Identifying First Half MVP Candidates
With yet another day to go before actual baseball returns to the field, I thought I would take a quick look at some of the potential MVP candidates in both leagues based on the first half of the season.
Identifying MVP candidates is certainly not a straightforward process, nor is the criteria universally agreed upon. Knowing this I will not begin or end this article with any claim to have identified the “proper” candidates. These are my candidates based on my way of looking at the term “valuable”.
So what is my criteria? Well, I like to think of MVPs as players that provide an exceptional amount of production in both an absolute and relative sense. This means identifying players that lead or are close to leading the league in production, but where there is also a sizable gap between their production and that of the second best player on their own team. This means that I do tend to discount great performances by players that happen to share the same uniform as equally great players. Is it their fault? Absolutely not. In fact, those players could likely be the best all around players in the entire league. But when it comes to value I think there is a relative component that should be considered. This isn’t to necessarily give credit to the player (i.e. they don’t “step it up” to make up for the gap in talent on the team), but rather to the performance itself.
Like I said, this is my criteria and I don’t claim that it should trump all others, nor would I say it is complete on it’s own. Rather, I think it’ a useful starting place.
Okay, enough with the preamble. Let’s get to the data.
I used both WAR and wRC+ to determine those players with the greatest absolute and relative value so far. WAR is a useful heuristic when one is trying to capture overall value (hitting, baserunning, defense) in a single number. However, while all of the components of WAR can vary across sample sizes (including 80+ games), we know that UZR is even more sensitive to small samples. To counter this a bit I also included the absolute and relative values of each player’s weighted runs created (wRC+). This is not to discount defense, simply to acknowledge that an offensive performance (or, at least, the offensive metric I am using) over roughly 80 games is closer to reflecting a player’s actual talent level than their defensive performance.
Here are the National League candidates:
| NL Player | WARDiff | WAR | wRC+Diff | wRC+ |
|---|---|---|---|---|
| David Wright | 3.6 | 4.9 | 48 | 169 |
| Joey Votto | 2.8 | 5.0 | 49 | 186 |
| Carlos Ruiz | 2.2 | 4.3 | 43 | 168 |
| Andrew McCutchen | 2.1 | 4.5 | 55 | 180 |
| Ryan Braun | 1.9 | 4.3 | 43 | 166 |
| Michael Bourn | 0.7 | 4.5 | -7 | 142 |
And here are the players in the American League:
| AL Player | WARDiff | WAR | wRC+Diff | wRC+ |
|---|---|---|---|---|
| Robinson Cano | 2.5 | 4.3 | 24 | 152 |
| Mike Trout | 2.0 | 4.8 | 12 | 172 |
| Austin Jackson | 0.8 | 4.0 | 10 | 160 |
In terms of absolute value I decided to include a cut-off of 4 WAR. It’s somewhat arbitrary and crude, based on my conception that a full season MVP will likely end the season with a WAR around +/-8. (Yes, I know that this isn’t always actually the case, but as I said it’s based on my view of MVP.)
What’s interesting is that the National League appears to have more candidates for first half MVP than the American League. This isn’t to say that the National League is therefore better, just that the league happens to have more 4+ WAR players over the course of the first half. Additionally, it illustrates that whether you are looking at WAR or wRC+, rosters in the National League are less balanced than those of the American League.
The player with an absolute value over 4 WAR and the highest difference between their performance and their closest teammate in the National League is David Wright. Wright is having an unreal season so far, and whether you look at WAR or simply his offensive performance (wRC+) he clearly has been the linchpin for a Mets team that currently sits a half game out of the Wild Card. Wright is essentially tied with the Reds’ Joey Votto in terms of WAR, but has accounted for 3.5 wins more than any other player on the Mets. Votto has been the superior offensive player this year, but the gap between both players and their teammates is virtually identical. I should also note that the Pirates’ Andrew McCutchen, who has been nothing short of spectacular, has accounted for a greater differential offensively than either Wright or Votto. Now you might say that’s not a hard trick given the offense in Pittsburgh, but McCutchen is sporting the second-highest wRC+ (180) in all of baseball.
