This diary continues some analysis from the first win expectancy diary, which explored a way of assessing how much each player contributed to our team's past success. Roughly speaking, your contributions get more credit the more 'clutch' they are. A typical play counts about -.050 to +.050 or so, a run scoring play or big threat counts about +.100 or so, and real game breakers count +.150-.200.  Plays above +.300 are huge turnarounds, and plays getting above +.400 are pretty rare. They are very often the ones you remember all season. For instance:

The Top 6 A's Plays of the Year (through Aug 31)

(R: runs ahead (or behind) the batting team is; B: bases occupied with runners. E.g. B: 13 = runners at corners)

May 11, 2005. OAK v BOS. Top 9.2. R:-1. B:1. E. Byrnes homered to deep left, J. Clark scored (0.829, +0.775)

Jul 27, 2005. CLE v OAK. Bot 9.1. R:-1. B:12. J. Kendall singled to left, J. Payton scored, M. Scutaro to third, J. Kendall to second advancing on throw (0.826, +0.436)

Aug 11, 2005. ANA v OAK. Bot 7.2. R:-3. B:23. E. Chavez homered to deep right, M. Ellis and J. Kendall scored (0.542, +0.428)

Jul 17, 2005. TEX v OAK. Bot 9.1. R:-1. B:2. A. Melhuse doubled to deep right, M. Ellis scored (0.703, +0.425)

Aug 29, 2005. OAK v BAL. Top 6.2. R:-1. B:12. D. Johnson homered to deep right, B. Kielty and S. Hatteberg scored (0.742, +0.417)

May 30, 2005. TAM v OAK. Bot 9.2. R:-1. B:13. J. Kendall singled to left, C. Thomas scored, M. Scutaro to second (0.631, +0.411)

Good times, huh?  I refuse to post the worst plays of the year for us, but I will say that 7 of the top 8 plays in A's games were bad for us, and that 8th one was Byrnes's play above and I think you know how that ended. Basically, it has been an unusually heart-wrenching season.

Okay, here's more analysis.

Update [2005-9-6 13:54:14 by Apricot]: Note that the player WX sums should be taken with a grain of salt as they depend on how one counts errors... but they are still interesting.

First, here is the year's total WX contributions of every A's player, as of Aug 31 2005. The columns are name, total win expectancy contributed, number of plays involved and WX/play (probably a meaningless number). Recall that bad defense is punished, but good defense is not rewarded (due to my physical limits, not the limits of the method). So Kotsay loses out here... Crosby probably loses a little too.

