEXT is not only essential to understanding baseball, it’s important for society as a whole. Really.

    That said, you can see why the casual fan has a knee-jerk reaction against it; It’s not just Cabrera-Trout…in MANY famous cases, it doesn’t seem to pass the smell test.


Here are two:

    1. In 1963, Sandy Koufax had one of the greatest seasons of all time. He won the NL MVP, and won the MLB Cy Young unanimously.
He went 25-5 with a 1.88 ERA, and 11 Shutouts. Eleven. A WHIP of 0.875 in 311 Innings Pitched (!! yeah, 20 Complete Games), a 5-to-1 strike-out to walk ratio…I mean it was ridiculous. He led the Dodgers to the pennant.

    Did I mention he also threw a no-hitter? Oh, and in the World Series against the heavily-favored, two-time defending champion Mantle-Maris Yankees, he threw 2 complete-game masterpieces, winning both the opener and the close-out game of the Series. His team won the world championship, and, of course, he was the Series MVP. Kind of a good year.

    And according to WAR, a guy named Dick Ellsworth had a better year. 

    And should have won the Cy. Hey, 22-10, 2.11 ERA, it was a darn good year. But over 100 fewer K’s, fewer IP…he basically trailed Koufax in every single category.


Joe Poz has previously written about why Koufax may have seemed even better than he actually was, because of context. But it IS difficult for the mind to wrap itself around the fact that Ellsworth was slightly more valuable that year.

    2. Tom Seaver in 1969. The Miracle Mets. Tom Terrific. In the heat of the pennant race, Seaver won 10 straight starts. 25-7, 2.21 ERA, led the Mets to perhaps the most shocking world championship of the 20th Century. Won the Cy Young almost unanimously, came a very close 2nd in the MVP race, and was subsequently featured on every big TV show and every sport-based magazine cover that existed in ’69.

    And according to WAR, he was the 5th best pitcher in the National League that year.

    Behind Bill Hands and Larry Dierker, and WAY behind (a full 4 WAR behind) Bob Gibson, who’d had a seemingly mediocre season.

    That kind of stuff makes it tough for the casual fan to digest WAR.

    • Mike says:

      I really don’t like rWAR for pitchers because of how they account for defense. It can lead to some weird results. Let’s look at fWAR for those seasons.

      In fWAR, Sandy Koufax is 2.5 wins ahead of everyone else in 1963, so you probably wouldn’t have any argument with that:

      In 1969, fWAR has Seaver as 19th:

      There are a couple of reasons for this. One, Seaver’s 273 innings is good but not great (for the time). A lot of pitchers ahead of him threw more than 300 innings. Two, Seaver got a lot of help from defense/luck, as his babip was .232…thus his own personal value was lower.

      “Bob Gibson, who’d had a seemingly mediocre season.


      I’d like to know how 300+ innings with a 2.18 ERA is a mediocre season.

    • Herb Smith says:

      Because the year before his ERA was 1.12, a full run lower?

      I certainly wasn’t ragging on Gibby, just pointing out that the consensus was that he’d come back down from his perch on Mt. Olympus from the season before. And the voters obviously thought that too; in ’68 he won the NL MVP, and won the Cy Young unanimously. In ’69, he came in 30th in the MVP vote, and didn’t get a single Cy vote.

      BTW, one reason that Koufax finished below Ellsworth in all-over WAR in ’63 was because he was a truly terrible hitter, and that negatively affected his number. (I always thought that was odd, because Koufax was known to be a genuinely gifted all-around athlete; in HS he was better at basketball that baseball, he had balance, strength, spatial awareness…it’s odd.)

    • Dinky says:

      Herb, you get a great big *huh* from me. According to bRef, Koufax’s WAR as a pitcher for 1963 was 10.7; Ellsworth’s was 10.2. Ellsworth had a better year at the plate (-0.2 versus -0.8) and both tied at 9.9 for total WAR behind Willie Mays. Ellsworth had a higher ERA+ from park adjustments, but Koufax pitched more, which is why his pitcher’s WAR is better. 20.1 more innings counts for WAR, but not for ERA+. So for Cy Young award, Koufax was clearly the winner.

      Ellsworth’s was a fluke season for a guy who never had an ERA+ above 108 in any other season, who was expected to be the #3 starter for the Cubs, who never excelled before or after. The Cubs were languishing in 7th place; the Dodgers were absolutely dependent on Koufax in winning the pennant and then the World Series, and would have won neither without Koufax. Koufax won the pitching triple crown in 1963 and pitched more innings than Ellsworth; he was already noted as an excellent pitcher from his prior two seasons. Ellsworth was the fluke. Willie Mays led the league in WAR, but the Dodgers won the pennant and Koufax was the best player on the best team. I would not have complained (much) had Mays won the MVP, but then Clemente jobbed Koufax in 1966 when Koufax actually led the NL in WAR (Clemente was 6th, Koufax was on the pennant winner) so things evened out with Mays holding the short end of the stick because his team couldn’t beat the Dodgers when Koufax was healthy.

      Without Ellsworth, the Cubs would have been, what, eighth? So context does count both in MVP and Cy Young awards. If Trout had played all of 2012 and Cabrera came up later in April (missing, say, 2+ weeks), that’s the comparison. WAR is a tool, like all statistics.

    • Dinky says:

      Herb, you get a great big *huh* from me. According to bRef, Koufax’s WAR as a pitcher for 1963 was 10.7; Ellsworth’s was 10.2. Ellsworth had a better year at the plate (-0.2 versus -0.8) and both tied at 9.9 for total WAR behind Willie Mays. Ellsworth had a higher ERA+ from park adjustments, but Koufax pitched more, which is why his pitcher’s WAR is better. 20.1 more innings counts for WAR, but not for ERA+. So for Cy Young award, Koufax was clearly the winner.

