Monday, May 04, 2009

Predictions Revisited

I'm planning something cool (read math related and not really cool for people who don't do math in their spare time) for this week regarding end of season results based on April records, but that's actual content and takes time. In the mean time, enjoy my terrible, terrible predictions from before the season.

AL East (Average Spots Off = 2)
Preseason Picks 5/4 Morning
Yankees Blue Jays
Rays Red Sox
Red Sox Yankees
Orioles Rays
Blue Jays Orioles

How bad are my predictions looking? Well, I thought that Yankees, Red Sox, and Rays would be battling it out. However, the Blue Jays are one of two teams that are in the opposite position of my prediction (last to first or first to last). Of the 6 divisions, my average spots off here were the second highest only to:

AL Central (Average Spots Off = 2.4)
Preseason Picks 5/4 Morning
Indians Royals
Twins Tigers
Tigers White Sox
Royals Twins
White Sox Indians

The Indians were my other complete mis-pick, but hey, at least I knew it was a homer call. What I really didn't expect was for the non-Indians in the central to all be within 2 games of each other.

AL West (Average Spots Off = 1)
Preseason Picks 5/4 Morning
Angels Mariners
Mariners Rangers
Rangers Angels
Athletics Athletics

The AL West was the best by my made up, completely meaningless metric of average spots off. I believe this is at least partially affect by there only being 4 spots, so less possible deviation. Basically, the Angels are underperforming to my expectations, but the rest of the order remains the same.

NL East (Average Spots Off = 1.2)
Preseason Picks 5/4 Morning
Mets Marlins
Phillies Phillies
Braves Braves
Marlins Mets
Nationals Nationals

Despite my made up metric (mmmmmmmm), the NL East looks the best for me, as 3 of 5 are in the right position with the Mets and Marlins just flipping spots. I'll take it.

NL Central (Average Spots Off = 1.7)
Preseason Picks 5/4 Morning
Cubs Cardinals
Brewers Reds
Reds Cubs
Cardinals Brewers
Astros Pirates
Pirates Astros

Opposite issue of the AL West where the number is probably inflated because of the presence of 6 teams instead of the normal 5. The only real surprise here are the Cardinals strutting (though I suppose the Cubs underperforming is the second part of that).

NL West (Average Spots Off = 1.2)
Preseason Picks 5/4 Morning
Diamondbacks Dodgers
Dodgers Giants
Giants Diamondbacks
Rockies Padres
Padres Rockies

No one cares about the NL West.

In summation: Overall, I was an average of 1.6 spots off, which sounds small except being off by 1.6 of 5 means I was closer to being 2 spots off than correctly guessing or being 1 spot off.

# of spots off # of teams
0 4
1 12
2 8
3 4
4 2

If I randomly selected each teams placement, while ignoring every other teams placement, I would have a 20% chance of guessing the placement correctly. To this point, I have correctly predicted 4 of 30 teams, or .13%. Not good. However, if you include the teams that are 1 spot off (and I do because it makes me feel better) then I predicted 16 of 30, 53%. Not bad.

Assuming there could be ties (which is possible, sort of), then I could randomly get the place of every team in a division 0.032% of the time. The math here doesn't really work, but I'll leave it to someone else to point out why.

No comments: