Let's Discuss WAR

Michael Ciaglo-USA TODAY Sports

Whenever I analyze a player's Hall of Fame case: I start with JAWS, which (if you are unaware) stands for Jaffe Wins Above Replacement Score. It's a system which combines a players career bWAR, peak WAR (or WAR7), and averages them to compile a number. The theory being you average the 'greatness' of a player, with how long their career lasted to give you an idea of how good a player was in their career. Invented by Fangraphs' Jay Jaffe: he invented it in an attempt to reduce the subjectivity inherent in Hall of Fame arguments.

Now, WAR is not a precise statistic and is not meant to definitively determine whether a player with a 5.5 bWAR is better than a 5.7 bWAR (for instance). In that same vein: JAWS is not meant to definitively say that one player is better than another. For instance: Larry Walker ranks ahead of Tony Gwynn in Right Field JAWS by 3.2 wins: does that definitively mean Walker was a superior player? Probably not (at least in the eyes of most fan). However, JAWS does a great job of breaking conversations out. For example; the Hall of Fame cases of Mark Texiera and Todd Helton are really not that similar, and JAWS shows that with Todd Helton ranking 15th all time for first baseman (just barely below the standard), while Mark Teixera ranks 10 wins below the standard. In shorthand you can use JAWS to explain that Helton is a good candidate, if perhaps a borderline case while Tex simply falls well short of the Hall.

This obviously takes what would be a lengthy explanation down to a couple sentences (in obvious cases), and that is where WAR really shines: to differentiate between relatively large differences. But while WAR is a useful tool it is not infallible, and also is only as objective as the subjective assumptions placed into it. Chris Bodig of Cooperstown Cred regularly discusses this, especially in regards to the defensive component of WAR, which is an evolving process. Unlike offense there remains much to be learned defensively on what is and is not valuable, which leads to a noticeable bias in defensive statistics in WAR both for and against modern players.

All of this brings me to the twin topics I would like to discuss today: Andruw Jones and Gary Sheffield. Both of their WAR scores are impacted heavily by how bWAR measures their defensive value (or in the case of Sheffield: lack-thereof). First here are their JAWS scores and the standards at their respective positions (Career WAR, WAR7, JAWS):

Andruw Jones, CF: 62.7/46.4/54.6 (11th); AVG: 71.9/44.8/58.3

Gary Sheffield, RF: 60.5/38.0/49.3 (23rd); AVG: 72.1/42.5/57.3

Both players fall below the JAWS line for the Hall of Fame, although Andruw Jones ranks far closer to the average standard than Sheffield. Jones' rank intuitively leads you to believe that he should be worthy of induction. How could anyone rank so high and not be worthy*? Gary Sheffield on the other hand is sandwiched between Bobby Bonds and Mookie Betts(!).

Now defensive (dbWAR) factors greatly into both players' cases. Andruw Jones ranks as the greatest defensive centerfielder of all time by dbWAR, six wins above Willie Mays (who's career was nearly 4,000 PAs longer). On the flip side Gary Sheffield ranks as the worst defensive right fielder of all time; five wins behind Dave Winfield (who's career was over 1,000 PAs longer). So if we take the WAR score at face value: we need to believe that Andruw Jones was not just the best defensive centerfielder of all time, but over 33% better tan Willie Mays. We also need to believe Gary Sheffield was not just the worst defensive right fielder of all time, but almost 20% worse than Dave Winfield (which in and of itself is troubling: Dave Winfield won seven Gold Gloves).

