Originally Published: September 28, 2006

Making sense of predicted win totals

Predicted wins: Where did we go wrong, what did we get right? We were close on many, but missed wildly on a half-dozen teams, writes Rob Neyer.

Share
Neyer By Rob Neyer
ESPN Insider
Archive

With the 2006 standings nearly in the books, we thought it might be infotaining to compare what really happened to what we thought would happen, "we" being ESPN.com's staff of writers and editors. What to use for "final standings"? I used Baseball Prospectus' Postseason Odds page, which lists projected final standings.

One problem, though: There's a lot of "play" in the standings, due solely to luck. You might predict a team is going to win 85 games, and with just average luck it would have won 85 games. But the team got stuck with lousy luck, and won only 80 games. Were you wrong? Yes. And no. What we really want to know is how accurate our predictions were, fundamentally. And one thing we can check is another Baseball Prospectus tool, "third-order wins," or in this case third-order winning percentage; both are based on the underlying events that typically lead to runs.

To continue reading this article you must be an Insider