Yes, I meant, 'density,' not, 'destiny.'

With the unavoidable talk of Big Wilkie coming to DC, what would the impact be?  Well, this shows the fun of the W.A.R. spreadsheet--we can plug in Dunn's numbers and see how it changes our expected wins!  What is the statistical case for (or against!) Dunn-K?

The shocking statistical conclusions after the jump!

UPDATED!  Since the original post, I added a comparison of how Dunn would change things if NJ misses the entire season.  I also added an estimate of Willingham's fielding runs to the original calculation.

We all know the upside of Adam Dunn.  He's a "guaranteed" 40 HR/100 RBI.  He strikes out a lot, but he walks a lot, too--a real "three true outcomes" (BB, K, HR) player.  Where do we put him, though?  The obvious choice is 1B, where we minimize (we hope) his defensive weaknesses, and his bat is a little bigger than the fragile NJ's (even if his OBP is lower).  There have been rumors that Dunn wants the OF, which would mean LF, where he would displace Josh WIllingham in our current projection.  Here are the Marcels batting projections for Dunn and the people I currently have pencilled in at LF and 1B:

Player		OBP	SLG	OPS
Adam Dunn	.374	.497	.870
Nick Johnson	.394	.464	.858
Josh Willingham	.355	.453	.808
Ronnie Belliard	.333	.420	.753
Dmitri Young	.352	.435	.788
Willie Harris	.337	.397	.734

Dunn is a clear improvement over anyone he'd replace in terms of offense--although not that much of an improvement over NJ.  Of course, we also have to project his impact on the defense.  CHONE has projections of how many runs a player will save/cost you in 2009, and here they are for the same set of players (<0 means player costs you runs, >0 means player saves you runs):

Player		1B	LF
Adam Dunn	-6	-13
Nick Johnson	+3	n/a
Josh Willingham	-3	-8
Ronnie Belliard	+1	n/a
Dmitri Young	-9	n/a
Willie Harris	n/a	+14

Okay, Dunn is better with the glove than Da Meathook, but that's hardly anyone's standard for fielding excellence (and I realize that technically I had Hammer with some time at 1B, but CHONE doesn't have IF projections for him (UPDATE--I gave him a value of -4 runs, since that's just under what Kory Casto has at 1B, and he's slightly worse than Casto in the OF--yeah, I know they're different skill sets, but it's good for a rough guess.)).  Now let's figure out how this changes the lineup. Dunn is durable, so we'll give him 590 PA in either LF or 1B.  I'm not sure what happens to the roster if we get him, so I'm going to assume that if he's a 1B, we trade Nick somehow. If he's a LF, Hammer becomes 5th OF and backs up 1B "when" NJ gets hurt. So we're talking either:

• LF: Dunn (590), Willingham (105) or
• 1B: Dunn (590), Belliard (55), Young (50)

In either case, the rest of the lineup is like the current W.A.R. spreadsheet in terms of playing time.

So, what happens?  I made a new copy of the W.A.R. spreadsheet and plugged in Big Wilkie, as above.  Here it is.  There are three worksheets, one without Dunn-K, one with him at 1B, and one with him in LF.

Here's the upshot (updated with Hammer's 1B fielding value):

• Nats without Dunn: 74.9 wins
• Nats with Dunn-1B: 74.5 wins
• Nats with Dunn-LF: 75.4 wins

WTF?! Big Wilkie only gets us half a game in LF, and costs us half a game at first? It makes sense: take a look at who he's replacing!  Dunn is a more of a slugger than NJ, but not even Dunn gets on base as much as Discerning-Eyed Nick Johnson. In sabremetrics, OBP gets more runs than SLG, so it turns out that NJ and Big Wilkie have the same expected run production! Dunn is better than the people who fill in "when" NJ gets hurt, but his stone hands cost almost a full game over the soft-gloved Johnson.  His bat can't overcome it.

The story is different in LF (which surprises me--I thought 1B would be the better fit). Here, he's only about half a game worse with the glove than The Hammer, and his bat is a full game better in terms of extra runs.  (As an aside, Dunn's exact value in 590 PA in LF is 1.9 Wins Above Replacement (WAR). Willie Harris, thanks to his speed and insanely sticky hands, is worth 2.0 WAR in LF with 590 PA...)

Chicks dig the long ball, I know, but the numbers like someone who can field*, as well (it's not the strikeouts that hurt, it's the terrible range).  According to the stats, Dunn ain't worth it even if we get him for free.  (Of course, if NJ pulls up lame in mid-April, Dunn looks way more useful.  But if Nick lasts past the All-Star break, it's a wash.)

Am I missing something, here?  Even though numbers don't lie, they can be misleading--what makes Dunn worth it (besides our NJ health fears)?

UPDATE--What if NJ breaks before Opening Day?

Nats News Network pointed out in the comments that JimBo is probably comparing Dunn to a lineup with a broken NJ, so I thought I'd try the same exercise.  My "worst case no-NJ Nats" lineup takes away NJ's 400 PA, and replaces them with 300 PA from Kory Casto, and 100 extra 1B PA for the Hammer.  TAWH picks up Hammer's missed LF PA, and gives up the 55 PA I gave him backing up Zimmy at 3B to Alberto Gonzalez.  (I also took a guess at Hammer's 1B fielding value as -3 runs.)

This lineup is the Robot Apocalypse: 72.3 wins.

We plug Dunn in at 1B, and it's just like above, 74.5 wins.  However, if Big Wilkie insists on playing LF, we'll figure Hammer takes most of the time at 1B, with TAWH backing up LF.  This lineup is 74.4 wins. (If we get a 2006 NJ who gives us 590 PA, the Nats get 75.4 wins without Dunn!)

Bottom line: if NJ is healthy (or even mostly healthy), Dunn adds nothing, and might even cost us a win.  If NJ gets hurt during Spring, Dunn adds 2 wins.

Here's the "broken NJ" spreadsheet if you're interested in the breakouts.

* Dunn had a decent year in the field in 2008, although he was firmly in the bottom half of NL LFers with the glove in 2004-2007. By comparison, CHONE projects Ex-Nat Alfonso Soriano (whom I personally instructed in proper fielding from my CF seats in RFK in 2006 YOU GOTTA CHARGE THOSE, YOU LAZY BUM!) to be worth -2 runs in LF in 2009, vs Dunn's -13.