Now, I am not suggesting that I completely discount dbWAR: I am saying that we should take those rankings with a grain of salt. To use this mathematically let's rank Jones as Willie Mays' equal, and Gary Sheffield as Dave Winfield's equal, just to set a baseline. If we drop Andruw Jones 6 wins (to make him Willie Mays' equal, to be clear) his career score drops from 62.7 to 56.5, which then drops his JAWS score from 58.3 to 50.6 (and this does not even touch his peak score) which ranks 15th all time behind Jim Edmonds. If we tack 5 wins back onto Sheffield's score he goes from 60.5 to 65.5 making his JAWS 51.8, moving him to 17th, which is right below Ichiro Suzuki. Reposting them here:

Adjusted Andruw Jones, CF: 56.5/46.4/50.6 (15th); AVG: 71.9/44.8/58.3

Adjusted Gary Sheffield, RF: 65.5/38.0/51.8 (17rd); AVG: 72.1/42.5/57.3

Obviously this is just a thought exercise, but I think these adjustments bring some serious ramifications to both of their cases. However, this exercise really hurts Andruw Jones who (I think) really benefits from his high JAWS score. Here is Keith Law:

The first is that he’s the best defensive center fielder in MLB history, and it’s not particularly close. By TotalZone, the best available metric for estimating defensive value for players prior to this century, he’s at 230 runs saved in center, 54 runs ahead of Willie Mays, in about 10,000 fewer innings played at the position.

But more importantly: Andruw Jones remains on the BBWAA Ballot and Jim Edmonds fell off after a single season. Why? I think the answer is dbWAR.

Look, Jim Edmonds and Andruw Jones are similar players, both were fabulous defenders with lots of power. Jones hit more homers (434 v 393), but Edmonds was the all around better offensive player (OPS+ of 132 v 111, .376 OBP v .337, etc). Edmonds also racked up lots of hardware (eight Gold Gloves, only two less than Jones). BUT: Edmonds dbWAR slots Edmonds as 15th all time in total JAWS while Jones ranks a far more satisfying 11th (ahead of Hall of Famers Richie Ashburn, Andre Dawnson and Billy Hamilton). Edmonds sits next to Willie Davis and Jimmy Wynn (not in the Hall of Fame).

I am not saying that Andruw Jones does NOT belong in the Hall of Fame. I am saying that I am not convinced the defensive statistics are accurate enough to differentiate that much between Edmonds and Jones. Furthermore, Edmonds did many other things well. Edmonds more frequently scored more runs, hit more doubles, and walked more than Jones. I am quite confident Edmonds was the superior offensive player. While I am confident that Jones was the better defensive player than Edmonds, probably much better, I am not convinced we've narrowed this down to a complete science.

Gary Sheffield's defense is another matter, and one I have written on personally, but I am far from alone. Several writers are strong advocates of Sheffield. Bill James, for instance, does not believe that he was such a poor defender. Jay Jaffe has personally written about Gary Sheffield's defense and questioning the defensive statistics writing:

I’m troubled by the extent to which those outlying defensive stats — largely estimates from the pre-batted-ball-type era — nuke Sheffield’s value. That goes double when they’re compared to his defensive numbers via alternative methodologies. Baseball Prospectus‘ Fielding Runs Above Average pegs him at -89 runs for his career, and Michael Humphreys’s Defensive Regression Analysis, which is available at the Baseball Gauge and has been incorporated into the sabermetric component of the past eight years’ Gold Glove voting, puts him around -108 runs. Both are bad, but neither is as extreme an outlier, and such figures push him much closer to the JAWS line for right fielders

So clearly there is wriggle room to look at defensive statistics, which are still less understood than offense. I personally have written extensively in the past on why I think Gary Sheffield's poor defense should be strongly considered against him. Here is what I wrote two cycles ago:

Which brings us back to the question: should we penalize Sheffield for (as its argued) his manager's decision to play him in the field? The answer is: yes, of course we should. We penalize Edgar Martienz for the Mariners' stupidity in not promoting him full time before his age 27 season, we penalize Omar Vizquel for Mike Hargrove batting him second in the lineup. How can we not penalize Sheffield for his poor defensive play? Every facet of the game matters, and we must account for each part of it. Of course hitting accounts for more, but the fact that Gary Sheffield was an insane, feared, Hall of Fame hitter is why we're even having this discussion in the first place. If Sheffield was any worse as a hitter: we would not be having this discussion. Heck, we penalize Edgar Martinez for playing DH: he was a good defender at third base, his primary position.