      Ellsworth’s was a fluke season for a guy who never had an ERA+ above 108 in any other season, who was expected to be the #3 starter for the Cubs, who never excelled before or after. The Cubs were languishing in 7th place; the Dodgers were absolutely dependent on Koufax in winning the pennant and then the World Series, and would have won neither without Koufax. Koufax won the pitching triple crown in 1963 and pitched more innings than Ellsworth; he was already noted as an excellent pitcher from his prior two seasons. Ellsworth was the fluke. Willie Mays led the league in WAR, but the Dodgers won the pennant and Koufax was the best player on the best team. I would not have complained (much) had Mays won the MVP, but then Clemente jobbed Koufax in 1966 when Koufax actually led the NL in WAR (Clemente was 6th, Koufax was on the pennant winner) so things evened out with Mays holding the short end of the stick because his team couldn’t beat the Dodgers when Koufax was healthy.

      Without Ellsworth, the Cubs would have been, what, eighth? So context does count both in MVP and Cy Young awards. If Trout had played all of 2012 and Cabrera came up later in April (missing, say, 2+ weeks), that’s the comparison. WAR is a tool, like all statistics.

  14. I like stats, but baseball is not a spreadsheet. It’s a game, played by human beings. Human beings who respond to each other in emotional ways. So when you have a guy who’s doing freakishly amazing things as a hitter—teeing off against the best pitchers in baseball, driving in runs in all sorts of clutch situations—9th inning home runs to tie games, extra inning home runs to win games—he’s making the whole team better, he’s making them believe that anything is possible so long as they have the best hitter on the planet coming up to the plate. The Tigers line-up is hardly a murderers row, but they feel like it with the mighty Miggy in the middle. As fine a player as he is, Trout is hardly having the same effect on an absolutely brutal Angels squad. You want context—that’s context.

    • Mike says:

      “The Tigers line-up is hardly a murderers row”

      The Tigers are 1st in the majors with a 115 wRC+ (meaning they are 15% better than average as a team). Yeah, that would go down without Miggy, but it’s still a really good lineup.

      “Trout is hardly having the same effect on an absolutely brutal Angels squad”

      Yeah, boy, Trout really shouldn’t have signed Pujols and Hamilton. Or maybe he should have resigned Hunter. He really needs to get more great hitters on that team, since he is the GM and all.

    • schuyler101 says:

      So the context is Miggy has better teammates?

    • Either you believe in stories and narrative or you don’t. I do. I’ve seen ballplayers transform whole teams. Keith Hernandez came over from the Mets, and they immediately went from being good to bad. Kirk Gibson kicking ass with the Dodgers in 88. George Brett pulling a team that had no business winning to the World Series. This year, the Dodgers were dead in the water, playing the most lifeless ball you ever saw, until Yasiel Puig came in and galvanized them with his energy and hyper-aggressive style of play, and suddenly the whole team lost their fear of failure and caught fire, and now they’re unstoppable. Is it only one person? No, it’s never just one person, but one person can create an aura, a presence, a confidence that teammates feed off of. Joe D had it in a way that Ted Williams didn’t, but now all you see are the stats and figure the Red Sox were the ones winning championships all those years. When people talk about Derek Jeter being a winner, it’s the feeling the Yankees had of—it’s okay, we’ve got the coolest kid in class on our team, look at him get a big hit or make a big play. Of course, an old and broken down Jeter doesn’t have the same effect—it’s always about performance first and foremost—but there are those intangibles that transform clubhouses. Pete Rose had a ferocious competitiveness that the Phillies felt they lacked, and he helped lead them to their first championship in decades. You want to parse numbers and tell me that X is better than Y according to some formula? Fine, you do that. There’s some interesting stuff to be had in those numbers, but they don’t tell the whole story. Milton Bradley led the the majors in OPS+ one year yet the Rangers wisely got rid of him, and every team that saw those numbers and took a chance on that paranoid nutcase lived to regret it. When Bill Simmons says that all these advanced stats have drained the fun out of baseball because they don’t allow you to have an argument based on what you see and feel—so that a bloodless and uncaring JD Drew is stamped a great player because the stats say so—you have gone a long way towards making it pointless to watch the games in the first place. People don’t watch baseball anymore—the ratings prove it—they debate box scores. “I made baseball as fun as doing your taxes” the Simpsons said of Bill James (which was unfair, because Bill James was always a writer first and a statistician second, but it certainly applied to his offspring). So when I see Miguel Cabrera carrying the Tigers to a championship with a season for the ages, and you point to Mike Trout (who is having a fine season, don’t get me wrong) on a team that is under-performing from top to bottom, finding new and excruciating ways to lose night after night (some via Trout’s own gaffes in the outfield), and you’re telling me that Trout is having the season we should celebrate, this is the season we should immortalize for posterity, I am switching to some other sport. I am a baseball fan, not a WAR fan. Miguel Cabrera is the MVP up down and sideways, and it’s not even close.

    • Rick Rodstrom made this statement:

      “Keith Hernandez came over from the Mets, and they immediately went from being good to bad.”

      I was curious about this. Unfortunately, I MISREAD the statement the first time through. I thought Rick was claiming that the Mets went from bad to good when Hernandez joined them; I researched this and found that, indeed, the 1983 Mets were 22-36 (.380) when Hernandez entered the lineup for the first time on June 17th. From that day forward, the Mets were 46-58 (.442). Not exactly bad to good, but at least from very bad to not all that bad.

      But — that’s not what Rick actually claimed. When I reread it, I realized he was claiming that the Mets went from good to bad (immediately!) upon Hernandez’ departure. In Hernandez’ final year with the Mets (1989), they went 87-75. The next year, after Hernandez joined the Indians, the Mets IMPROVED to 91-71. It wasn’t until 1991 that the Mets slid to 77-84.

      So…Rick? Have I completely misunderstood your claim?

    • Paul Zummo says:

      Keith Hernandez came over from the Mets, and they immediately went from being good to bad.

