How This Year’s Best Players Compare to History

Yesterday afternoon, Dave told me he was going to write about the National League MVP. Specifically, he said he was going to write about how there’s just no real good way to pick a winner, because all the best players are almost equally good. Here is the post. It went live in the morning. It’s all sensible, and it does a good job of laying out the landscape.

I said to Dave it would be possible to compare this year’s crop of the best players to how the crops have looked throughout baseball history. I added that I could look into it if he won’t. His response, and I quote: “I won’t.” So now here I am! I’m running through some quick analyses, going back to the start of the integration era in 1947. I’ve prepared plots to examine all of Major League Baseball, and I’ve prepared plots specific to the American and National Leagues. Off we go.

Everything in here, by the way, is based on WAR. As imperfect as it is, especially in the earlier decades of baseball history, there’s no better way of summing up player value in one number. I did this for purposes of simplicity, and so, with that in mind, let’s start simple. Here is the highest player WAR, each year, since 1947.

There are a few games left this season, so we could see a bump. Aaron Judge leads the majors in WAR, at 7.8, and maybe he’ll get to 8.0 by the end of the weekend. Maybe someone else will pass him by. No matter what, this year is fairly unusual — the average high WAR over the several decades is 9.7, with a median of 9.6. The last time the highest WAR was under 8 was 2014, when Mike Trout finished at 7.9. That’s not very long ago. But before that, it hadn’t happened since 1994, when the season was cut short. Before that, it hadn’t happened since 1981, when the season was also cut short. As great as Judge has been, he wouldn’t be No. 1 in most seasons. It’s too bad that Trout had to get injured.

Let’s make the numbers a little more complicated. The graph, too. On this plot, you see three lines. One shows the WAR difference between first place and second. One shows the standard deviation of the five highest WARs. The last shows the standard deviation of the ten highest WARs.

Unsurprisingly, because the best player’s WAR is relatively low, there’s a relatively small spread between first and second. The spread this year is 0.1 WAR; the average is a full win higher than that. The standard deviation of the top five WAR figures is smaller than the average, and the smallest since 2009. The standard deviation of the top ten WAR figures is also smaller than the average, and the smallest since…ever! Or, at least, 1947. By this measure, we’ve never before had such a tightly-grouped top-10. It goes from Judge, at 7.8, down to Charlie Blackmon and Giancarlo Stanton, at 6.4.

That’s for all of MLB. To get league-specific, let’s start with the AL. Again, here are the high WARs.

Both the average and the median are 8.9. Compared to that, this year’s top player has been underwhelming. As recently as 2008, though, there was a worse top player, when Grady Sizemore finished at 7.4. And then, in 2004, Ichiro Suzuki finished at 7.1. Judge is likely to clear that by almost a win. Not the worst season ever! Moving on, now, to the more complicated stuff.

The usual difference between the best and second-best player has been about 1.2 wins. The gap right now is just 0.1 WAR, which is the same as it was two years ago. Looking at the spread among the top five players, the standard deviation is half of the average, although it was lower as recently as 2008. Looking at the spread among the top ten players, the standard deviation is right on the average. The spread among the top ten players has been normal. It stretches from Judge, at 7.8, to Mookie Betts and/or Carlos Carrasco, at 5.0.

We’ll wrap up with the NL. Here are the high WARs.

Do you know who leads the National League in WAR? It’s Kris Bryant, at 6.7. And, it’s Joey Votto, at 6.7. And, it’s Anthony Rendon, at 6.7. All of them are great, but still, this year is atypical, because the average high WAR over the whole span has been 9.0. The last time someone led the NL with such a low WAR was 1988, when Andy Van Slyke finished at 6.4. Since then, no one has led with a WAR lower than 2010 Votto’s 7.0. As you probably already knew, the best players in the NL are all bunched up. It’s further revealed in the following plot.

The average distance between first and second has been 1.0 WAR, with a median of 0.7. At this point, this year, we’re looking at double zeroes. It’s the same as the difference between second and third. So then, think about the spread among the top five players. The standard deviation is 0.2 WAR. The average is 1.0. There’s never before been less top-five spread, at least since 1947. What about looking at the top ten players? The current standard deviation is 0.5 WAR, while the average is 1.1. The standard deviation has only before been lower in 1988 and 2010. That is, we’re looking at a total of 71 baseball seasons, and among the top ten players in each year in the NL, this year has the third-lowest spread.

Depending on how you look at it, then, either this is the most even year in modern NL history, or it’s just close to it. As far as WAR is concerned, there’s never been so little difference among the top five players, and there’s seldom been so little difference among the top ten. Not that MVP voting should just come down to a list of the best WARs in the first place, but if you figure that voting and WAR have a pretty strong relationship, you can see why this year’s NL MVP voting results are likely to be mixed. No one has singularly emerged; plenty of players all have compelling arguments. I don’t think we can expect much separation over just the final few days, with all but one playoff spot locked up.

Van Slyke didn’t win the NL MVP in 1988. He finished fourth, without a first-place vote. Kirk Gibson finished first, although the first-place votes were spread among three names. In 2010, somewhat surprisingly, Votto won, and it was almost unanimous, with just one first-place vote escaping to Albert Pujols. Even though Votto wasn’t head and shoulders above the rest of the field, the voting results put him on a pedestal. I guess it’s possible that, this year, the voters could all rally behind one single player. It’s happened before, and it could happen again. Nothing, I suppose, should be taken for granted. But I still find picking an NL favorite virtually impossible. Good luck to the people who will actually have a voice.

