In case anyone was curious, this is the 6th version of Skynet... Skynet has only agreed with the humans ONCE (last year - Auburn vs. Oregon).
Here are the other years, if anyone was curious:
2010: Skynet and Humans: Auburn vs. Oregon
2009: Skynet: Bama vs. Cincy (got destroyed by Florida in a different game); BCS: Bama vs. Texas (got smoked by Bama)
2008: Skynet: Oklahoma vs. Texas; BCS: Oklahoma vs. Florida (won champ, despite Skynet snub)
2007: Skynet: LSU vs. Va. Tech; BCS: LSU vs. Ohio State
2006: Skynet: Ohio St vs. Michigan; BCS: Ohio St vs. Florida (won champ, despite Skynet snub)
What is most interesting is the narrow difference this year in the actual BCS rankings. The difference between Bama and Ok State was a mere .08 points (in the actual BCS). In Skynet this year Ok State finished ahead by 0.64 points. How does that rank relative to past years? Take a look at this table:
2009 - Cincy over Texas by 0.6
2008 - Texas over Florida by 1.02
2007 - Va. Tech over Ohio State by 1.6 (and over LSU - but this year was all screwed up with upsets)
2006 - Michigan over Floriday by only 0.03
Conclusion: It's a fairly big margin, but not that big in comparison to the limited data we have
If people were curious what the margins were in each of the computer polls (in terms of normalized values, in Skynet) here is the table:
But if anyone was curious, the two other computer rankings that are "biased" because they use human rankings at the beginning of the season for the initial rankings and are tainted, had normalized values of: Massey - Ok State - 1.740, Bama 1.988; Billingsley - Ok State 2.511, Bama 3.411. Amazingly, even in those rankings, Ok State was better than Bama!
But the thing that concerns me most about the computer polls? How in the world did Sagarin and Wolfe not end up with a greater change from Ok State beating Oklahoma? This is a strange question. Because the way the BCS is calculated, the computers only take 4 of the 6 computer rankings - throwing out the highest and lowest values. However, while the BCS takes into account every single vote, it merely just averages the values for the computer rankings (i.e. if you are 2nd in Sagarin, you get 24 points; if you are 3rd you get 23 points - no matter how close or far away the two teams are). So, let's play a BCS hypothetical: what happens if the BCS is still worth 1/3 but it uses a methodology to take into account normalized values? For this hypothetical I'll consider all 6 computer rankings:
|Total||Harris||Coaches||Computers - normalized|
Thus, it puts Ok State 0.002 closer to Bama. Not by as much as I thought it would actually! The more interesting part is how it would take very few ballots to be switched for Ok State to have made the big game. Once again, we realize that Skynet is supreme, because it's the only system where we can be certain that no one is paid off to switch their vote or keep their vote a certain way! That's it for this year. Skynet will have to remain angry once again!
Most underrated by the humans: Ok St, K STate, Oklahoma, Baylor, Georgia (The Big 12 was easily the best conference this year, unanimous by all the computers; the Big 12 won like 90% of its games out of conference)
Most overrated by the humans: Stanford, Oregon, Boise St, Wisconsin, Va. Tech, Mich St (wow, how things have changed. Boise State ended up on the overrated chart!)