In terms of the American League, my criteria yields three candidates. The Yankees’ Robinson Cano has managed to stand out on a roster that does not lack for star power. This year, Cano has accounted for over 2.5 wins more than his closest teammate (Mark Teixeira) and has been 23% better offensively than Curtis Granderson. As good as Cano has been, Angels’ rookie Mike Trout has arguably been more impressive. Trout has amassed almost 5 WAR since he was called up in late April — over half a win better than Cano. Additionally, Trout has been an offensive force, posting aan AL-leading wRC+ of 172.
So who would I actually pick as the first half MVP from each league?
It’s really tough in the National League. My inclination and admitted bias as a fan leads me to say Wright, but given what McCutchen has been doing to keep that Pittsburgh team in contention it is very hard to argue against him. In the American League, it’s much easier for me. Combine Trout’s offensive contribution with his phenomenal defense and all-around stellar game and it’s hard for me to argue against the 20 year old being the first half MVP in the American League.
I think you had a crucial flaw in your end result in that you forgot to include Delmon Young for consideration.
Me agrees with Trout, but Joey Vottomatic takes my vote in the NL. I also tend to agree with discounting guys surrounded by stars. It hurts a team a lot less when one of these guys gets injured. That said, Cincy would be down around Milwaukee’s neighborhood without Votto.
How dumb do you feel now?
This is clearly an article that belongs on NotGraphs – I mean really, you give the readers all of these ‘facts’ and ‘analytical observations’ and ‘contextualized statistics’ and you STILL don’t end up with Michael Young as AL MVP? He’s only the most grit-soaked player on a first place team. D’uh…
And, he’s classy.
“Have some class, don’t be an ass.”
MY
Would the writers give Trout ROY and MVP?
Ichiro got it.
They gave Verlander the MVP, when he didn’t even have the best ERA or FIP in baseball…
You’re forgetting the difference between the AL and NL average ERA. Verlander did have the best ERA- in baseball!
The writers gave Verlander the MVP when he didn’t have the best FIP in the majors? O RLY??? Do you think the writers even know what FIP stands for?
No way, never happens.
But it HAS happened, with Ichiro.
Just as importantly, it’s not as though it *should* happen all the time but writers just don’t like doing it; Trout is a legit candidate this year and I think he will be treated as such should he put up as monstrous a 2nd half as his first.
Burleigh: It happened with someone else too… hint hint
burleigh, that sound you’re hearing is the joke going over your head
MVP is about performance, not talent level, and thus the defensive numbers should count equally. While they are less predictive of future performance, they still accurately represent what has actually occurred.
This is not true. There is massive measurement error in the defensive numbers. They do not accurately reflect what happened. The defensive numbers are much different than the offensive numbers. We never miscount home runs. We have no idea if any single defensive play credited as being difficult is actually difficult or routine.
You don’t know if a homerun was a laser over the CF wall or a pop up down the line (well, we might, but no one cares). There is uncertainty in everything. The stats we have to differentiate the actual performance of defensive players are pretty good and the offensive ones aren’t perfect either.
Are we discounting for BABIP?
The fact is defense is valuable and mistakes in valuation can go one way or another (just like with hitting), but to ignore it or place less weight on it is foolish.
TKDC,
There is difference between home runs to be sure. But we are talking about what actually happened. We measure the number of home runs that were actually hit (whether they are “good” home runs or not) with absolute certainty.
This is not true of defensive plays. It is a routine occurrence for a player to be docked for missing a play that is judged that 20% of players would have made, but in reality no one could have made, etc.