      Mark already spoke to this, but this is an extreme claim (switching good and bad in the sentence above). You know who else joined the team at about the same time as Mex? First Darryl Strawberry, who debuted about a month before the trade for Keith, and then the next year Dwight Gooden debuted. Backt-o-back rookies of the year, not to mention this is about the time that Darling and Fernandez became mainstays of the organization. So while the Keith Hernandez trade was undoubtedly a turning point for the organization, you are missing, dare I say it, context.

    • Gene Claude says:

      I’m guessing those who “believe in the power of narrative over numbers” ceasing doing so when picking a brain surgeon.

    • djangoz says:

      Rick – numbers and narratives can very happily co-exist. The problem is when sportscasters and fans create narratives that the numbers don’t support, which happens all the time.

      “Kobe Bryant is a great clutch shooter” – a completely false narrative.

      “Ted Williams is one of the best hitters of all time” – a very true narrative.

      People like Bill James and others are just separating the wheat from the chaff when it comes to these narratives. But if someone grew up thinking Kobe Bryant was a great clutch shooter and bought his jersey and had his poster up on the wall, well, their emotional reaction is often very negative to hearing this. Cie la vie.

      Some people seek the truth, some clutch their lies close to their chest.

    • Dinky says:

      Uh, Puig helped. But overall, how much has he helped? Hanley Ramirez came off the DL and the Dodgers’ worst position, by FAR, became their best position. Greinke came off the DL and the #2 started went from Ryu to an ace. They traded for Nolasco who has been 6-1 with a 2.20 ERA since coming over; their #3 starter is also pitching like an ace. All those extra starting innings improved the bullpen, as did replacing League with Jansen. When Kemp comes back, it’s possible that Puig sits as much as anybody else (at least against RHP) because he’s insanely good when good, but also incredibly stupid when bad. Attributing the Dodgers surge to Puig is very short-sighted. They have the best pitching in baseball since Puig came up, but not the most runs.

      As for Kobe, Kobe takes *EVERY* tough shot. He is usually double teamed, sometimes triple teamed. He often tries letting somebody open take a shot, and the rest of the team has less success open than he does in double or triple coverage, at least since Fisher left. When Kobe had Shaq as an option, he had no triple coverage and less double coverage; teams needed to keep two guys on Shaq else the inevitable foul would often be a missed free throw to make a three point play in single coverage. The only Lakers I would prefer to have the ball in the clutch than Kobe, as long as I’ve been following the team, are West, Magic, and Kareem, and Magic often had good team mates to pass to: Kareem, Wilkes, Worthy, Scott. In this article, which I suspect is part of your opinion:

      there is a picture of Kobe taking a shot despite being quadruple teamed. What that article kind of glosses over is that the league as a whole only shoots 29.7% in those kind of game on the line situations; Kobe at 31.3 is better than that, and the best on the Lakers. What the article also doesn’t mention is that a big part of Kobe having only one assist in clutch situations in the time covered is that he passed the ball to guys who missed the shot (or cheated to the passing lane to Derek Fisher, cutting off his only decent second choice, in a triple team situation). Kobe is clearly better than average, best on the Lakers, so calling it completely false shows bias on your part. This more recent article:

      has lots of analysis. In the last league wide table, it ranks only the players who take the most shots in clutch situations, the clutchest of the clutch, regular season or playoffs, and Kobe doesn’t do badly here. In the tightest situations (last 24 seconds of the game or overtime, to tie or take the lead) Kobe’s 34.9% beats the average of the most common clutch options (31.4%) and improves to 37% versus 26.1% at shooting 3-pointers, third best in his sample. So he is better than the league average at guys having to take shots when the opposing team is completely focused on defense with no intention of getting a fast break off a miss, and much better at three pointers. Kobe is probably no longer the best in the league (and may never have been) but he has always been very good and taken and made (and missed) a lot of shots in clutch time.

      Remember, it’s better to get off a miss than it is to not get off a shot at all. I think Kobe’s the best the Lakers have ever had at getting off his own shot, because he can get the ball in the back court and take it himself. Kareem was the best at getting off a shot IF HE GOT A PASS. Magic was the best at dishing when a man was open. West was Mr. Clutch. Kobe is not nearly as weak as you say he is, and I cite statistics supporting it.

    • From 1977 to 1982, the Mets weren’t just bad, they were a joke, a moribund franchise playing to an empty stadium in the same town as the Yankees during a period when the Yanks won multiple World Series. The best season the Mets had was 1980 when they went 67 and 95. Ask other teams that have been epically bad for years how difficult it is to turn around a losing sports culture. Hernandez came aboard, and the Mets winning percentage improved by 62 points. Strawberry was the other notable addition to the team, but he started off very slowly, as rookies who are rushed to the bigs at a young age often do. Hernandez became something of his hitting coach, as well as a coach on the field to the pitchers and infielders. The attitude changed—they were becoming pros, learning how to win—and next year, with the addition of some good young pitchers, the became contenders, and the year after that they became a juggernaut and won it all. One guy can’t do everything, but the attitude changed when Hernandez joined the club.

      The most dramatic attitude change I can remember came with this years Dodgers. Given their skyrocketing payroll, the Dodgers came into the season with enormous expectations, and they played tight for the first two months. They were afraid to fail, and so they often did. Their hitting with runners in scoring position was abysmal. They were dead in the water, in last place, 9 games out in June, with Mattingly seemingly a loss away from being fired. Then came Puig. Puig added not only his extraordinary physical gifts to the club, and a level of energy that had gone AWOL, he also added a level of fearlessness. If there was one thing Yasiel Puig was not afraid to do, it was fail. He was the most aggressive player I have ever seen. He took insane risks that forced the opposition’s hand. His game was one of permanent attack, and it freed the rest of the Dodgers into taking risks themselves, going for the extra base, always pushing, always fighting back. It led to the greatest streak of winning baseball in the last 70 years. Sure Puig had help. Hanley Ramirez returning to the lineup was key—but how much of Ramirez superb play was in response to the challenge thrown down by Puig?—and when Ramirez went down again the Dodgers didn’t miss a beat. The pitching, too, has been lights out during their great run, but I have to wonder how much of the Puig effect has rubbed off on the pitchers, filling them with confidence and energy and goading them into attack mode? Yasiel Puig is like a living, breathing steroid, raising everybody’s testosterone levels. Whether he burns out or breaks down is an open question, but he has certainly changed the Dodgers season.

