2009 Week 4 "Upset" Review

With the conclusion of week 4 of the college football season, the press would like us to believe there were 7 upsets in the top-25.  Let's take a look at each one to see if each was truly an upset.


Then #4 in the polls Mississippi lost at South Carolina and while my computer didn't predict the upset, it did have Mississippi only #17 so considered the #4 rating a tad high and did pick South Carolina plus the points.  Computer one up on the polls.


Then #5 in the polls Penn State lost at home to Iowa and again the computer didn't pick the upset but had then #16 Penn State by only 3.7 so it again had then #20 Iowa plus the points and got it right.  Computer two up.


Then #6 in the polls Cal lost on the road, bad, to Oregon.  The computer had Cal #5 so shame on it, but their performance to that point in the season gave no indication they'd lay an egg like that.  But the computer thought enough of Oregon to still pick them plus the points so got it right.  But call this one a push between the computer and polls.


Then #9 in the polls Miami lost on the road to then #11 in the polls Virginia Tech which probably wasn't really an upset with the game being in Virginia.  But my computer still picked Miami although it again had the right side of the spread.  Since the polls don't take home field advantage into account, we'll actually give this one to the polls as the computer had Miami #7 and Virginia Tech #21, a bigger gap.  Computer only one up now.


Then #18 in the polls Florida State lost at home to South Florida, and again the computer didn't pick the upset but had South Florida plus the points.  Since the computer had Florida State #14 probably call this a push too.


Then #22 in the polls North Carolina lost to Georgia Tech on the road, and the computer did pick this upset having Georgia Tech by 0.5.  Computer up two.


Then #24 in the polls Washington lost on the road to Stanford and the computer also picked this upset and had Washington no where close to the top-25 at #51.  Computer up three.

So, and I am probably a bit biased but I think the data supports it, the polls are still suffering from early season bias and at least as of last week are rating certain teams higher based solely on tradition or name.  Yet another reason to give computer ratings a look.  Check out my computers ratings through week 4 here.