Reply to thread

Message

<blockquote data-quote="Samuel S" data-source="post: 129411537" data-attributes="member: 1754664">It's FAR more realistic than the FPI. Models like that do much to illustrate the massive flaws in models that attempt to reduce team comparisons to quantifiable statistical calculations. To a limited extent, such statistical models can be useful in deciding among teams which your eyes tell you are close. Beyond that, it's pretty silly stuff, especially when many games remain to be played. To the extent the models are well constructed they still are of little worth until all the games are played. The FPI is obviously not well constructed. GIGO still holds true and it doesn't matter if you run 10,000 or 10,000,000 &quot;simulations&quot; with garbage.Not only is it patently absurd to think 4 of the 5 best teams are from the Big 12 (which also serves, obviously, to inflate the ratings of the other conference teams), that it has 3 loss USC at number 7 is laughable. Is beating Utah that impressive?</blockquote>

[QUOTE="Samuel S, post: 129411537, member: 1754664"] It's FAR more realistic than the FPI. Models like that do much to illustrate the massive flaws in models that attempt to reduce team comparisons to quantifiable statistical calculations. To a limited extent, such statistical models can be useful in deciding among teams which your eyes tell you are close. Beyond that, it's pretty silly stuff, especially when many games remain to be played. To the extent the models are well constructed they still are of little worth until all the games are played. The FPI is obviously not well constructed. GIGO still holds true and it doesn't matter if you run 10,000 or 10,000,000 "simulations" with garbage. Not only is it patently absurd to think 4 of the 5 best teams are from the Big 12 (which also serves, obviously, to inflate the ratings of the other conference teams), that it has 3 loss USC at number 7 is laughable. Is beating Utah that impressive? [/QUOTE]

Name