      WATCH THE GAMES PEOPLE, that’s my message. The numbers are just a byproduct to what is happening on the field.

    • simon says:

      Rick, Yasiel Puig happened to arrive in the Dodgers’ lineup at almost exactly the same time as Hanley Ramirez returned from injury. Ramirez has hit .343/.392/.628 and put up 4.3 WAR in 68 games, essentially a 10 WAR pace. It’s *possible* that the Dodgers resurgence doesn’t have everything to do with Puig.

      Not that Puig hasn’t been important, of course.

    • Ian R. says:

      “,,,I have to wonder how much of the Puig effect has rubbed off on the pitchers, filling them with confidence and energy and goading them into attack mode? Yasiel Puig is like a living, breathing steroid, raising everybody’s testosterone levels. Whether he burns out or breaks down is an open question, but he has certainly changed the Dodgers season.”

      You know, Rick, I have to hand it to you for keeping it pretty believable for so long. This, though, is far too over-the-top to be a serious claim.

      It’s possible that Puig’s presence has energized the Dodgers. It’s also possible, and far more likely, that they were an incredibly talented team to begin with and finally started playing up to their talent.

    • Which Hunt says:

      Ryu has a 3.1 WAR this year according to Baseball Reference. Grienke averages a 3.1 WAR over his career. This year Grienke has a 3.6 WAR and 1.1 of that is his HITTING! It’s not so much that the Dodgers are replacing a #2 pitcher for an ace, its that they have Kershaw (7.1 pitching!!!!) followed by two ace caliber pitchers, rather than Kershaw then Ryu and whoever they were cycling up from the Albuquerque Isotopes (replacement level). Why all the hate for Ryu who is having an excellent season even by Grienke standards?
      Also Greinke batting is must watch television.

    • Which Hunt says:

      This comment has been removed by the author.

    • Before Puig arrived, Zack Greinke had a 4.80 ERA. Since Puig arrived, it’s been 2.27. Perhaps it’s coincidence, perhaps not.

  15. J Hench says:

    One thing that may make the whole context part easier to understand (at least it does for me) is that it is not saying that Cabrera (or Koufax in ’63, or whoever) did not actually create all those runs. It’s not taking anything away from what Cabrera did.

    Instead, the context is that, in the run environment in which Cabrera played, the 145 runs or so he created has a certain amount of value. And in the run environment in which Trout played, his 130 runs actually has more value.

    If the average team is scoring 5 runs a game over 100 games, and Player A creates 7 runs/game, then he’s a valuable player. But if Player B creates 6.5 r/g in an environment where the average team is scoring only 4 runs/game, then the runs Player B creates end up being more valuable to his team winning games, because it is harder to score runs in that environment. No one is saying that 7 runs is lower than 6.5. Just that in a 4 run/game environment, each run is more valuable than in a 5 run/game environment.

    • djangoz says:

      Well said.

      I made the point below that if the Cowboys played on a field that was only 90 yards long things would be different. They would score more touchdowns, more field goals and give up more of each too. I think most sports fans can understand that pretty easily and would also think it was somewhat “unfair” if Tony Romo has 10% more TD passes than someone else and was named MVP. “He plays on a short field!” would be screamed by every non-Cowboy fan.

      The wild variances in baseball parks is just something that has been around so long that most fans never stopped to think about it and it stretches their brain to do so now. Plus ti would possibly make them reevaluate some other memories they fondly cherish. So they close their mind to it and dismiss it as that propeller-head nonsense.

    • Dinky says:

      I agree that Cabrera did not create all those runs. Having Prince Fielder behind him means he must be pitched to far more often.

      Koufax, on the other hand, was the best pitcher in baseball from 1962-1966. His career OPS at home was .573; away was .616. Both are superb. Perhaps more telling, his career as a starter was .590, as a reliever .668, suggesting either that Koufax was mismanaged (which has been suggested) before the Dodgers moved to Dodger Stadium. Since his homer rate was better as a reliever (mostly in Ebbets Field, some in the Coliseum) than as a starter, it suggests that mismanagement was the factor. He only had five games in relief in the years he was in Dodger Stadium, 78 before then. Since his motion was notoriously complex and he known to throw up to 100 warm up pitches to prepare, it seems probable that had he just been used as a starter all the time, he’d have had even better career stats. Yes, it helped Koufax to have a lot of speed behind him in the peak years, but broadly speaking, it didn’t help Koufax to have other good pitchers on his staff. In 1963-1965 his road ERA was under 3; in 1966 it was under 2, so crediting all his success to Dodger Stadium would be shortsighted. Also, for such a complex windup, greater familiarity with the mound would lead to a bigger swing than for a pitcher with a simpler release.

    • J Hench says:

      Re: Koufax, and anyone else for that matter, the point is not (and forgive me, because I did not make it too clearly the first time)…. The point is not “Koufax’s dominance is an illusion caused by pitching in Dodger Stadium – put Mudcat Grant there and he’d be just as good” (just using grant as an example, feel free to substitute any other name there). The point is that Koufax’s dominance, due to the fact that fewer runs were scored overall in Dodger Stadium, is not worth as many wins as it would have been if he had put up the exact same numbers in Fenway at the time, or in Wrigley, or in the 1990s.

      That doesn’t mean I’m crediting the success to Dodger Stadium, or to the 1960s (and of course, Koufax didn’t benefit from really low offensive numbers of the late ’60s). It means that, given the context of the park, the numbers that Koufax posted did not contribute to as many wins as they would have had they taken place in an environment where runs were less scarce.

      And let’s not kid ourselves. WAR and other performance metrics love Koufax. He had two seasons with over 10 bWAR. He was great.

      Also, this isn’t to say that parks do not influence performance. It’s just to differentiate that concept -“the Dodgers moved to Chavez Ravine and suddenly Koufax was a HOFer” from how park factor is used in WAR, as I understand it.

  16. Ozsportsdude says:

    Joe does Hitters or Pitchers park take into account the players on the team

    Seems a coincidence that Detroit goes from being a Pitchers to a Hitters Park around the same time Detroit gets Prince Fielder and has an awesome hitting team.

    Just for ‘context’ I am an Australian with very little knowledge of Baseball, I just love Joes writing so read even when a lot of the stuff doesnt mean much to me. But I am also an Analyst in my job, so numbers interest me and am just interested whether the Hitters or Pitchers part thinks factors in the players who play in them

    • Dinky says:

      It takes into account the players on the team as a whole compared to league wide adjustments as a whole. Thus, hitters everywhere (on average) hit better at home, but hitters in good hitting parks tend to gain more compared to the average home/road splits than hitters in bad hitting parks. But it does not account for the differential between being in a good offense (Cabrera) or bad one (Trout).

  17. Noah Hunter says:

    Just a brilliantly, simple, well-directed and effective article. There is a reason you are revered.

  18. Vidor says:

    This year it feels like Posnanski really wants Trout to win the MVP and is straining for a reason to justify it.

  19. Alejo says:

    Dear Blogger,

    This “explanation to the layman” is something I needed and I reckon the math makes sense and is indeed sensible but, for me, Cabrera still has an edge (MVP voting-wise) because of baseball tradition: home runs “feel” (I know, it’s absurd, to have this kind of feeling nowadays) more valuable than stolen bases and good defence by a left fielder playing for a team completely out of contention. I mean, we are talking about a core infielder that has been demolishing the league for two years in a row and is playing for a division leader (then again, context).

    One thing: you didn’t take into account Cabrera’s splits:

    Home .366 .455 .648 with 83 hits, 16 Hr and 57 RBIs

    Away .349 .443 .719 with 87 hits, 27 Hr and 73 RBIs

    I don’t really see how context, as referred to home field, is a factor in this conversation. The guy is breaking bones all over the country.

    • Jovins says:

      He also plays in some really good pitchers parks in his division. It’s why some of the Colorado guys have such extreme splits – not just because they play in a great park for hitters, but because in their division they play in LA, San Diego, and San Francisco.

    • Alejo says:

      So now it turns out that Cabrera only plays in hitters parks, then.

      That is why he is so good.

    • He’s so good because he is a fantastic hitter. If I am understanding the advanced metrics, if Cabrera was playing for the Angels his offensive numbers would still be better than Trout’s, though not by as much. Perhaps the only way to verify this for sure is to have him play for the Angels. Which, as an Angels fan, I heartily endorse. Though it seems more likely that they’ll sign him to a five-year deal when he’s 38.

    • Alejo says:

      I was being ironic.

    • djangoz says:

      I can understand your feelings about the tradition. But I’d suggest that those feelings will not be nearly as common 20-30 years from now. Just as Pitcher wins will not be as important or batting average.

      It’ll take time, but we’re moving toward better metrics. When kids grow up thinking about OBP and reading Moneyball it will affect how they view the game as adults.

    • Dinky says:

      Cabrera is a great hitter. He’d be a great hitter in Petco or anywhere else. But it’s easier to see his greatness in a good hitting environment. Look at a couple of guys whose careers were suppressed by the Astrodome: Jose Cruz and Cesar Cedeno. Cruz had a career OPS+ of 120, which is to me the break point at which All-Stars get chosen and five full-time seasons above 130, yet he was only chosen to the All-Star team twice. Cedeno’s 123 is even better, with five Gold Gloves and 550 steals in his career. But neither reached even .800 career OPS, and both had stats badly suppressed in Houston.

  20. My overall problem with WAR is that, from what I have read and seen, I do not quite trust the defensive stats. I DO trust that Tom Tango and Bill James and others have figured out the difference between, say, a double and a single, precisely. But I have much, much less confidence that fielding “value” has been so accurately determined, and how much it is “worth” compared to a player’s offense.

    • I would be very pleased if standard error numbers were included, especially since WAR is being being listed down to the tenth! 5.1 WAR is essentially the same as 5.0 WAR due to the error in estimation.

      I would then presume that those players who have a large chunk of their WAR coming from their fielding would have larger standard error numbers. So it might be 5.0 WAR +/- 0.5 WAR for a great hitter while it’s 5.0 WAR +/- 1.5 WAR for a great fielder. Though now I’m troubled by the 6.5 WAR “peak” for the latter, even if the low point is 3.5 WAR.

      But the point is please provide a range of what their WAR likely is and don’t be so presumptuous to think you calculate this number down to a tenth.

    • I agree. I just cannot get beyond the seemingly arbitrary defensive measurements and I am not at all sure the proper balance between offense, defense and base running has been established. Park Factor makes sense on some level, but I don’t think it is as big of a factor as some do. The bottom line is WAR doesn’t do it for me – I’m not opposed to advanced metrics, but I’m waiting for a better formula.

    • invitro says:

      I love WAR and am trying to learn and understand how to use it and its components. Last year, my confidence in dWAR was shaken a little bit when some players had historically high dWARs because of the increasing prominence in defensive shifts. Then sometime in summer, the dWAR formula was changed, which resulted in a change of… I don’t remember, maybe two in dWAR, which is immense.

      That made me feel like dWAR, while being the best available measure of defensive value, is still going through major growing pains, and major changes are still likely to occur. Is this a reasonable viewpoint? Is it a reasonable reason to be wary of dWAR?

    • Chad says:

      Yes, yes, yes, and yes.

    • Ian R. says:

      Absolutely yes. The biggest issue with WAR as it stands is that it uses single-season defensive stats, and defensive stats aren’t accurate in a one-season sample.

      Having said that, if we drop the defensive component from WAR, it looks like Trout and Cabrera are at identical 8.6 marks. Identical.

