I wish they would get rid of the last 10 games criteria. It's a really poor measure for selecting teams, because it's so dependent on strength of schedule more than anything else. It disadvantages teams in tough conferences, and benefits teams in sim-dominated leagues. I know they use it in real-life (actually in real life, they use last 12 games), but this is one of those instances where I think they should depart. 

Posted by professor17 on 8/1/2010 8:41:00 AM (view original):
I wish they would get rid of the last 10 games criteria. It's a really poor measure for selecting teams, because it's so dependent on strength of schedule more than anything else. It disadvantages teams in tough conferences, and benefits teams in sim-dominated leagues. I know they use it in real-life (actually in real life, they use last 12 games), but this is one of those instances where I think they should depart. 

Not only that, but it's used in real life because teams can really start playing well/gelling/clicking, whereas in HD it's just the same #'s/ratings. It's a senseless criteria for a computer game.

Not entirely true:  A human coach could make lineup/settings adjustments that make his team suddenly start doing much better.  An injured player could return.  Ratings and IQ improve over the season.

Something to keep in mind -- when comparing RPI the SIM likely uses the raw RPI number, not the RPI rank (which is what most people call RPI here).  Those numbers could be a lot closer so the RPI advantage might not look so great when viewed through that lens.

Is the information still available? I'd love to BPI it and see what happens.

I have Kansas by a nose...

W L Team BPI SOS
15 13 KANSAS 107.86 253.39
17 13 PURDUE 107.67 242.83

IRL, what we have been observing is that the biggest factor is record versus 1-25, 1-50, 1-75, 1-100 not RPI.  For the most part, RPI should take care of itself if you look at these two records.  RPI becomes secondary if you look at those records because it's already included in RPI and those games should carry more weight.  Conference records imho should only matter when comparing teams from the same conference, not teams from different conferences.  While their 1-100 records were the same (8-10).  Purdue's 1-50 record was better.  Do we know what their 1-25 record was?  Perhaps their conf. records made the difference.  To me that doesn't make sense because I'd guess Pudue's conference was much more difficult than Kansas'.  What were the conference RPIs?  How did B10 conference teams do h2h with the B12?

Rails, to answer your questions, final conference RPI (after the NT) were Big 12 .5725, Big 10 .5671.  Records against the Top 25 are Purdue 1-5, Kansas 0-8.  I don't have the time to do conference head to head right now, but I might have a chance later.

Posted by colonels19 on 8/2/2010 10:29:00 AM (view original):
I have Kansas by a nose...

W L Team BPI SOS
15 13 KANSAS 107.86 253.39
17 13 PURDUE 107.67 242.83

^

colonels - I'm not even arguing that Kansas wasn't the better team - they probably were.  And this was shown in your system.  But that's not what WIS uses so I'm not sure what it has to do with this. 

It seems like the common answer from this thread is that this makes no sense.  I guess what I'm trying to figure out is if the conference record is THAT important, which it sure seems like the answer is "yes."