Reply to thread

Message

[QUOTE="jncuse, post: 4917092, member: 1969"]
"luck" is not an input into the formula or his ranking, it is only an output.    It won't impact your adjusted EM which is what he ranks teams on.   "Luck" is an output of the formula, comparing your EM to your actual record.

That being said our predicted ranking may still have some "input" factor into our current ranking.   But not too much.  Our predicted record may still hurt us a tad... I had seen earlier from an article in 2011 that around Christmas that year KP's predicted ranking still weighed around 20% of your ranking, but kept on dropping down to zero over the next little while.   Since we started at 107, it probably hurts us around 5 or 6 spots right now by deduction (assuming it is a 20/80 split) but not more than that.

As others have mentioned "luck" is similar to the "pyth wins" concept in major sports like baseball and basketball.   I "Pyth" variance is easier to comprehend -- typically teams with high pyth variances from pyth wins to actual wins (good or bad) tend to be very good or very bad in close games, or have an abnormal amount of large blowouts (good or bad).    Its similar for the "luck" factor in the NCAA, but it doesn't work out very well early in the NCAA season because there are so many cupcake games that are not competitive -- its becomes more relavant to look at later in the season.    Typically if you are losing a bunch of close conference games, your "luck" will be deemed bad.    As I said I don't like the word luck because with certain teams in can be more of a skill arguably.   But analytics guys always view performance in close games as diverging back to the mean.
[/QUOTE]

Verification

What is a Syracuse fan's favorite color?