Jeff made Lookout Landing a thing, but he does not still write there about the Mariners. He does write here, sometimes about the Mariners, but usually not.

Newest Most Voted
Inline Feedbacks
View all comments
Da Bum
6 years ago

Obviously not a leader or top player with all facets of the game but JD Martinez is having the greatest offensive September ever. His slugging, just slugging, is over 1.000 with an OPS close to 1.500.

In the 2nd half he has hit a HR every other game (31 in 63 games), more than a RBI per game (73 in 63) and has an OPS of 1.149. His 2nd half wRC+ is an absurd 182 and he’s done it all with a BABIP of .321.

6 years ago
Reply to  Da Bum

In Martinez’s case, the unsustainable bit isn’t going to be BABIP, it’s going to be HR/FB%.

Votto’s 2nd half OPS in 2015 and 2016 were both higher than 1.149, with wRC+ over 200.

Martinez is having a fine, fine run…but there’s nothing historic about it.

Da Bum
6 years ago
Reply to  JDX

His September is.

6 years ago
Reply to  Da Bum

Just in the past decade or so, there have been ten players who have managed a wRC+ of 200 or more in a pre-ASG or post-ASG half:

2006 Howard 202, 2d half
2008 Ramirez 201, 2d half
2011 Bautista 209 1st half; Napoli 213 2d half
2012 Posey 200, 2d half
2013 Cabrera 207, 1st half
2015 Harper 211, 1st half; Votto 213, 2d half; Encarnacion 202,2d half
2016 Votto, 202, 2d half

Trout was well over 200 at the ASB this year, but I wouldn’t count him since he had missed 45 games at that point.

Notice there have been more in the second half than the first half. Small sample size, but makes sense because the second half is fewer games, there is larger variation in hitting stats. In fact, the list of players who have had a wRC+ of > 200 over a stretch of 81 or more games in that period is limited, AFAIK, to Bautista, Cabrera, Harper, Votto (2015), and Judge this year. Harper has the longest such stretch, 147 G, 638 PA. He fell just a few games short of pulling it off for a full season. There have been only four players with 200 wRC+ seasons in the past 60 years, two of them known dopers (Bonds and McGwire), and the other two at least suspicious, and in any case did it in a strike-shortened season (Thomas and Bagwell).

Da Bum
6 years ago
Reply to  WARrior

Right. You just ignoring my first paragraph. I led with September for a reason. I know his 2nd half overall isn’t in the convo for best ever. His September is.

I mentioned his 2nd half stats because 31 HRs in 63 games is crazy, as is his 73 RBI. They are anecdotes to note but not my main point which is why I didn’t lead with them.

6 years ago
Reply to  Da Bum

Yes, JDM’s September is awesome, but still not unprecedented. Harper had a comparably productive month in May 2015, Trout in July of the same year, and Encarnacion in August. If you want to restrict the discussion just to September, we might have to go back to Bonds, 2001: 16 HR, SLG 1.078. Far better OBP, his BB/K just about the inverse of JD’s.

Mean Mr. Mustard
6 years ago
Reply to  WARrior

As far as I know, there was nothing suspicious about Thomas at all, besides his physique. He was a known proponent of stronger testing.
You may say, “he who protests loudest has most to hide” in response, but to that i say “Pfah!” His Topps rookie card stats don’t differ much from his last playing days, insofar as height/weight goes.
You may say “Pfah! Who cares about listed height/weight on a baseball card?”
To that, I say, “His physique as far as photographic appearance didn’t change much, if at all, throughout his playing days”.

Regarding Bagwell, well…there’s no doubt he bulked up through the years, but he didn’t just show up one day looking like Popeye. Looking at photographs, there’s a natural weight-training progression. Maybe I’m wrong and just don’t want to be, given that I live in Houston and loved that messed-up swing. But it’d be hard to convince me, outside hard evidence, that he did anything more than live in the gym.

As far as the strike-shortened season, that magical ’94…well, you got me there.

6 years ago

Never heard any suspicions of Thomas. He’s just as big now as he was when he played. Then again, I don’t care.

6 years ago

Let’s be honest about Bagwell for a second:
-He was a low pedigree prospect with a strong batting eye but virtually no HR power. HR power suddenly emerged when he made the jump from AA to the Majors. Slightly suspicious, nothing more.
-He achieve almost unparalleled heights for a prolonged period of time. By itself, that means little, but the “steroid era” saw many rare-to-unique accomplishments due to “assistance”, making this a tiny but real piece of the puzzle.
-His peak was extended fairly long for the typical aging curve (extraordinarily common during the steroid era, for obvious reasons). Again, a tiny piece.
-His body and production suddenly fell off, starting in the year when the secret test was held and then ending (with the abrupt end of his career) in the first year of public testing. This is suspicious as heck, though not a complete indictment by any stretch.
-Withing a VERY short amount of time, Bagwell’s physique shrunk by an astonishing amount (similar to what we saw with Ivan Rodriguez, except Bagwell had already left the game when his happened). This, combined especially with the previous item, is pretty damning.

I’m not particularly interested in semi-officially labeling anyone a PED-user without stronger evidence than this, so I do not support calling Bagwell one. Pretending that his PED-use case is flimsy or non-existent, though, is almost insulting. Given the above case, and the general circumstances of the era itself, it is more likely than not that Bagwell was a user. Again, I do not believe that is a good enough standard for guilt, and I wouldn’t bring it up at all if not for your comment, but you need to be fairer than that, I think.