Many of you are already aware of the fact that I have, on more
then one occasion, voiced my criticism of the RPI. About three
years ago I wrote a column on the subject and some modifications
have been made to the formula since.
However, along with many of my fellow coaches, I believed some
things were still not being weighted as much as they probably
should be. Beginning this past fall, with the help of the staff
of collegeinsider.com, I corresponded with numerous coaches to
get their input. The result is our version of the Ratings Power
Index.
As most of you know, the RPI consists of three main
calculations.
1) Team's own winning percentage (25%)
2) Team's direct opponents winning percentage (50%)
3) Team's opponents' opponents winning percentage (25%)
The formula weighs neutral court wins, home wins and road wins,
with the highest value being put on a road victory. The winning
percentage is then determined by the number of wins divided by
total of wins and losses.
While the calculation has been improved upon, many coaches still
believe that it overlooks certain criteria. Thus, the newly
created Coaches RPI takes into consideration the following.
1) Closely contested losses. There should be a slight value put
on such games. The idea being, from the outset of a contest, you
want to be in a position to win in the final few minutes. While
closely contested games don’t count in the win-loss column, they
are ordinarily and indication of a better performance that a
lopsided loss.
2) Impressive road wins. Winning in college basketball is not
easy, regardless of venue. But winning on the road is a greater
achievement then some may understand. The numbers prove that the
home team does win the majority of the time. While the RPI does
consider road wins, it doesn’t calculate impressive road wins.
This point should help to separate some from the pack. Dominant
teams will win impressively on the road so this gives them an
advantage over the team that wins on the road by a slim margin.
It doesn’t diminish a road win, but rather rewards the
impressive ones.
This should also, in theory, help dominant teams in leagues with
teams that are at the low end of the RPI. It’s not uncommon for
a team to dominate a mid-major conference, but despite winning
convincingly they still drop in the RPI because of their
opponents. This calculation can help to offset that to a degree.
3) Conference road wins. Again this would further benefit the
dominant teams in smaller conferences, as well as the
high-profile leagues. Combined with an impressive road win, a
team can improve two parts of the calculation on a given night,
should they post an impressive conference road win.
Again, this helps to provide a slight advantage. All wins are
equal in the standings, but all wins are not equal in reality.
As difficult as it is to win on the road it is even more
challenging to win on the road in your respective conference.
And these challenges are relative.
In the ACC winning at North Carolina is comparable to winning at
Old Dominion in the CAA. It favors the achievement and really
rewards teams that consistently win on the road in conference. A
couple of years ago Davidson went unbeaten in Southern
Conference play (16-0). Combine a perfect road mark in
conference with some impressive road wins and this calculation
would have improved their profile.
Based on our three additions to the calculation, it would stand
to reason that the difference in the Coaches RPI and the
traditional RPI would be most evident through January and
February.
Whether our ratings system is viable is yet to be determined. We
are essentially using this season as a trial and error period.
At season’s end we can more closely examine the numbers as they
relate to the team’s perceived evaluation. It is very possible
that the calculations can be changed prior to the start of the
2007-08 season.
Our goal here is simply to offer another option. As human beings
we often complain about things, but rarely offer alternative
solutions. Since many of us think it’s broken, we are trying to
fix it.