All Season through Aug 31st             Jun 1st through Aug 31st

R. Harden*      2.880   490 0.006       E. Chavez       2.426   350 0.007
H. Street*      2.181   284 0.008       J. Blanton*     2.220   460 0.005
B. Kielty       1.792   406 0.004       D. Johnson      2.191   295 0.007
D. Johnson      1.751   317 0.006       H. Street*      2.062   150 0.014
B. Zito*        1.582   757 0.002       B. Zito*        1.719   500 0.003
E. Chavez       1.335   562 0.002       R. Harden*      1.647   317 0.005
J. Duchscherer* 1.263   278 0.005       E. Byrnes       1.127   68  0.017
J. Blanton*     1.114   687 0.002       N. Swisher      1.017   309 0.003
A. Melhuse      1.054   83  0.013       J. Duchscherer* 0.961   193 0.005
J. Payton       0.796   174 0.005       A. Melhuse      0.894   57  0.016
J. Kendall      0.640   551 0.001       J. Payton       0.796   174 0.005
N. Swisher      0.372   414 0.001       B. Kielty       0.748   273 0.003
K. Calero*      0.364   184 0.002       M. Ellis        0.381   224 0.002
M. Kotsay       0.345   521 0.001       K. Yabu*        0.184   96  0.002
E. Byrnes       0.275   217 0.001       D. Haren*       0.166   439 0.000
M. Watson       0.017   34  0.000       R. Colon*       0.159   21  0.008
R. Flores*      0.011   22  0.001       D. Haren        0.066   3   0.022
J. Garcia*      0.003   4   0.001       R. Flores*      0.011   22  0.001
B. Reames*      0.000   31  0.000       J. Garcia*      0.003   4   0.001
F. Bynum        0.000   1   0.000       B. Reames*      0.000   7   0.000
J. Clark        0.000   2   0.000       F. Bynum        0.000   1   0.000
J. Witasick*    -0.003  75  -0.000      K. Calero*      -0.003  142 -0.000
M. Ellis        -0.079  351 -0.000      J. Witasick*    -0.003  75  -0.000
S. Etherton*    -0.100  53  -0.002      J. Kendall      -0.011  361 -0.000
T. Harikkala*   -0.104  58  -0.002      C. Thomas       -0.019  2   -0.009
K. Yabu*        -0.125  235 -0.001      R. Rincon*      -0.027  73  -0.000
J. Kennedy*     -0.158  102 -0.002      T. Harikkala*   -0.091  23  -0.004
R. Rincon*      -0.215  132 -0.002      J. Kennedy*     -0.158  102 -0.002
C. Thomas       -0.488  55  -0.009      M. Watson       -0.175  15  -0.012
M. Scutaro      -0.596  328 -0.002      J. Cruz*        -0.184  20  -0.009
K. Saarloos*    -0.639  511 -0.001      K. Saarloos*    -0.208  322 -0.001
J. Cruz*        -0.680  155 -0.004      K. Ginter       -0.338  52  -0.006
E. Durazo       -0.701  156 -0.004      S. Hatteberg    -0.360  210 -0.002
B. Crosby       -0.835  341 -0.002      M. Scutaro      -0.413  161 -0.003
O. Dotel*       -0.838  73  -0.011      M. Kotsay       -0.418  304 -0.001
D. Haren*       -0.908  716 -0.001      B. Crosby       -0.999  332 -0.003
S. Hatteberg    -0.930  401 -0.002      R. Glynn*       -1.446  82  -0.018
R. Glynn*       -1.446  82  -0.018      
K. Ginter       -1.474  149 -0.010      

Rookie of the Year

Winexp says they should be ranked:

H. Street*      2.181   284 0.008
D. Johnson      1.751   317 0.006
J. Blanton*     1.114   687 0.002
N. Swisher      0.372   414 0.001

and it's not really close: Street is the ROY for us.

If you like to imagine the season started Jun 1 (not completely unreasonable for rookies working their way in), the standings look like:

J. Blanton*     2.220   460 0.005
D. Johnson      2.191   295 0.007
H. Street*      2.062   150 0.014
N. Swisher      1.017   309 0.003

Which is a three-way tie with Swish way behind. Probably what most people would say, since we've purged May (what?) and before from our brains.

Here's how they progressed over the year (running totals for the entire year).

Injuries

Win Exp is good at looking back and assessing past performance. There is no proof one way or the other whether it assesses forward performance. However, I can't help but think it does, given that clutch and non-clutch situations should even out over time, so Win Exp should approach actual future contributions for large samples.

So what does whole-season Win Exp tell us about the injury situation?

• Harden (2.880): huge blow to our MVP (by a long shot), especially since Kennedy (-0.158) has to replace him. Kennedy's start in LA rated +.019, which is pretty good for a 6th starter, but a real comedown from Harden who averaged +.152 per start.
  
• Kielty (1.792): underrated due to recent cooling off (he broke even in August) and kRaZy hAiR, but he's been responsible for a lot of game-changing hits and ranks #3 for the year.
  
• Zito (1.582) : don't need numbers to tell you he's huge, but the numbers agree.
  
• Kotsay (.345): probably underrated by winexp. It's also unclear how big a drop-off it is to Matt Watson as replacement.
  
• Crosby (-.835) has had a bad year in raw results. His injury results in either a tie or improvement from playing Scooter (-.596) at the plate.
• The New Guys

I can't help throwing in a graph of the new guys' performance.

It's too early to tell what it means, but Payton has continued to contribute, and Witasick had a great burst of play before matching it with a burst of crummy play.