So, if a player hits 12 home runs, we know with certainty that his team scored those runs, whether all 12 were lucky Green Monster pop-ups, or whether all 12 were no doubters. They actually happened either way. With respect to the fielding numbers, all we know for sure is if the player made or didn’t make a play. But because we have no good way to evaluate what it means to make a play, we then guess about whether other players would have made the same play. Sometimes the guess is good, sometimes the guess is bad. At any rate, we are not accurately representing what actually happened.
No doubt defense is important. But we have to take the difficulties associated with evaluating it into consideration. It is not the same as hitting, and we have to accept that.
That’s what you said to your girlfriend, right – massive measurement error. Hiyuah!!!
We miscounted Hack Wilson’s 191st RBI for a half-century, but otherwise your point stands.
Folks really need to stop the “UZR is what happened on the field” narrative.
UZR is a MODEL and compares the difference between balls that the fielder played vs an aggregate distribution of balls across three (?) years of major leaguers.
So outside of any bias issues, UZR is assuming that the distribution of balls that a fielder in a particular zone is similar to what Joe Average sees… as the sample gets larger this is a fairly reasonable assumption. Over a small sample it is not (which is why UZR is “noisy”). If the zones were infinitely small you wouldn’t have to worry about zone size, but they aren’t so a fielder over a 1 year sample may see a greater frequency of balls at the edge of the zone of in the center of the zone, so the delta calculation is not exactly a fair representation of “what happened on the field”
UZR is an APPRXIMATION of what happened on the field, not a direct measurement of what did… and the key is whether the distribution of plays in a specific zone is really representative of the aggregate distribution.
WAR, wRC, etc, are all approximations of a player’s value in the batters box. I’m not sure I see a meaningful distinction here.
It is amazing both how far Josh Hamilton has fallen in the last month and a half, but also how his numbers still look as good as they do. As an Angels fan I’d love to see our boy Trout become the youngest MVP ever, but it is also difficult to imagine Hamilton not turning it around and at least getting to 40+ HR and 120+ RBI; couple that with a .300+ BA and the likelihood of the Rangers winning the division, and I think he gets the nod from the voters, unless the Angels steal the division and Trout is even better in the second half.
I don’t agree with trying to find a definition of “valuable” other than “best”. It is clear that when this award was established, it was meant to go to the best player, but an unfortunate word choice was used, which was not intended to have the meanings that people, unfortunately including voters, attach to it.
Nit-picks over the definition of valuable are counter-productive, as each person will have his own definition.
Just give the award to the best player, as originally intended. “Best” itself leaves plenty of ambiguity to nit-pick about, but is at least clearer than “valuable”.
Minor nitpick, it should be the player with the best season, not the best player. If in-his-prime [pick your choice as greatest player ever] missed half a season with an injury, the fact that he’d still have been the best player (no real argument otherwise) does not mean he’d still deserve the MVP.
How is “it clear that when this award was established, it was meant to go to the best player”?
I personally prefer defining it as stated, value. Its much more interesting to reward a player for his value in helping his team win than simply saying player X is the best player in year Y. The best player is already rewarded by winning batting titles/ HR crowns, etc.
Its preferance to be sure, there is no “right” answer.
When deciding on an MVP, I would include contextual stats. That’s just me.
Yeah, the award was not intended to be context neutral, that’s a bias that is applied by statisticians.
So the belief that, in a 4-3 win, a first-inning home run is just as valuable as a ninth-inning home run is a bias?
I was referring to the tendency around here that the guy with the best WAR should be MVP regardless of team performance.
remember when ryan zimmerman was dave cameron’s MVP candidate in 2010? that was funny
See my comment 2 slots down from here…I could not agree more.
I am astonished by how good Carlos Ruiz has been this season. He has been the only reliable hitter all season in the Phillies lineup. He’s a fantastic defender-hey, he can even catch a knuckleball! I look forward to seeing if he can maintain it.
He’s my vote simply because even great catchers have never been even close to this good-and there he is, near the top. Too bad the Phillies stink. He’d be a household name otherwise.
