March 19, 2007
Division Projections
John Beamer puts together his predictions for the division races at The Hardball Times. The most interesting thing about the standings is the National League is so bad. Arizona posts the best record in the senior circuit with an 86-76 mark. While I don't disagree there are no clearly dominant teams in the NL, just by chance one of these teams will win 90+ games. Run this a few times and see how often you get a 90 win team. (I get a 90 win team about 96% of the time.) By random chance, some team in the NL is going to much better than this projection.
I have a feeling THT won't want to be reminded of these projections at the end of the year.
Conversely, their lows in the NL aren't low enough. If the Nats manage to win 70 games this year, they should be popping the bubbly. They might be challenging the 1962 Mets in terms of futility.
David,
What John is looking at is true talent, not random chance. Of course, by random chance some team will win 90. Heck, it might be the Nationals! But random chance is, well, random, and it's nothing something you want to include in your projections.
I would concentrate on the predicted order of finish, which I think is pretty accurate.
Dave, I got less than 90 5 times out of 12...but only 5 out of 50 as well (38 in a row over).
This same concern comes up when player projections include a bunch of players who are projected to hit 38-40 home runs. Of course there's a good chance someone will hit 45 or 50 home runs, but it's difficult to predict which one of those players will be the one to do it.
Similarly, it's likely that one of the four NL teams projected to win 84-86 games will win 90+ games, but we're projecting the average numbers of wins for those teams will be around 85.