The "luck" column in Kenpom is one potentially problematic data point for us:
[source]
The problem is, we had a ton of turnover so our preseason rank was very low. Because it was low to begin the season, it will stay lower than it might (should?) have due to our "luck" factor. His algorithm doesn't assume that his "expected record" might have been wrong; instead it assumes that it was right all along and that you are just over-performing. So we will "tend to be rated lower" than "our record would suggest".
Our luck factor is the 8th highest in the country at +0.157.
It's quite possible that we may not be that lucky. Instead, we might actually be as good as our record indicates (meaning our luck factor should be zero), and he might have just had us WAY TOO LOW in the preseason rankings.
The easiest one to understand is Luck, which is the deviation in winning percentage between a team’s actual record and their expected record using the correlated gaussian method. The luck factor has nothing to do with the rating calculation, but a team that is very lucky (positive numbers) will tend to be rated lower by my system than their record would suggest.
The problem is, we had a ton of turnover so our preseason rank was very low. Because it was low to begin the season, it will stay lower than it might (should?) have due to our "luck" factor. His algorithm doesn't assume that his "expected record" might have been wrong; instead it assumes that it was right all along and that you are just over-performing. So we will "tend to be rated lower" than "our record would suggest".
Our luck factor is the 8th highest in the country at +0.157.
It's quite possible that we may not be that lucky. Instead, we might actually be as good as our record indicates (meaning our luck factor should be zero), and he might have just had us WAY TOO LOW in the preseason rankings.
Comment