“He’s my vote simply because even great catchers have never been even close to this good”
Um…Mike Piazza? That was a more offensive environment, but even if you wouldn’t call Mike’s ’97 performance better than Ruiz’s ’12, you certainly can’t say it’s “not even close.”
If you want to make the claim that catchers get routinely underappreciated in MVP consideration, then I’ll be the first to agree with you.
Agreed. I count three Piazza seasons that are either better or, at least, “close” to Ruiz’s current season. Joe Mauer in 2009 was probably better and certainly in the same neighborhood too.
Of course, Mauer won the MVP so maybe that helps your case.
Mike Piazza, I’ll give you. I just don’t recall him being as good defensively and Ruiz is just as fantastic behind the plate as he is at the plate. I’ll agree that catchers ARE routinely overlooked, forgotten, unappreciated, the one player that no one seems to pay attention to. Whatever.
I still vote my mid season award to Carlos Ruiz.
do we reward ruiz for being absurdly lucky, as his BABIP 70 points above his career BABIP, and HR/FB more than twice his career rate would seem to point towards?
Yes, he certainly has benefited from some degree of good fortune, but he also has a career-high line drive rate which is probably another factor in his higher babip. Additionally, his line drive rate has been progressively climbing for the last four seasons.
When it comes to MVP—which is loosely defined as “What player was most integral to the success of his team?”—I like to operationalize this as the player whose actual (context dependent) results on the field had a high impact on his team’s chances. I like WPA for this, and you’ll notice this approach implicitly gives a player more credit to the player when other teammates’ success makes this player’s performance less critical. It penalizes production that comes at meaningless game contexts. And it also (at least partially and implicitly) takes into account league and ballpark effects.
Whether or not “clutchness” actually exists is debatable, but I consider the MVP a wholly results-oriented award. And call me naive, but if two players each have 40 HR, but player A’s were literally all walk-off homers and player B’s all came in the 9th inning of blowouts, I’m giving the MVP to player A.
I will concede that this approach doesn’t take into account defensive and positional considerations, the way that a stat like WAR does. That’s a problem, so it’s fairly clear to me you can’t *only* look at WPA. But WPA shares a strength of WAR, which is that pitchers and hitters can be compared.
I’d like to see the award split into two:
Barry Bonds Award – best player in the league
Kirk Gibson Award – player most integral to their team’s success
That way the first would be relatively straight-forward, while the second could have all sorts of fun things going on, and probably a lot more pitchers too.
Pigs will fly before MLB names an award after Bonds ;-)
Seriously, dude, ever heard of Babe Ruth?
I think Ruth would win both. Point is, one is for best player regardless of how his team finishes.
I see your point. Although hitting-wise, there is already such an award (the Hank Aaron award).
There is also award you get when you are considered to be the best player in baseball (context-independent) and are anticipated to continue to be a great player in the future. It is called the Alex Rodriguez award, and he won it once in 2000 and again in 2007. The design of the trophy is somewhat “out there” and creative, it is a pile of cash you can’t even see over, made entirely out of legal tender.
Gibson’s rWAR in 1988 was just 0.6 behind the league leader, giving him a pretty sound argument for best player in the league, grit and leadership aside.
Gibby was far-and-away the league leader in Jason’s patented, ultrasecret Grit Units Above Replacement (jGUAR) with 1.63/100 appearances. A lot of that was accumulated in the postseason, however.
Trout has been amazing this season. It would be pretty cool if he won the MVP. It’s fun to think about what a 20 yo rookie on pace for a 9 WAR season could do in his prime, which is still 3 or 4 years away from even starting.
I appreciate the disclaimer that we all have our own definitions of value, but I don’t find the “WARdiff” charts particularly informative. Has Robinson Cano been more valuable because Curtis Granderson can’t catch a fly ball? Granderson has certainly shared the offensive load, and while Cano’s league-average defense may be needed more with a statue playing to his right every day, I don’t see how he’s more like an MVP because none of his teammates (except Sabathia, whom you ignore here) have been worth more than 1.8 WAR. wRC+ may be more useful here, but I still don’t see how Cano’s differential advantage trumps Trout’s absolute advantage.
