Is too much weight given to RPI ratings?

[Hanmudog]

Looking at the SEC I see Alabama with an RPI of 83 and a record of 17-8 (9-2). Kentucky has an RPI of 15 and a record of 18-7 (6-5). Tennessee has an RPI of 25 and a record of 16-10 (6-5).

I realize schedule strength is a big part of the RPI but Alabama has beaten both UK and Tennessee and performed much better in than those two teams in SEC play yet if the NCAA tourney started today Alabama could be left out while UT and UK are considered locks. Just seems like the formula needs to be tweaked somehow to give greater weight to head to head matchups.

 

[MadDawg.sixpack]

is that it places too much emphasis on early season wins/losses. And there is an inherent problems with teams ranked in the 75-200 range. I would argue there isn't a dime's difference between a team ranked #115 and one ranked #175, but playing the latter would hurt your RPI much worse than the former.

 

[Henry Kissinger]

i think so. it'd benefit us as a team in general if rpi was less important. we typically play better at the end of the year. you can put that on coaching, but that's a whole separate issue. i think msu has done a poor job giving consideration to rpi when scheduling. i like scheduling a top 25 non conference game each year, even if it's probably a loss. but i think the real killer for us is scheduling sub 200 teams.

i know it's good to make the entire season important, and i know the non-conference play early on is what you base conference strength on, but i wish conference play determined which teams made it to the tournament. i'm not saying this should be implemented at all because it'd be a ton of work, but i'd like the nonconference play to determine how many teams each conference gets into the ncaa tournament. then i'd like the standings from each conference to determine which teams get in (e.g. sec gets rated as 5th strongest conference, it gets 5 teams in. either take top 5 overall [could dissolve east/west to make it more fair] or take top 2 from each division and go to head to head for third place teams. if conference tournament winner is 4th place in east, the third place teams aren't considered).

small schools could suffer if some sort of major conference cartel is formed, but i think i'd be fun.

even though that'll never happen, i like the general idea: how many teams does each conference deserve? which teams were best in conference play at the end of the year?

 

[RebelBruiser]

Really, it's not a flawed formula so much as it's flawed in its use.

I could see the argument for weighting wins and losses more later in the season rather than early, maybe adding 0.2 wins/losses to the calculation for a team's last 10 games, but if you did that, teams from weak conferences would see a big boost in their ratings by playing cupcakes later and getting win totals inflated in that stretch.

The RPI rewards you most for playing teams with good records and playing them on the road. You can play a team with a low RPI, but if they have a good overall record against a really weak schedule, you're going to still receive more benefit than you might receive playing a team with a better RPI but a worse record.

 

[Hanmudog]

Really, it's not a flawed formula so much as it's flawed in its use.

I could see the argument for weighting wins and losses more later in the season rather than early, maybe adding 0.2 wins/losses to the calculation for a team's last 10 games, but if you did that, teams from weak conferences would see a big boost in their ratings by playing cupcakes later and getting win totals inflated in that stretch.

The RPI rewards you most for playing teams with good records and playing them on the road. You can play a team with a low RPI, but if they have a good overall record against a really weak schedule, you're going to still receive more benefit than you might receive playing a team with a better RPI but a worse record.

It seems like lately the RPI is the Holy Grail for the selection committee so I agree that it's use is flawed.

Who is the better team, Alabama or UK? Alabama or UT? Alabama is clearly better than either yet the selection committee will probably look at the 60+ spot difference in RPI and give UT and UK bids over Bama without thinking twice unless Bama wins the overall SEC title.

 

[dawgs.sixpack]

is that it places too much emphasis on early season wins/losses. And there is an inherent problems with teams ranked in the 75-200 range. I would argue there isn't a dime's difference between a team ranked #115 and one ranked #175, but playing the latter would hurt your RPI much worse than the former.

yeah i've said this for years, could apply to CFB and the BCS formula too. weight team by team for the top 25, then teams 26-50 get a flat value, 51-100 a lower flat value, 101-200 a lower flat value, and 200+ the lowest flat value. football you could do team by team for the top 25, 26-50 a flat value, 51-75 a flat value, and 76+ a flat value. no bcs birth or ncaa birth should turn on whether you played #85 or #101, a top tier team should beat both teams easily (football) or #210 or #276 (basketball).