Monday, January 4, 2010

Projected 2010 Wins Above Replacement

As I did last off-season, I will begin tracking the Nationals' projected wins above replacement (WAR) as the roster evolves.

I'm using Tom Tango's calculations for WAR, Bill James's projections for hitting, and CHONE for pitchers. Why these projections? Because I like them (and they're free and handy). The playing time projections are my own. For fielding, I used each player's UZR runs above average for the last three seasons, regressed to the mean.

A replacement level team would win about 47 games, so by adding the individual WAR of each player, I project the current team would win about 74 games.

For comparison, the Nationals were a 66-win team last season based on run-differential, so my projection says they're about eight wins better than that. You shouldn't compare the 74 wins to the team's actual win total of 59, since WAR projects wins based on pure run differential--the bad luck, lack of clutchness, bad managing, or whatever it was that accounted for the seven game gap in the Nationals' 2009 actual win total and Pythagorean win total will not be captured here.

Most of that is improvement in the bullpen, less because the new guys are so good, but because last year's group was so bad that it's statistically unlikely that any group of major league pitchers will repeat that performance.

Here are the detailed individual projections:

Position players


PT wOBA* FRAA** WAR
C Jesus Flores 60% .323 0 1.66
C Ivan Rodriguez 40% .304 0 0.65
1B Adam Dunn 65% .391 -8.43 2.30
1B Mike Morse 35% .342 0.34 0.51
2B Cristian Guzman 50% .305 0 0.34
2B Willie Harris 50% .318 -1.06 0.68
SS Ian Desmond 65% .341 -4.51 1.89
SS Cristian Guzman 35% .305 -1.95 0.35
3B Ryan Zimmerman 85% .374 10.09 4.96
3B Willie Harris 15% .318 0.52 0.23
RF Elijah Dukes 80% .348 1.15 1.91
RF Mike Morse 20% .342 -0.29 0.38
CF Nyjer Morgan 85% .327 8.48 2.42
CF Willie Harris 15% .318 -0.56 0.21
LF Josh Willingham 80% .361 -6 2.77
LF Adam Dunn 20% .391 -10.79 0.76

Pitchers


IP ERA
WAR
SP Jason Marquis 200 4.60 1.98
SP John Lannan 180 4.53 1.92
SP Craig Stammen 130 5.00 0.72
SP Garrett Mock 130 4.47 1.47
SP Ross Detwiler 130 4.71 1.13
SP J.D. Martin 100 4.54 1.06
SP Scott Olsen 90 5.18 0.32
RP Matt Capps 70 3.63 0.67
RP Brian Bruney 65 3.92 0.41
RP Sean Burnett 65 4.18 0.23
RP Jason Bergmann 65 4.17 0.24
RP Tyler Clippard 65 3.88 0.44
RP Collin Balester 50 4.99 -0.27
RP Eddie Guardado 40 4.26 0.11
RP Victor Garate 40 4.85 -0.15
RP Drew Storen 20 5.88 -0.30

*If you're not familiar with wOBA, it's a composite offensive value metric similar to eQA. That doesn't help you? Sorry. Here's how to think about it: wOBA adds together all a players offensive contributions and then translates it to a 0-1.000 scale weighted like OBP. So about .330 is average, .365 or above is good, and .300 or below is bad. If you scan the Nationals team page on Fangraphs you can get a pretty good feel for what the scale is like or read more here.

**FRAA=Fielding Runs Above Replacement. There's actually a Baseball Prospectus stat called this, but since I like to be confusing, that's not what this is. As explained above, this is my own calculation of the three-year average of the player's UZR runs above average, regressed to the mean to correct for statistical noise.

5 comments:

Positively Half St. said...

I would take 74 wins, I guess, but I would rather see the Nats sign another player who would boost the total 2 or 3 more. Good luck to make up for last year's bad luck could then offer the possibility of an interestig team.

Anonymous said...

I see no way that Lannan pitches under 200 innings and posts a 4.53 ERA. He's way too good to post a 4.53 ERA. I think 4.60 ERA is a little too high for Marquis. I think he finishes around 4.30.


I think your projections have Willie Harris playing way too much. I just don't see him playing half the time at 2B.

Unknown said...

Does anybody remember what our pre-season projected wins were for last year? How did that compare to the 66 we "earned" based on run differential? Thanks.

Steven said...

I don't see Willie playing that much either, but only because I think there's someone not yet in the fold.

Are you thinking that Guzman plays more? Alberto Gonzalez or Eric Bruntlett? If the latter, then the WAR goes down.

On the Lannan and Marquis projections, CHONE is pretty much in line with other projections that have come out. There's certainly a chance that they outperform these projections, but the idea is to provide a mid-point projection, not a 90% best-case scenario that's unlikely to occur for every player.

Steven said...

Sass--Here's the link you're looking for: http://firejimbowden.blogspot.com/2009/03/updated-nationals-war.html.

The projections had them at 77 wins earned, a full 11 wins too optimistic.

Cabrera and Olsen were much worse than the projection systems expected. And the bullpen was far worse than projected, which makes sense. It's hard to project the team being below replacement in an entire phase of the game.

I personally guessed the Nationals would be at 74 wins, mainly because I doubted Olsen, Cabrera, and the bullpen. But I didn't doubt nearly enough.