Football: How did the simulations do?

And you thought we were done with football for the season. Think again!

Remember this, the predicted 2009 football standings based on a crude simulation of the season done one million times?

Some of the predictions based on that post were obvious: Laval to finish first in the Q, and so on. But let's look at everyone's results.

Below, the number after each team is the percentage of times the sim said they would finish in that spot in the standings. The overall score for each conference is the sum of the percentages divided by the number of teams in that conference, and, if it's not substantially more than the baseline number (100 divided by the conference size), that means I was really, really bad at this.

AUS
1. Saint Mary's (72.2%)
2. St. Francis Xavier (45.6%)
3. Acadia (20.0%)
4. Mount Allison (8.8%)
Overall score: 36.7 (baseline of 25), 2/4 correct

Mount Allison was worse than I thought; I had them finishing second a quarter of the time, never mind third.

QUFL
1. Laval (90.9%)
2. Montreal (27.2%)
3. Bishop's (13.9%)
4. Concordia (20.6%)
5. McGill (27.4%)
6. Sherbrooke (0.7%)
Overall score: 30.1 (baseline of 16.7), 2/6 correct

In the Q, I only had two teams right, but you'll recall how crazy those 3-6 seedings became.

OUA
1. Queen's (63.7%)
2. Laurier (19.7%)
3. Western (15.7%)
4. Ottawa (16.1%)
5. McMaster (16.6%)
6. Guelph (14.5%)
7. Waterloo (16.7%)
8. Windsor (10.4%)
9. Toronto (18.7%)
10. York (55.2%)
Overall score: 24.7 (baseline of 10), 6/10 correct

Top and bottom was easy to pick here. In the middle, it wasn't. While most others in the OUA were watching two quarterbacks chase a passing record during the final week of the season, those at Laurier (one of four 6-2 teams) were continually updating their tiebreaking situation.

Canada West
1. Saskatchewan (38.3%)
2. Calgary (24.3%)
3. Alberta (6.5%)
4. Regina (19.7%)
5. UBC (20.9%)
6. Manitoba (16.4%)
7. Simon Fraser (1.6%)
Overall score: 18.2 (baseline of 14.2), 1/7 correct

Yikes. Needless to say, the simulation wasn't built to predict the use of seven million ineligible players (give or take) in one season.


Putting only 11 teams in their correct positions is not great (although it's better than total random guessing), but the west messed that number up a bit (it's 10 of 20 without that conference). And I guess what this whole exercise means is that the OUA and AUS are the more predictable conferences. Which can be good or bad, depending on your view.
Next PostNewer Post Previous PostOlder Post Home

0 comments:

Post a Comment