It's time for my WAR projection. Unsealing the envelope...the answer is 77 wins. That is 1 win more than last year.
Most of you who read TCB are aware of "Wins Above Replacement Player" or WAR. The WAR method can be used to project a team's win record, by summing players' projected WAR values for the season. The key inputs for individual players (wOBA, baserunning runs added, defensive runs above average, ERA, playing time, pitching leverage), though, can be subjective. Forecasts, by definition, are uncertain. That's a nice way of saying, "I gave my best shot at predicting the future, but I could be very wrong." I think this particular Astros' team is difficult to project, and the range of realistic possibilities around the 77 win projection probably is larger than usual as a result.
Last year, Stephen Higdon (DQ) and I prepared a joint WAR projection, which pointed to 79 wins. I have used the same spreadsheet as a starting point for 2011. (Thanks, Stephen!) Frankly, I was surprised that that my result for 2011 is as high as 77 wins. Given that many analysts foresee an Astros record of 70 wins or less, and the fact that the Astros outperformed their pythagorean record by eight games in 2010, I expected a win total in the low 70's. But, after reviewing my inputs, I feel that the 77 win projection is realistic.
Generally speaking, I selected wOBA and ERA projections for players which fell within a range encompassed by the Bill James projection, fans' projection, and Marcel, as shown the fangraphs' players pages. Some players don't have each type of projection (or no projections), and I tried to select reasonable numbers based on past performance. In most cases, the projections for given players are similar. If they are not, I selected the values which seemed most reasonable. For defensive values, I examined players' previous DRS and UZR performance, tweaked those numbers to reflect my own opinions, and ultimately made a judgement call.
I feel that the WAR calculation approach is particularly useful in considering "what if?" or sensitivity questions. For example, what if Brett Myers and Wandy Rodriguez experienced injuries which affected their performance and innings pitched? If Brett and Wandy run into trouble which hikes their ERAs to 4.25 and reduces their workload to 130 and 160 innings, respectively, then the Astros' wins decline by four to 73 wins. (That struck me as a lessor impact than I expected.)
On the offensive side, I calculated a couple of alternative possibilities for a higher Astros' W/L record:
- The projected offense for Carlos Lee is lower than his career levels, and I think there is room to believe that Lee could have a bigger bounce back. Optimistically, if Lee's wOBA increased to his 2008 level, the Astros' wins would rise to 80.
- I also calculated a "high offense" scenario, based upon Wallace and Chris Johnson breaking through with low to mid .800's OPS, Lee posting a 2008 type offense, and Hunter Pence boosting his OPS into the high 800's. This would put the Astros over the 82 win mark and a winning record. While the assumptions, individually, are not unrealistic, the combination probably represents a good deal of luck.
A noteable point is that I have assumed that Jeff Keppinger will be used liberally as a super-sub when he returns from his injury. Because of the earlier rumors about the Astros' interest in trading Keppinger, I wondered how important his role would be to the ultimate W/L results. It turns out that it's not critical--so long as Joe Inglett, with his .325 wOBA is available as a replacement if Keppinger is traded.
I assumed that Carlos Lee will get 25% of the playing time at 1st base, with Wallace getting the bulk of the remainder. I don't think the results are very sensitive to this assumption, though. I have pegged Wallace as a +5 fielder at 1st base, and assumed that Lee is an average fielder at 1st base. As for Lee's defense in LF, I put him at -10, which assumes some regression to mean. Chris Johnson is the second worst defender, behind Lee. Keppinger and Sanchez are also pegged as negative defenders. Michael Bourn's high level defense in CF is expected to continue. Barmes is assumed to be the team's second best defender at +6. Downs, Quintero, Hall, and Borgeois are assigned slight positive results on defense. Overall, my results for the team are +1.8 on defense, and -0.2 in baserunning.
At the last minute, I replaced Rowland-Smith's starter innings with Aneury Rodriguez. However, the difference in the two pitcher's ERA doesn't cause a significant difference in the W/L record.
My results indicate that the probability of 70 wins or better is 87%. The probability of a winning record (82 wins) is 22%. And if you think that 89 wins will be required to get in the playoffs, the Astros' probability is 3%.
The ERA and WAR is shown for each pitcher below.
|
Pitchers |
ERA |
WAR |
|
Myers |
3.90 |
3.35 |
|
W-Rodriguez |
3.65 |
4.00 |
|
Norris |
4.25 |
2.11 |
|
Happ |
4.05 |
2.44 |
|
Figueroa |
4.10 |
1.62 |
|
A-Rodriguez |
5.00 |
0.3 |
|
Lyles |
3.75 |
1.15 |
|
Bullpen: |
||
|
Lopez |
3.50 |
1.23 |
|
Lyon |
3.50 |
1.45 |
|
Melancon |
3.90 |
0.51 |
|
Fulchino |
4.45 |
0.03 |
|
Abad |
3.73 |
0.28 |
|
Del Rosario |
4.00 |
0.24 |
|
Villar/A-Rod |
4.50 |
-0.01 |
|
Other LHP |
4.80 |
-0.09 |
|
Total Pitching |
4.04 |
18.6 |
The wOBA assumptions and ultimate WAR is shown for each position player below.
|
Hitter |
wOBA |
WAR |
|
Quintero |
0.28 |
-0.03 |
|
Towles |
0.3 |
0.32 |
|
Corporan |
0.315 |
0.17 |
|
Wallace |
0.32 |
0.15 |
|
Lee |
0.349 |
0.56 |
|
Keppinger |
0.32 |
0.30 |
|
Hall |
0.318 |
0.99 |
|
Sanchez |
0.299 |
0.02 |
|
Downs |
0.306 |
0.2 |
|
Barmes |
0.306 |
1.08 |
|
Johnson |
0.341 |
1.30 |
|
Bourgeois |
0.299 |
0.05 |
|
Bourn |
0.325 |
3.49 |
|
Pence |
0.35 |
2.25 |
|
Michaels |
0.315 |
0.08 |
|
Unspecified |
0.295 |
-0.02 |
|
Pitchers |
0.135 |
-4.89 |
|
Total |
6.0 |