For the first time since he came into the league, Mike Trout is not the slam dunk choice for American League MVP.

As I often have to remind myself because the numbers in his favor were so compelling, Trout did not actually win the award in his first two seasons in the league. He probably should have though, considering he was almost four wins better than the actual 2012 MVP winner, and three wins better the 2013 winner (in terms of fWAR).

Last season, Trout was finally rewarded for his excellence with an MVP and was 3.4 wins ahead of Alex Gordon, who finished second in the league in fWAR.

Trout once again finished at the top of the AL WAR leaderboards at FanGraphs, Baseball-Reference and Baseball Prospectus in 2015 (he spent most of his season in Double A the last time this was not the case), but the margin was not nearly as wide as his first three years in the game.

He was worth 9.0 fWAR, 9.4 rWAR, and 10.0 WARP. Josh Donaldson is second according to all three formulas, and was worth 8.7 fWAR, 8.8 rWAR, and 7.6 WARP.

While the margin is fairly significant in terms of BP’s WARP, Trout and Donaldson are close enough in terms of the other two WAR measurements that a closer look is warranted.

WAR is a good stat, but it is not perfect (nobody is claiming it is, of course). It is a very good estimate of value, but is still an estimate, and a good rule of thumb is to assume a 0.5-1.0 win margin of error when comparing two players.

When a player is four wins better better than another, we can comfortably assume he has had a better season. When the difference is only a third of a win, though, we cannot be nearly as sure, so this year’s MVP discussion is not as cut and dry as it has been recently.

Before moving on to each player’s case for the award, here are their offensive stats from this season.

And here are their defensive and baserunning numbers.

The Case for Trout

Trout’s career has evolved in an interesting way, as he has gone from a guy who derived much of his value from defense and baserunning to one of the game’s best power hitters in the span of four years.

This is not to say he was ever not a strong power hitter (he hit 30 home runs and had a .238 ISO in his rookie season), but in 2015, he was second in the majors in slugging, third in ISO (.290), and tied for fifth in home runs.

The combination of this power and a .402 OBP gives Trout an American League-best 172 wRC+, and only Bryce Harper (197) was better in the Majors.

Trout’s edge in terms of offensive production arguably puts him over the top relative to Donaldson.

As mentioned, we cannot be super confident in WAR when the margins are slim, but we do have more faith in offensive statistics than defensive ones.

Sports-Reference.com found that past batting runs produced correlate with future ones at a greater rate than defensive runs produced per inning (from one year to the next, r=0.57 for batting runs, and 0.32 for defensive runs). There is also greater disagreement in terms of different defensive metrics than offensive ones, where there is virtually none.

Donaldson had a 154 wRC+ in 2015 (seventh in baseball, fourth in the AL), so while he had a great year at the plate, he was worse on a per plate appearance basis than Trout.

Despite having 29 fewer plate appearances than Donaldson, Trout was also a more valuable offensive player overall.

The Angels outfielder was worth 57.1 park-adjusted runs above replacement with his bat and produced 3.3 baserunning runs above average, according to FanGraphs.

Donaldson produced 44.3 batting runs above replacement and 4.0 BsR, and after positional adjustments, the gap widens further.

Trout produced 60.4 offensive runs above average, behind only Harper in the Majors, while Donaldson ranks fifth with 48.3. He was also worth 8.9 offensive wins above replacement, compared to 7.7 for Donaldson.

Defensive performance narrows the gap in terms of fWAR and rWAR, as Donaldson was worth 1.5 defensive wins above replacement and 10.7 defensive runs above replacement, compared to 1.0 dWAR and 2.1 dRAR for Trout.

(In terms of WARP, incorporating defense actually increases the difference between the two players, as Trout was worth 11.1 Fielding Runs Above Average, compared to 4.4 for Donaldson).

Given Trout’s lead in the three major total value statistics, as well as the components of WAR we can be most confident in, he has a compelling case for MVP.

The Case For Donaldson

Before looking at what could give Donaldson an edge for the voters, let’s first look at something that should not.

