The Simplicity of the Game:
Bill James Analyzes the 2005 ALOOO Season
Bill James, the Father of Baseball
Statistics, has devised a mathematical formula to predict how many games a
team should win in a given year. This formula is based on the
simplest element of baseball: to win games, a team must score more
runs than they allow. As simple as this sounds, this formula is
amazingly accurate. This article looks at using this formula to
analyze the 2005 ALOOO season to see which teams overproduced and
underproduced wins.
The formula used to predict wins is as
follows (RF=runs for, RA=runs against):
Predicted Win % = (RF^2)/(RF^2+RA^2)
for the mathematically challenged, this is runs for
squared divided by the total of runs for squared and runs allowed squared
Predicted win total = Predicted
Win % x 168 games
Here are the results, by
division:
| ALN |
Runs
For
|
Runs
Against
|
Actual
Win % |
Calculated
Win % |
Actual
Wins |
Calculated
Wins |
+/-
|
| PKL |
1027 |
966 |
0.542 |
0.531 |
91 |
89 |
2 |
| BUM |
852 |
909 |
0.435 |
0.468 |
73 |
79 |
-5 |
| COL |
859 |
990 |
0.435 |
0.430 |
73 |
72 |
1 |
| SAM |
826 |
1157 |
0.345 |
0.338 |
58 |
57 |
1 |
| ALS |
|
|
|
|
|
|
|
| PBT |
1117 |
707 |
0.726 |
0.714 |
122 |
120 |
2 |
| MON |
1212 |
1061 |
0.649 |
0.566 |
109 |
95 |
14 |
| FAT |
821 |
1017 |
0.417 |
0.395 |
70 |
66 |
4 |
| HUN |
921 |
1063 |
0.411 |
0.429 |
69 |
72 |
-3 |
| NLE |
|
|
|
|
|
|
|
| CS |
1247 |
1078 |
0.583 |
0.572 |
98 |
96 |
2 |
| CHI |
1137 |
1117 |
0.482 |
0.509 |
81 |
85 |
-5 |
| DOG |
977 |
1062 |
0.464 |
0.458 |
78 |
77 |
1 |
| 8M |
892 |
1111 |
0.369 |
0.255 |
62 |
66 |
-4 |
| NLW |
|
|
|
|
|
|
|
| GB |
1114 |
865 |
0.595 |
0.624 |
100 |
105 |
-5 |
| QGR |
1063 |
907 |
0.589 |
0.579 |
99 |
97 |
2 |
| PIT |
903 |
906 |
0.482 |
0.498 |
81 |
84 |
-3 |
| DUK |
964 |
1016 |
0.476 |
0.474 |
80 |
80 |
0 |
Summary:
Bill James average predicted win
percentage for the ALOOO league was right on the money. The average
in the wins over predicted/wins under predicted column is zero.
This is incredibly accurate!
The team that over performed the
most was the Montreal Shipwreckers who won 14 more games than James
predicted based on the runs for/runs against formula.
How did Vince get his team to win 14 more games than Bill James
predicted? It may be a result of the Washington Nationals'
phenomenon. Vince managed his team to an incredible 34 - 12 record
in 1-run games in 2005. As you may recall, the Nationals in 2005
were in first place in the NL East for quite some time, despite the fact
that they had scored fewer runs than they allowed. Baseball experts
correctly predicted that the lack of offense would catch up to the Nats
and the numbers would win over the long haul as the James formula
suggests, as a team cannot continue to win despite being outscored.
With the Shipwreckers great record in 1-run games, the formula gets skewed
a bit as a 1-run win can offset a large loss. While Montreal did
score more runs than they allowed, it was an amazing feat that they won
109 games with such a small run differential.
The rest of the league fell very
much in line with the James predictions. The most under performing
teams, all at 5 fewer wins than predicted, were Da Bums, the Chicago
Champions, and the Green Bay Glory.
If you haven't fallen asleep yet,
congratulations! League comment regarding this article is
welcome. All league members are encouraged to contribute to the
league web page. Send ideas and submissions to David.
Take me back home!
|