      I don’t think it’s too much of a stretch to say that Trout is a better defensive player than Cabrera, even if the margin is smaller than 7 runs (if anything, I’d say it’s probably bigger than that). So when WAR puts Trout ahead of Cabrera, I’m inclined to believe it.

  21. Dave Gabel says:

    On the subject of Context: Doesn’t the quality of a guy’s team play an even bigger role than the size/conditions of his ballpark? Trout is flanked by Erick Aybar (.310 OBP) and a badly under-performing Josh Hamilton in a mediocre Angels lineup. Cabrera is flanked by Torii Hunter (.343 OBP) and a somewhat under-performing Prince Fielder in what I assume is a better overall Tigers lineup. Not suggesting Trout would be a lot closer to Cabrera in the hitting categories if they switched teams, but maybe a little bit??

    • Robert says:

      In some of the traditional stats, you’d obviously be correct. A better supporting cast would certainly help Trout’s runs scored and RBIs. But for the more advanced stats like WAR, those lineup differences are accounted for mathematically…as explained by lxcila down the page.

  22. Paul Zummo says:

    This is a great explanation of what goes into WAR, and why the numbers appear the way they do. One thing I would say is that perhaps the contextual elements of advanced stats – the ballpark adjustments that go into WAR, as well as FIP more generally – make them more useful as tools to evaluate players overall than they are as concrete barometers for determining post-season awards.

    Let me try to explain a little further. Let’s say a pitcher has a FIP of about 2.2 and leads the league, but his traditional stats are a bit worse – 3.00 ERA, .500 winning percentage. Meanwhile, another pitcher has the reverse situation and his ERA is much lower than his FIP. What this tells us is that if you had to take either pitcher going forward, you’d go with Pitcher A because the contextual stats tell us that he ran into some bad luck, had a bad defense, etc. However, FIP is really just telling us, in essence, that Pitcher A should have better numbers than Pitcher B if all things were equal. But in terms of awards we can only judge what did happen. So I think FIP (and FIP is, I believe, included in pitcher WAR) won’t be as useful in determining who should win a Cy Young Award than in measuring who the better pitcher is.

    In the case of Trout and Cabrera, what the park factor and other contextual adjustments tell us is that Trout is having a mildly better season in a vacuum than Cabrera. But the players don’t play in a vacuum, and so it’s not really fair to punish Cabrera for actual performance that may have been aided by the parks he played in.

    This has all been a long-winded way of saying that WAR, as great a statistic as it is, shouldn’t be the sole criteria for choosing MVPs, especially when the two players are close.

  23. invitro says:

    Some prominent baseball writer, maybe Bill James, wrote somewhere that the Hall of Fame (and maybe other major awards like MVP, Cy, etc.) should require its voters to pass a rigorous test. The test would cover all baseball history, all rules of baseball, and all of the common sabermetric methods. If you couldn’t pass this test, you couldn’t vote. If you couldn’t explain WAR, you couldn’t vote.

    I wonder what you folks and Joe think about this idea (if I remember it correctly).

    • Rob Smith says:

      If there were a test like this, we’d end up with four qualified HOF voters.

    • invitro says:

      I didn’t remember it correctly… James recommends a baseball scholars group be -part- of an extended group of HoF voters (that includes fans, players, professionals, media), and he doesn’t mention “advanced statistics” at all as part of the baseball knowledge test. So never mind.

  24. Chad says:

    You are absolutely correct. I came away certain that WAR is just wrong and Cabrera absolutely is having the better season.

    • djangoz says:

      Please don’t reproduce. That’s what it really comes down to.

      We need to weed this anti-math and science way of thinking out of the gene pool. I wish you a great life…I just hope you don’t create more of you.

    • Chad says:

      Classy. I’m sorry that because I don’t buy into a flawed system used to measure the value of a baseball player as to who is better that you think I shouldn’t reproduce. Just because someone comes up with a number that says Player A is better than Player B and I don’t buy into it I’m anti-math??

      I’m all about baseball stats. I’m not about people who insist that WAR is the only measure of who is better. It’s flawed.

    • clashfan says:

      Sure it’s flawed. But other stats are *more* flawed. Using WAR as the one and only stat may be misguided; others should be used as well, to gain a better picture.

      But it’s a whole lot better than RBI.

  25. matt david says:

    I don’t buy this: “Mike Trout has 9 more doubles and 7 more triples than Cabrera. That is worth roughly 14 runs to a typical team, if those events occurred in typical fashion.”

    Does this refer to an RBI or a run scored? If it is an RBI, RBI is a team dependent stat which Joe has gone over and over about how little we should put into it. If it is run scored, we already have a stat for that.

    • What do you mean “you don’t buy” it? Read the sentence you quoted again:

      “That is worth roughly 14 runs to a typical team, if those events occurred in typical fashion.””

      It doesn’t refer to RBI or runs scored. It refer to exactly what was just stated — the number of runs that those hits would typically produce. Obviously some doubles are hit with no one on base, and the runner is stranded, and no runs are actually recorded. On the other hand, other doubles are hit with the bases full and maybe 2 or 3 runs are recorded. So this is an aggregate of all those possible situations. The reason for looking at it from this angle is that the hitter doesn’t control the circumstances in which he hits — he shouldn’t be penalized (or rewarded) for the success of his teammates in getting on base in front of him (or driving him in once he’s on base).

    • Ixcila says:

      It’s neither. It refers to the “run value” of each of those individual events. It’s derived by looking at all the events that can happen when a batter steps to the plate, and calculating – using something called linear weights – how much each event contributes to his team’s runs.

      For example, a home run is obviously the best outcome a hitter can come away with from a plate appearance, and it’s a relatively simple example, so let’s start there. How many runs is a home run worth? The old answer would be “it depends,” because it varies with how many people are standing on base. Solo shot? One run. Grand slam? Four runs. But that answer obscures the fact that the batter has no control over how many guys are on base when he comes up. None. Zero. So if one batter hits three grand slams, and the other hits three solo shots, we can’t say grand slam guy is better, just that he was luckier.