As the MVP debate heats up, you might read about why Donaldson should win the award because his team won the division and Trout’s did not make the playoffs.

You won’t get that argument here.

Yes, Donaldson was a big reason the Blue Jays were one of baseball’s best teams, but it’s probably not a good idea to hold Anaheim’s struggles against Trout. It’s not Trout’s fault only two other Angels position players were worth more than 2.0 fWAR (roughly what we should expect from an average player), or that his team ranked 21st in park-adjusted ERA.

No, the Angels did not make the playoffs, but in a context-neutral setting, replacing Trout with even an average center fielder would cost the team roughly seven wins. The 85-77 club would look more like a 78-84 team.

Then again, we don’t live in a context-neutral world and this actually leads into Donaldson’s best case for MVP: context is important to winning and losing baseball games.

Consider a single, a play that is worth, on average, 0.77 runs above the run value of an out. It is treated as such by wOBA, and by extension, fWAR.

Not all singles are equal, though, as a base hit with two outs and no one on is worth about 0.5 runs, while a single with two outs and the bases loaded is worth roughly 2.3 runs relative to an out.

Depending on the score, the hit could have even more value to the team (the two-out, run-scoring single is more valuable in a tie game than one where the player’s team is already winning 15-0).

When calculating wOBA and WAR, it makes sense to use the hypothetical run values, since they are more indicative of the player’s true talent/performance. He cannot control when his plate appearances occur, and there is little statistical evidence “clutch hitting” is a repeatable skill.

This does not mean the “big” hits did not happen though, and for a backwards looking award, it makes sense to give a player full credit for the plays that help his team win.

Michael Lichtman, co-author of “ The Book,” made the point for taking context into account on his website, while arguing against the use of WAR as a means of determining MVP.

“I think that most reasonable people will agree that an MVP has to have had some – no a lot – of articulable performance contributing to actual, real-life runs and wins,” Lichtman says.

If we take this context-dependant approach, Donaldson’s MVP candidacy has some more juice.

Donaldson has the edge over Trout in terms of both win probability added and RE24 (which measures the “change in run expectancy from the beginning of a player’s plate appearance to the end of it”).

Donaldson added 6.27 wins worth of WPA to lead the American League, while Trout is in second with 5.26 WPA. Donaldson also led the league in RE24 (60.35), and was 7.7 runs ahead of Trout, who was again in second.

Trout did hit better than Donaldson with men on base (174 wRC+ to 168), while Donaldson was slightly better before park adjustments (.418 wOBA to .417; neither WPA nor RE24 is park-adjusted).

However, while Trout may have been the better hitter with men on, Donaldson was seemingly more productive in these spots. He drove in 19.7% of all men on base during his plate appearances, the top rate in the American League (and fifth-highest in baseball) among players with at least 300 plate appearances, according to Baseball Prospectus.

Trout, meanwhile, ranked 129th in MLB with a 14.4% rate.

Both players posted a comparable wOBA with men on and Trout had the higher wRC+, but this has much to do with his 23.5% walk rate with men on. Donaldson walked in 15.3% of his plate appearances in this split.

Walks are great, but it is considerably harder to drive in a run with a walk than a hit. Donaldson had a greater rate of singles, doubles, and triples per plate appearance with men on (while Trout had the edge in home runs).

WPA and RE24 are great tools, but they also have a limited role in this discussion. As mentioned they are not park-adjusted, nor is there a league or positional adjustment, and defense does not factor in at all.

It is probably misguided, then, to say “Donaldson is MVP solely because his offense had the greatest impact on his team’s run expectancy.”

That said, it makes much more sense to claim that since two of the three best context-neutral value stats are close enough that we can consider the players even, context-dependant stats should be used to break the stalemate.

This is probably a good approach, and probably the best one for advocates of Donaldson’s candidacy.

In any event, both candidates are deserving, and with no one-dimensional sluggers reaching arbitrary counting-stat milestones, it will be almost impossible for the BBWA to mess this one up.