Here is a mid-season update on the baseball predictions posted March 11, 1999:
This posting is being made July 2, 1999.

The average number of games played per team is currently 77-1/3 right now.
To the best of my knowledge, even if the probability of each team winning each
game were exactly correct, the expected error at this point in the season should
be 3.51 games (wins) per team on average. My current error is 3.78 games per 
team (so I am reasonably close).

At this time, we should expect about 20 teams (2/3 of all teams)
to be within one standard deviation of exact (4.4 games)  if the model
is good and that is exactly what I have. We should expect 95 per cent 
to be within 2 standard deviations (8.8 games) -- I have 93.3 percent.

Here are the results (error) for each team rounded to the nearest whole 
number -- i.e., number of games the model is off.

ERROR      No.     Teams
in         of
GAMES      Teams
0          0
1          4       Devil Rays, Mets, Expos, Padres
2          7       Royals, Rangers, Braves, Marlins, Astros, Cubs, Giants
3          7       Red Sox, Blue Jays, Tigers, A's, Angels, Brewers, D'backs
4          2       Yankees, Rockies
5          2       White Sox, Twins
6          3       Indians, Phillies, Reds
7          3       Mariners, Pirates, Cards
8          0
9          1       Dodgers
10         1       Orioles

Some comments: I think almost everyone predicted the Dodgers and Orioles 
to perform much better than they've been up to this point in the season.
The Yankees and Indians have fairly large errors but both are in first place, 
as predicted -- with the Indians doing better than expected and the 
Yankees worse.  The Twins are in last place, as predicted but are a better 
last place team -- so far -- than my prediction.