Basketball Rankings Changes!

After looking at my systems projections for the NCAA tournament, I have made a quick check on the accuracy of the ratings and found some issues. Turns out, it is underrating the best teams, while overrating the worst. This requires just a slight shift in the model's constants so it's not a big issue, but it will cause a little delay in getting the percentages online. You'll notice also that the favorites chances will be higher than if you calculated them from the old rankings (12.8% for Duke was the highest). Is this the cleverest way to spend my finals week? No, but it sure beats writing history papers.