      So we strip away context. How? Calculate how often, in the entire league, batters step in with the bases empty, with one on, two on, or the bases loaded. Those situations aren’t equally probable – bases empty is a much more likely situation than bases loaded – which is where LINEAR WEIGHTS comes in. We weight each situation by multiplying its run value (1 run for bases empty, 4 runs for bases loaded) by it’s probability (as a fraction of 1), and out pops a number: about 1.44. That’s the average value a home run contributes.

      So when a hitter smacks a dinger, for purposes of calculating WAR, we don’t pay attention to how many guys were on base. The batter doesn’t control that. Instead, we credit him with 1.44 runs. And we do similar math to calculate the average value of a single, a double, a triple, a walk, a strikeout, a GIDP, and every other possible outcome. Add all those up, and you get a stat called wRC (weighted runs created), which shouldn’t be confused with the traditional stat “runs created”, obtained by summing runs and RBI.

      Normally, you just see all these numbers lumped together as wRC or wRAA (same stat, just using league average as the baseline), but Joe’s splitting them out piece-by-piece, so we can see where the value comes from.

    • matt david says:

      Thanks for the reply…however, why do we count a double play against a player then? Isn’t that all about context?

    • Andrew says:

      I’m not an expert, so someone please correct me if I’m wrong. I believe the “double play” component is included so we can add value to batters who are better at avoiding a double play, all else being equal. Let’s say two hitters both come up to bat with a man on first ten times, and all ten times hit a ground ball to an infielder. If hitter can beat out the double play 3 out of 10 times, but hitter B gets thrown out all 10 times, then hitter A made 3 less outs than hitter B, again all else being equal. The trick is to figure out the “all else being equal” part. So I don’t think WAR counts up the number of double plays grounded into and detracts this from the hitter, because you’re right, that would be context dependent. Instead, I believe it looks at the ratio of GIDP to double play opportunities, and from that ascertains whether one hitter is better at avoiding a double play in a context-free environment…ie “all else being equal.”

      That’s how I imagine it works, anyway, but if someone knows better please correct me

    • I can’t find anything on fangraphs explaining how/if GIDP are calculated into fWAR, but B-Ref absolutely does incorporate GIDP rates (as opposed to absolute values) into bWAR.

      Therefore, it’s not that Miggy is penalized because he hits into more DP’s than Mike Trout — it’s that Miggy is penalized because he hits into more DP’s than expected compared to an average player in those same circumstances, while Mike Trout hits into fewer DP’s than expected compared to that same average player in his circumstances.

    • Andrew says:

      I think I was trying to say the same thing as Mark, only he said it much more succinctly and clearly.

    • Dinky says:

      Mark has it right. It’s all runs expected versus outs used up. A double play eats up two outs to no benefit; players who create fewer outs (either by greater speed cutting down on DPs, or more fly balls, or better situational hitting) provide more wins to their team, all other things being equal.

  26. Aaron Ross says:

    A wonderful post Joe, as always, but there is something wrong with the logic of the context part of the argument. Context can tell us what a player would do if everything else was equal. In other words, if Trout and Cabrera played the same number of games in the same parks (e.g. if they were on the same team), then Trout would be more valuable. This is a fine way to determine “fairness” or to project how valuable a player is if he is going to move teams and therefore switch context.

    However, something such as MVP is based not on what a player does given his circumstances or what he might do if the circumstances were different. It is based, as it should be, on what he actually does. That being the case, the context piece of this seems somewhat irrelevant.

    • Gene Claude says:

      The point is that the context (park) determines whether what he does (hit the ball 398.1 feet, for instance) is good (a home run) or bad (an out). He “did” the same thing in either event, it is something entirely outside of his control that determines its value. If MVP means “most valuable player,” we are trying to isolate the value attributable to the player’s actions.

      One thing about park effects that makes my head hurt…presumably some players are hurt more by certain parks. Couldn’t a park be a pitchers park, but not affect some players as much? I assume Miggy’s average home run distance is farther than others’. Wouldn’t that mean that he is affected less by his park than others? Or does that get factored in by the process?

    • djangoz says:

      You’re actually trying to make a case that the park a baseball player plays in is not a factor in their performance?

      I give up.

    • Mark Daniel says:

      djangoz, I seem to think he said that park is a factor in a player’s performance, but in the case of the MVP award it’s irrelevant.

      I think this idea stems from the various ways you can interpret park factor adjustments. In one interpretation, park factor means what a player actually did in a pitcher’s park is more valuable than comparable numbers put up by a player in a hitter’s park, because the pitcher’s park made runs harder to come by.
      In the other interpretation, park factor could mean that a 400 foot out at a pitcher’s park is better than a 375 foot home run at a hitter’s park.
      One, I think, is a misinterpretation, but I at least can see how that happens.

    • Dinky says:

      Djangoz, the park matters. But for MVP awards, he feels the team context is as important. The Tigers are in the playoffs, the Angels are not, so Trout needs to be a whole lot better (Ernie Banks better) to deserve the MVP for the weaker team. I think pennant races matter less than most, but when it’s at all close I think it’s a valid tie breaker.

  27. djangoz says:

    If the Dallas Cowboys played on a 90 yard field and the Redskins played on a 110 yard field; the Celtics played with 10.5 foot hoops and the Heat with 9.5 foot hoops…

    Of course the very different nature of ballparks in baseball makes a difference!

    Joe, you’re far more tolerant than I am. Any “friend” who said “eh, too much math for me” wouldn’t be a close friend for very long.

    I can understand fans wanting to enjoy a sport without thinking too much, that it is entertainment. But if you’re going to argue who should win the MVP and someone makes a compelling argument, that’s pretty easy to understand, and you dismiss it that easily – well, I don’t have much respect for that person.

    • Chad says:

      That’s because you believe that WAR is perfect. Others do not.

