About the RPI rankings

[jimdahl]

I've received a couple questions about why the RPI published on USCHO differs from the RPI table on this site. Since the questions came so back-to-back, I assume USCHO just started publishing theirs, so it was a good time to look into it.

The most obvious difference is that it looks like USCHO isn't removing "negative wins."

e.g. They list Bemidji St as .6722

http://www.uscho.com/rankings/rpi.php

In the SiouxSports RPI details for Bemidji St, you can see that .6722 would be Bemidji St.'s RPI with all games counted, but after removing its wins over Air Force and Northern Michigan, their RPI rises to .6903.

http://siouxsports.com/hockey/rankings/rpi...s.php?teamid=24

Since these questions seem to come up every year, until everyone gets their calculations right, I've tried to provide a lot of supporting details so those inclined can see for themselves how SiouxSports.com is calculating RPI (and other ratings). However, it is sometimes still a challenge to determine the differences when comparing to just a ranking table like USCHO's and I have lots of practice, so definitely continue to ask questions if you think you see a problem.

[Goon]

I've received a couple questions about why the RPI published on USCHO differs from the RPI table on this site. Since the questions came so back-to-back, I assume USCHO just started publishing theirs, so it was a good time to look into it.

The most obvious difference is that it looks like USCHO isn't removing "negative wins."

e.g. They list Bemidji St as .6722

http://www.uscho.com/rankings/rpi.php

In the SiouxSports RPI details for Bemidji St, you can see that .6722 would be Bemidji St.'s RPI with all games counted, but after removing its wins over Air Force and Northern Michigan, their RPI rises to .6903.

http://siouxsports.com/hockey/rankings/rpi...s.php?teamid=24

Since these questions seem to come up every year, until everyone gets their calculations right, I've tried to provide a lot of supporting details so those inclined can see for themselves how SiouxSports.com is calculating RPI (and other ratings). However, it is sometimes still a challenge to determine the differences when comparing to just a ranking table like USCHO's and I have lots of practice, so definitely continue to ask questions if you think you see a problem.

I am glad that you understand all of this stuff but is there still an incentive to play tough teams and win?

Or is it better to just rack up wins against say some cupcakes?