Or, in short, I’m glad you picked Trout. I would too.
Trout has been amazing this season. It would be pretty cool if he won the MVP. It’s fun to think about what a 20 yo rookie on pace for a 9 WAR season could do in his prime, which is still 3 or 4 years away from even starting
Actually he is on pace for 10.4 WAR (as of today) if you project WAR/G over remaining games.
I think it’s amazing how good Austin Jackson has been this year. If he hadn’t missed 20 games or whatever it was he’d have a 5+ WAR already. Crazy considering he was the joke of baseball last year.
You could say the same thing about Mike Trout, except he wasn’t the joke of baseball last year (also, either was Jackson…), because he wasn’t even playing Major League baseball.
I agree with your philosophy in the intro, but I’d (respectfully) suggest a slightly different criteria to get at the “relative value” you talk about. Instead of comparing the WAR and wRC+ of the “absolute” best players to the second best player on their teams, why not look at their percentage of the total team’s WAR? (and wRC?). That could be a better way to capture the “relative” value of each player to the whole team.
Using your methodology above, Cano should win first half MVP since he is further ahead of the second best player than Trout is for both WARDiff & RC+Diff. However, Trout has accounted for 19% of his team’s WAR, while Cano produced 15% of his team’s WAR (including pitchers). In other words, Cano’s whole “supporting cast” is better than Trout’s (not just the second best player), and Trout’s contribution to his team is much more important than Cano’s.
The same applies in the NL – McCutcheon accounts for a whopping 25% of the Pirates total WAR, Wright is 23% of the Mets, and Votto 21% of the Red, even though he is having the best “absolute” season of the 3. So, while the argument “the Reds would be nowhere without Votto” is true, the Pirates would be even worse off without McCutcheon (and they’re in first place because of him).
I’m not sure whether “percentage of team WAR” is the best method, and there’s more math needed for wRC+ (you’d have to remove that player & see where the team’s offense would be without him). But I do think it’s better to compare a player’s contribution to the whole team, not just the second best player, to get at relative value.
I like your approach.
25% for McCutcheon. Nice.
Wouldn’t this method, taken to its logical conclusion, mean giving MVP’s to decent players on incompetent teams? Or would we limit it to playoff contenders, so only good players on decent teams in lousy divisions (or a league with parity and two wild cards) would get a boost?
“giving MVP’s to decent players on incompetent teams”
This is at the crux of MVP arguments a lot of years. (“WHATTAYA MEAN, JOSE BAUTISTA?!? THEY’RE IN FOURTH PLACE!!1!” is the common refrain.) Personally I don’t have an issue with regarding the “best” player with the MVP award, whether his teammates are awesome, awful, or somewhere in between.
But interpretations of what constitutes “value” or who was “best” may differ widely, and reasonable minds may still disagree even if they are working from the same definition of “value”.
To Jon L – I’d still start with a list “absolute” best candidates, as Bill did above (my preference is by WAR & WPA) to avoid the mediocre players on poor teams. My method/goal was trying to narrow down the list of “best” players to show which one showed the most “value” to his team. I.e., for McCutch, he’s one of the top five players in the NL, AND provides the most value to his team, AND they’re in playoff contention because of him.
who’s mccutcheon?
Interesting methodology. I quibble though–there’s only one win between Ruiz and Molina, and that doesn’t include their (IMO) substantial differences in defensive ability.
Is Molina a good defensive catcher too?
In a ten team ten keeper league I was offered
Granderson, and A-Gonz for Cargo, I have Harper, J Upton, Rasmus, and Rios in OF, and Cuddyer and Ike Davis at 1st. Is this a smart trade for me from a keeper perspective?
