Is Josh Donaldson on a Hall of Fame trajectory?

Since the start of the 2013 season, a look at the leaderboard of position player WAR will show Josh Donaldson in second place of all of baseball, trailing only the second coming of Mickey Mantle. After Trout/Donaldson, there's a big gap down to the next tier, a bunch of excellent baseball players who have put together Hall of Fame careers or are well on their way in the case of the younger ones.

Looking at that got me thinking about Donaldson's chance of one day making it to Cooperstown himself. The general thinking I've seen is that his late start - his first productive season came at age-27, a point when most Hall of Famers have already put up a lot of value - makes it very much an uphill climb unless he ages very gracefully. So I wanted to take a closer look.

Indeed, a look at the post-WWII leaderboard of third basemen through age-30 shows this. The top is populated by Hall of Famers and active future Hall of Famers, many of whom had already clinched their spots in Cooperstown or were very close to it, with upwards of 50 WAR.

Donaldson sits in 32nd with 30.5 fWAR, surrounded by the likes of  Edgardo Alfonzo, Al Rosen and Troy Glaus. Very good players all, but no one's idea of Hall of Famers. A strong end of the season could bump him up a bit, but we're still looking at Matt Williams and Graig Nettles comps. So clearly, he's got a lot work to do.

But those players didn't have the peaks that Donaldson has had over the the last four seasons, and therefore neither the projected future production that Donaldson does. In FanGraphs 2016 Trade Value series, Dan Szymborski's ZiPS system projected Donaldson at 6.4 WAR in 2017 and 5.6 WAR in 2018. Extrapolating that, from 2017-24 through his age-38 season, Donaldson would project to add another ~30 WAR.

That's just a projection, but it would give Donaldson about 60 career WAR, which is at least bubble territory. However, WAR is a counting stat that doesn't discriminate between compiling production over a long amount of time, and producing at a very level over a shorter period of time. How to balance those two factors is far from a settled debate, and it will likely be a fundamental factor in assessing Donaldson since he will likely have a shorter career by HOF standards due to his late start.

To try and incorporate both these factors, Jay Jaffe's JAWS system averages a player's peak 7 year WAR with their career WAR, and compares it with the average of all Hall of Famers at the position. For third base, it's about 55. Using the projections to fill out his career, Donaldson would come in at 53 (average of 60 career WAR and 46 WAR seven year peak). That would put him just below the average, which actually would probably give him a pretty good chance, especially considering he'll have won at least one MVP award.

While JAWS is a pretty good system, my issue is that it uses WAR - that is, wins above replacement. The replacement level concept is necessary when to valuing and paying major league players in the major leagues, since there is a shortage of average players and below average players have significant value beyond the league minimum.

But for Hall of Fame purposes, I don't think it makes sense to use WAR. Every year, plenty of players who had long careers producing at least league average levels hit the ballot and barely attract any support. Having looked through this a bit, it seems to me that from a Hall of Fame perspective, the replacement level is at least the overall league average. Therefore, when looking at the Hall of Fame, it makes far more sense to consider WAA (Wins Above Average) than WAR.

Last winter I was thinking about productivity versus time played in the context of relief pitching, so I charted ERA- against innings for all players post-WWII payers (the modern, integrated era). Since ERA- compares to the league average, this uses average as the baseline. To my surprise, I found a surprisingly clear dividing line between those chosen for Cooperstown and those not chosen. So I thought doing a similar thing for position players might shed some light on Donaldson.

The first issue is that there's no similar stat to ERA- for position players. There's wRC+, but that's only for batting runs. So I translated wRC+ to a stat that considers all runs - batting, fielding, baserunning, positional - instead of just batting. I'm calling it TRC+ (Total Runs Created), so we can at least roughly compare all aspects of a player's value to each other on a rate basis.

Below is a chart of all post-WWII players with at least 2500 plate appearances, with active players excluded as well as those who remain on the ballot (eg Tim Raines) as well those yet to hit the ballot (eg Derek Jeter):

There's a lot here, so let's break this down. The first group, in green, is players not in the Hall of Fame, clustered along the left and bottom. These are basically players who were quite productive, but had short careers; players who had long careers but were not that productive (by HOF standards); or both.

The next group is the light blue, and that's players elected by the Veteran's Committee. I separated them out because over the years it acquired a reputation for cronyism in selecting inductees (especially the interwar period), and while there's some worthy inclusions, there's also a bunch of light blue points surrounded by green points. These are players - Phil Rizzuto, Nellie Fox, Bill Mazeroski to name some of the more egregious - for whom it's really hard to make a case compared to others excluded.

Then we have the dark blue, the writers. For all the criticism of them, the reality is over the years if nothing else, they're done a reasonably consistent job. Jim Rice squeaked in, and is surrounded by green, and the one green point surrounded by dark blue Hall of Famers is Rafael Palmeiro, who is a special case.

There's actually a reasonably clear dividing line, starting at about 8,500 PA and around 35% above average run creation and extending through 10,000 PA and about 25% above average before it plunges down. Above that line, a player is likely to be elected. Below that, it's unlikely and along the line are the borderline.

The red triangle in the upper left is Donaldson. In terms of run productivity, would rank in the top-10 over the last 70 years, though this only includes the prime of his career and will drop as his decline years get factored in. The question is how much.

One major takeaway is that no player who has produced runs at 50% above the league average is excluded from Cooperstown, even with very short careers. On the other hand, there are very few players in the Hall of Fame with less than 8,000 plate appearances, and for most there were extenuating circumstances that wouldn't apply to Donaldson. Jackie Robinson, Larry Doby and Roy Campanella's careers started late due to the colour barrier. Johnny Mize and Joe DiMaggio each lost three prime years to military service; Joe Gordon and Ralph Kiner lost two.

That leaves Lou Boudreau as a potential comparison for Donaldson. He had a short career, only 7,023 plate appearances, but was highly productive, producing runs 49% above league average. He too won one MVP, though he was also a player-manager for most of his career (and credited with inventing the infield shift). Still, it took him a half dozen elections before he was elected by the BBWAA.

***

To make the Hall of Fame, the first thing Donaldson needs is at least another 4,000 plate appearances, and probably more like 5,000. That's about another seven full seasons, so he'll have to stay reasonably healthy. If he does that, without a catastrophic dropoff in performance, he should have a pretty good shot to be inducted in Cooperstown.

One final angle: what if Donaldson remained at MVP levels for the next couple years, and then had a catastrophic injury or production dropoff? There's always the Sandy Koufax precedent. He played 12 seasons, but the first six his ERA- was exactly league average. If you agree that average production does nothing for Hall of Fame purposes, then his case is all about those last six dominant years, over which he had a 156 ERA+. That's almost exactly Donaldson's level of excellence, with a 159 TRC+. If that was enough for Koufax, why not Donaldson?