I stand by my belief that we must consider the entirety of a player's contributions to the game (at least on the field). Gary Sheffield was a tremendous hitter and an atrocious defender. Perhaps my zeal to call him the worst defensive outfielder in history is premature. So instead of desperately trying to accurately measure his poor defense let's take a look at Sheffield offensively.

First, let's use a modified form of JAWS by isolating the batting component of bWAR (henceforth called obWAR) and split Hall of Fame careers into career obWAR, obWAR7 and average for a tweaked JAWS. By this ranking (which you can find here) Gary Sheffield ranks 7th in RF in our modified JAWS ahead of, among others: Al Kaline, Reggie Jackson, Tony Gwynn, Roberto Clemente and the two most recently inducted right fielders Larry Walker & Vladimir Guererro. In short: Sheffield was a Hall of Fame hitter for a long time.

Now, I to also keep in mind, we are still ranking Sheffield as the worst defensive player at his position of all time, tied with Dave Winfield. As noted, Gary Sheffied ranks as the superior hitter to Winfield, who ranks 11th. Let's compare Winfield to Sheffield more closely:

Gary Sheffield
.292/.393/.514 (OPS+ 140) with 2,689 H, 509 HR, 1,676 RBI, 467 2B, 1,636 R, 1,475 BB-1,171 SO, 253 SB (71%) in 10,947 PAs

Dave Winfield
.283/.353/.475 (OPS+ 130) with 3,110 H, 465 HR, 1,833 RBI, 560 2B, 1,669 R, 1,216 BB-1,686 SO, 223 SB (70%) in 12,358 PAs

The larger stat line does follow what obWAR suggests: Sheffield was a better hitter than Winfield, roughly 10% better by OPS+ (wRC+ rates Winfield 128 and Sheffield 141). Overall the body of evidence favors Sheffield. Looking at Gray and Black Ink it's a bit more muddled. Gary Sheffield's Gray Ink is 123 while Dave Winfield's is 152 (Black Ink is equal at 4). Why is this? Gray Ink counts the number of times a batter ranked in the Top 10 in: Home Runs, Batting Average, RBIs (4 points each), Runs Scored, Hits, Slugging % (3 points each), Doubles, Walks, Stolen Bases (2 points each), and Games Played, ABs, and Triples (1 point each). Here are their rankings:


HRs: 7, BA: 4, RBIs: 6
R: 5, H: 2, SLG%: 5
2B: 0, BB: 9, SB: 0
G: 1, AB: 0, 3B: 0


HR: 7, BA: 4, RBI: 10
R: 7, H: 3, SLG%: 7
2B: 3, BB: 2, SB:
G: 3, AB: 1, 3B: 3

In short: Winfield was a slightly better power hitter for his day and played more games, while Sheffield got on base more. I really struggle to see much of a difference between them, and certainly not one where I can find a strong advantage for Dave Winfield over Gary Sheffield offensively.


Overall my position has softened on both players. While I am not particularly enamored by either case (they are both borderline for me): I am 100% open to both players entering the Hall of Fame, which is a switch for me. Last year I left both Jones and Sheffield off of my virtual ballot (both just barely off). This year I would be more inclined to include both. Andruw Jones may be only the 5th best center fielder outside the Hall (as I ranked him last year), but I would put in basically every one of those players. Same with Gary Sheffield.

I remain adamant that generalists are quite underrated, and specialists are quite overrated, but that does not mean specialists cannot earn induction too. While Jones & Sheffield are both specialists (in their way): that should not cloud my thinking.

FanPosts are reader-generated, and do not necessarily reflect the views of Covering the Corner or the Covering the Corner staff.