I am going to answer this question 2 ways:
1. Back in 2003, my grad school statistics professor at Indiana helped me create a sports handicapping model that uses non-linear regression and a handful of variables to predict the score of sports contests (NCAA, NFL, NBA). I used this model for a few years, and made a decent amount in Vegas betting with it on the NCAA tournament, before abandoning it in 2006 as it takes a very long time to run every week. ANYWAY, I set up the model again this year, and while it generally takes 3-4 weeks of data to be accurate, I can get a good 'estimate' after 2 weeks. With all that said, the model favors KSU by 12 points in this game. So, answer 1 is 31-18 KSU
2. As mentioned, #1 relies on data - the more it is given the more accurate its predictions are. 2 weeks is hardly enough to start betting paychecks in vegas on the predictions. My GUT says we are going to be an angry bunch on Saturday, and will come out firing early and often. It has been a really long time since we have POUNDED a good oponent - probably Texas A&M back in 2005 was the last one. Therefore, I think we get all the breaks, and Arnaud and Co play their ***** off to a 42-17 victory!
How is that for polar opposite predictions!