I would take that trade if I were you. Granderson and Agonz in your lineup is much better than Cargo and Davis.
Cargo has the most value long-term, but Granderson and Agonz should be good for a few more years.
yona? lol
Thank you for not writing a single line about Austin Jackson. No, seriously- one sentence would have been pretty exhausting for you, and we don’t want anyone getting hurt out there.
why waste his time on Jackson?
try looking at his numbers?
considering his well documented batting mechanic change resulting in much increased power, those numbers are almost definetly going to stay where they have been and he has been an absolute force out of the leadoff spot.
Now if only Leyland would let him steal a base or two….
While Trout and Cano on paper look like the most deserving candidates, I think voters are going to have a tough time turning away from Hamilton. Barring injury he should still finish with big numbers, even with injury aka 2010 he’s shown he’s very capable of being elite if not the best.
Maybe I am not reading this right but on the leaderboards Bourn’s wRC+ is only 127, not 142
Subtracting the WAR of the second best player from that of the best? That’s not analysis, that’s arbitrary silliness.
If you want to talk about player value in context of the season then put in the effort and build a “Playoff Percentage Added” metric. If you’re not going to go that far, at least put in the time to consider whether the cheesy metric you’ve derived actually means something.
(Or if you don’t believe in extremely timing-sensitive stats but want to add some team/season context there are lots of other ways to do it that aren’t just ridiculous on their faces.)
even as a Yankees fan I have to admit that right now there’s no way Trout doesn’t win the MVP award. Cano was been awesome but Trout has been an alien so far. As a fantasy player I’m still mad about myself for not picking up Trout in any of my leagues, I thought he isn’t ready yet. I think I was slightly wrong ;)
I have to ask, realistically, do Trout or Jackson really deserve it?
Both of them are beneficiaries of vastly inflated BABIPs. Do we want to reward luck in MVP voting?
Should Jackson and/or Trout really get MVP just because a few more of their grounders found their way through the middle?
I argue that while it’s important to look at wRC+, it’s important to look at it in context – and try to determine if that’s a “true” number or if that is really measuring what the player did + some luck.
If they keep it up all year, then yes. For two reasons:
a) Luck is an inextricable element of the game, and very valuable.
b) BABIP isn’t “inflated” just because it’s an outlier. Elite hitters have maintained elite BABIPs over long periods of time. For some recent-ish examples, look at Tony Gwynn Sr., Wade Boggs, Rod Carew.
And I would add that just because the numbers say a player got “lucky” with their BABIP and/or may be overdue for some heavy regression, it shouldn’t matter a bit in the voting (in my mind, anyway). I would cast my vote based on what *did* happen, not what *should have* happened.
I would think most MVP candidates in most years are playing unsustainably well or “over their heads” to some extent, but they’re still putting up awesome numbers nonetheless.
That should have been my first point. Not to mention I doubt anyone ever got voted MVP because they hit the most singles!
Do we start voting against unusually high or “lucky” HR/FB ratio’s too? If that’s the case, then Cano shouldn’t be considered either. This site seriously needs some batted ball f/x so that we can analyse BABIP better. Otherwise, these short-sighted BABIP generalizations are going to continue to rule the land.
If Jackson’s BABIP regresses to his career average, he’s still on pace for 7 or 8 WAR due to the walks, power, and great defense. Trout obviously has less of a career track record of elite BABIP, but he definitely fits the mold of being someone who can sustain very high BABIPs – maybe not as high as it is now, but certainly much higher than the lazy regression standard of .300.
My sense coming into the article was that McCutchen and Trout would probably be at the top of my own lists. After reading, Votto and Wright are serious challengers in the NL, but I’m not sure anyone could resist giving Trout the ROY/MVP combo if he keeps things up in the second half. If he tails off to “merely awesome,” then it’s anyone’s guess.
Just to be a homer and stuff the ballot box again, Melky has to get some consideration.