Kenpom #60 takes down Kenpom #4

[madtownwarrior]

ND takes down Pitt at Pitt...

remember, kenpom's justification article says Pitt is the most underrated, hence his awesome statistical model ranking them at #4

maybe time for another kenpom justification article as to why Notre Dame is over-rated  

[honkytonk]

Pomeroy tries hard. I guess there is a certain entertainment value to what he sells...but Im not sure it's anything beyond that.

[Galway Eagle]

I think there's a lot of merit in his calculations normally.  Though in a year where there is so much parity and inconsistency with the teams it makes Pomeroy's calculations much less accurate. 

[lab_warrior]

I'm pretty sure this would meet the textbook definition of cherry-picking.

You are all VERY welcome to generate your own sabermetrically based
ranking system if you want.     

Just an FYI, if you would have bet on all the higher ranked KenPom ranked
teams last night, you would have made a DUMP TRUCK full of money. 

In a related note, I know that Nate Silver got all the Electoral Vote
predictions right, but he didn't get that one Congressional race correct. 
CLEARLY HIS FIVETHIRTYEIGHT SITE IS GARBAGE.

[madtownwarrior]

interesting, so when it's obvious which teams are better, it makes Pomeroy calculations more accurate.

I think there's a lot of merit in his calculations normally.  Though in a year where there is so much parity and inconsistency with the teams it makes Pomeroy's calculations much less accurate. 

[MU82]

I'm pretty sure this would meet the textbook definition of cherry-picking.

You are all VERY welcome to generate your own sabermetrically based
ranking system if you want.     

Just an FYI, if you would have bet on all the higher ranked KenPom ranked
teams last night, you would have made a DUMP TRUCK full of money. 

In a related note, I know that Nate Silver got all the Electoral Vote
predictions right, but he didn't get that one Congressional race correct. 
CLEARLY HIS FIVETHIRTYEIGHT SITE IS GARBAGE.

If only kenpom was anywhere near as reliable as Nate Silver.

Remember: This system always hypervalues Wisconsin. He tries to explain it away, but all it does is hurt his credibility.

Others say that, outside the team rankings, he has lots of good stuff. I will take their word for it, but his rankings are unreliable.

[martyconlonontherun]

I'm pretty sure this would meet the textbook definition of cherry-picking.

You are all VERY welcome to generate your own sabermetrically based
ranking system if you want.     

Just an FYI, if you would have bet on all the higher ranked KenPom ranked
teams last night, you would have made a DUMP TRUCK full of money. 

In a related note, I know that Nate Silver got all the Electoral Vote
predictions right, but he didn't get that one Congressional race correct. 
CLEARLY HIS FIVETHIRTYEIGHT SITE IS GARBAGE.

1. I would assume part of the reason Nate Silver is more accurate is due to the resources that he has behind him. More people care about polls for elections that trying to make sports into statistics.

2. There is a huge difference in using statistical samples and metrics in different demographics of millions of people, then it is trying to compare a limited number of games across a wide-range of talent in college basketball.

[akmarq]

And if not cherry picking then it's a big old case of confirmation bias.

You have a pre-concieved notion of where PITT should be and so you dismiss his rankings on the basis of him disagreeing with you. Unless you can show WHY his rankings are flawed and show HOW your ranking system is better, then you're just confirming your own biases.

I'm not saying that he isn't wrong on PITT, but every model is going to bias certain data points/metrics. It's the nature of the beast. I'd much rather have a transparent methodology than the 'eye test' or the 'plays like a winner' metric. Remember that polls are also a model for rankings teams thave have their own (well-documented) bias.

Point is - you can disagree with him on PITT, but don't attack the whole model based on a SINGLE case of it not being true to life. It's a hell of a lot better than most other methodologies.

[Lennys Tap]

Aren't plodders like Pitt and Wisconsin routinely overrated by Pomeroy? Doesn't he admit there's a glitch he hasn't figured out? Pomeroy is good, but I think even he would say Caveat Emptor when it comes to the "slow ballers".

[madtownwarrior]

With that thrashing of Nerdwestern, Kenpom vaults Wisconsin up to #5.

For the record, their RPI is arpund 26.

Yep, I see the 8 loss Bucky team as statistically the 5th best team in the country.   

He needs to adjust his overall ratings or else they will continue to be dismissed as a bad joke.

[madtownwarrior]

interesting read on the teams ranked vastly different between RPI and KenPom

http://www.thebiglead.com/index.php/2013/02/05/the-rpi-vs-ken-pomeroy-a-look-at-which-likely-tournament-teams-are-valued-most-differently/

[MuMark]

No rating system can be perfect. Lots of different systems but when you look at consensus they have most of the teams that will end up as high seeds near the top.....Pomeroy is as good as any of them for picking games.

http://www.masseyratings.com/cb/compare.htm