      I prefer to use OPS+, which makes adjustments for parks, as a basis of hitting value. Cabrera is clearly a better hitter than Trout using this method. Trout has more value as a baserunner and defender. I just happen to be one of those people that believes that Cabrera’s advantages as a hitter outweigh Trout’s advantages.

    • Ixcila says:

      First off, there’s no reason to use OPS+ instead of wRC+, because they’re both park-corrected, and wRC+ weights outcomes by actual run value, as opposed to OPS+, which overrates slugging over OPB, and suffers from mismatched-denominator-syndrome.

      So, looking at wRC+, Cabrera has a 204, compared to Trout’s 178. So yeah, Cabrera’s a better hitter. Fangraphs credits him with 15 extra batting runs, so he’s somewhere in the ballpark of 15 runs better, after park adjustment. However, by Fangraphs’ numbers, Trout has a nine-run edge in baserunning, and a seventeen-run edge in fielding, with Cabrera making up three runs on positional adjustment. Result? Trout has an fWAR of 8.8, against Cabrera’s 7.6.

      So basically, Fangraphs (or B-Ref, which are the numbers Joe used, which like Trout a little less and Cabrera a little more) supports your statement that Cabrera is a better hitter, but they offer actual quantitative evidence that Trout’s other attributes overwhelm that gap. You’ve offered your gut.

    • Chad says:

      I don’t believe the quantitative “evidence”. Pretty simple really. I (and many others) believe it’s flawed. I think it’s one measure to use, not the end-all be-all that so many want to make it.

    • Okay, but why? I hesitate to say out of fear of appearing to speak on your behalf (and please correct me if this does not apply to your position), but often the most vehement anti-WAR voices offer nothing more than “it’s flawed” or “I don’t trust it” as an argument.

      Fine, it’s flawed. We know. It’s a work in progress — but the point is exactly that — we are progressing toward the goal of accurately measuring a player’s value to his team. If you are going to say that you don’t trust WAR, then offer an alternative and EXPLAIN why.

    • I just happen to be one of those people that believes that Cabrera’s advantages as a hitter outweigh Trout’s advantages.

      Why? Because that’s what you want to think? Because that’s what your subjective gut says? Very rarely does “feeling” about something win out over actually crunching the damn numbers.

    • Rob Smith says:

      People who go with their feelings often have tough lives. Feelings play tricks on us. Using our brain and evidence is a much more solid way to go. As for metrics/stats, it’s an age old problem to obtain good reliable numbers. Often, you realize you can’t get perfect numbers, but if you get something reasonably good, then at minimum it can be used for comparisons. 7.6 WAR doesn’t mean much of anything by itself. But when compared to a 5.4 WAR it means something. I have an issue with career WAR however, since it’s cumulative, rather than an average per season. But it still does tell you something. Plus nobody said you can’t also use traditional stats. Those voting on awards and HOF these days largely look at both.

    • Chad says:

      Cabrera, with a strong finish and a little luck, and some health, could end up leading not just the American League, but all of baseball in the following stats:

      Batting average (leads .359 to .333)
      On base % (leads .450 to .435)
      Slugging (leads .683 to .679)
      OPS (leads 1.132 to 1.062)
      Runs (Tied for 2nd, trails 100-95)
      Hits (leads 173 to 169)
      Total bases (329 to 323)
      Home runs (2nd, trails 47-43)
      RBI (leads 130-121)
      OPS+ (leads 201-181)
      Runs created (leads 147-131)

      He won’t lead in walks or extra base hits, but he is in the top 5 for both.

      Again, that’s for all of MLB. So sue me if I think Cabrera is a better hitter. I’m not disputing Trout’s greatness, but I won’t accept that he is better because “WAR says so”.

    • Ian R. says:

      Nobody is arguing that Trout is a better hitter than Cabrera. The hitting components of WAR agree with you. What some people are saying is that the difference between Trout and Cabrera as hitters is smaller than the combined difference between Trout and Cabrera in all other aspects of the game.

  28. Brian says:

    Wait,why don’t intentional walks count?

  29. Jon W says:

    The response of your friend sounds a lot like the people who are against instant replay and appreciate how the ‘human element’ of the umpires affects games.

  30. WAR is a joke. Also Comerica, have you ever watched a game that’s played there? This week Feilders hit a couple of 410 ft fly outs. That’s not a hitters park.

    • Herb Smith says:

      Do you realize that if a ballpark has really deep fences ( which takes away homers) that is thus means there is a LOT more outfield grass (for hits to fall in). Thus, a park could be a bad home run park, but a great hitter’s park.

      PERFECT example: Fenway Park for left-handers. It’s really hard to pull a ball over the right-field wall. But all that open space means it’s a GREAT park for having high batting averages.

  31. I don’t have an opinion on MVP as between Cabrera & Trout, but I do have a question about the use of park effect. For offense, the wholly reasonable use of park effect is to assign relatively greater value to runs scored in a run-scare environment, and relatively less value to runs scored in a run-rich environment. But for defense, a run saved may have relatively less value in a run-rich environment, and greater value in a run-scarce environment. (Actually, the same might hold for negative offensive attributes such as GIDP.) The extreme case would be a direct relationship, i.e., since your team scores 5% more runs in this park a run saved is worth 5% less there–that would understate the contributions of good defenders in offensive parks. Still, as a first thought if there will be a lot of HRs and very few balls caught in foul ground, there will be more chances for your team to make up for the results of your lesser range, etc., on defense.

    Maybe one issue with park effects is the feeling that once a hitter has shown he can outplay a tough park, there is less interest in “reducing” his numbers when he moves to a favorable one. (Once you were Adrian Gonzalez in 2010, you never needed any help from Fenway Park.) I don’t know how I feel about that, either.

  32. I had to postpone trying to read all the contributions to this discussion, to protect my eyes and fading powers of concentration.

    Thanks to Joe for starting such a stimulating thread.

  33. This comment has been removed by the author.

  34. H Romero says:

    WoW! Many important info i found from this blog post. Thanks a lot from Mobile Marketing

Leave a Reply

Your email address will not be published. Required fields are marked *