ACC Power Ratings - Games Through 1/22/98

ACC Power Ratings - Games Through 1/22/98

Post by George W. Harr » Sat, 24 Jan 1998 04:00:00


        What follows are power ratings for ACC men's
basketball teams based solely on the results of games
among conference members through Thursday, January
22th 1998.  

Team           Rating      Record       Previous
                       Conf    Total

1)Duke          86.08   6-0     17-1    1)86.44
2)UNC           77.63   5-1     19-1    2)77.58
3)Clemson       75.19   3-3     12-6    3)75.55
4)FSU           68.96   3-4     14-1    5)67.21
5)Maryland      67.68   4-3     11-6    4)67.94
6)GTech         62.91   2-4     12-6    7)63.19
7)Wake          61.63   2-4      9-7    6)64.33
8)NCState       60.87   1-5     10-7    9)60.81
9)***ia      60.68   2-4     10-9    8)61.27

Recent Games:

Thursday, 1/22/98
        Wake Forest at FSU
                Prediction: FSU by ~3
                Result:     FSU by 24
Predictions, immediate & cumulative: 1-0, 12-8

Upcoming Games:

Saturday, 1/24/98
        Duke at ***ia
                Prediction: Duke by ~17
        Clemson at Maryland
                Prediction: Clemson by ~1
        FSU at UNC
                Prediction: UNC by ~9

--
Real men don't need macho posturing to bolster their egos.

George W. Harris  For actual email address, replace each 'u' with an 'i'.

 
 
 

ACC Power Ratings - Games Through 1/22/98

Post by Donnie Barn » Sat, 24 Jan 1998 04:00:00

Quote:
>    What follows are power ratings for ACC men's
>basketball teams based solely on the results of games
>among conference members through Thursday, January
>22th 1998.  

>Team               Rating      Record       Previous
>                   Conf    Total

>1)Duke              86.08   6-0     17-1    1)86.44
>2)UNC               77.63   5-1     19-1    2)77.58

What does your algorithm predict for the upcoming Duke/UNC game
in CH if it were played now?

--Donnie

--

   Challenge Diversity.  Ignore People.  Live Life.  Use Linux.  879.
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
_Things You'd NEVER Expect A Southerner To Say_ by Vic Henley:    
**  I don't want a flannel shirt for Christmas.

 
 
 

ACC Power Ratings - Games Through 1/22/98

Post by George W. Harr » Sat, 24 Jan 1998 04:00:00


wrote thusly:

=>Team              Rating      Record       Previous
=>                  Conf    Total
=>
=>1)Duke             86.08   6-0     17-1    1)86.44
=>2)UNC              77.63   5-1     19-1    2)77.58
=
=What does your algorithm predict for the upcoming Duke/UNC game
=in CH if it were played now?

        Simple enough to calculate; adjust the home
team's power rating upwards by 4.5, the team with the
higher adjusted rating is predicted as the winner with a
predicted margin of four less (since I value a victory as +4
points) than the difference in the margins, or one if the
margin is less than 5, so

        (86.08-(77.63+4.5))=3.95, so Duke by 1, but really
too close to call (the algorithm's record for this type of game
is 3-4, so it's really a toss-up.  If the algorithm predicts a
margin of >1, then the record is 9-4.

=--Donnie

--
Real men don't need macho posturing to bolster their egos.

George W. Harris  For actual email address, replace each 'u' with an 'i'.

 
 
 

ACC Power Ratings - Games Through 1/22/98

Post by Donnie Barn » Sat, 24 Jan 1998 04:00:00

Quote:
>=What does your algorithm predict for the upcoming Duke/UNC game
>=in CH if it were played now?

>    Simple enough to calculate; adjust the home
>team's power rating upwards by 4.5, the team with the
>higher adjusted rating is predicted as the winner with a
>predicted margin of four less (since I value a victory as +4
>points) than the difference in the margins, or one if the
>margin is less than 5, so

>    (86.08-(77.63+4.5))=3.95, so Duke by 1, but really
>too close to call (the algorithm's record for this type of game
>is 3-4, so it's really a toss-up.  If the algorithm predicts a
>margin of >1, then the record is 9-4.

That's pretty damned good...are those records any better if you
throw out another set of games (I assume you couldn't and didn't
make predictions on any games where one or more teams had never
played)?  What if you throw out predictions made where one or
more teams had only played one game?  What would the record be
then in predictions?

--Donnie

--

   Challenge Diversity.  Ignore People.  Live Life.  Use Linux.  879.
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
_Things You'd NEVER Expect A Southerner To Say_ by Vic Henley:    
**  I don't want a flannel shirt for Christmas.

 
 
 

ACC Power Ratings - Games Through 1/22/98

Post by George W. Harr » Sun, 25 Jan 1998 04:00:00


wrote thusly:

=>   (86.08-(77.63+4.5))=3.95, so Duke by 1, but really
=>too close to call (the algorithm's record for this type of game
=>is 3-4), so it's really a toss-up.  If the algorithm predicts a
=>margin of >1, then the record is 9-4.
=
=That's pretty damned good...are those records any better if you
=throw out another set of games (I assume you couldn't and didn't
=make predictions on any games where one or more teams had never
=played)?  What if you throw out predictions made where one or
=more teams had only played one game?  What would the record be
=then in predictions?

        The algorithm cannot work unless one can trace a
path between each pair of teams, so the first rating I had was
after the minimum of eight games, when each team had
played two games except for Georgia Tech, which had lost
at home to NCState by two, and Wake Forest, which had
lost in Richmond to ***ia by six.  In those teams' s
ubsequent games the algorithm picked Tech to lose at UNC
by 32, and UNC won by 21, and Wake to lose to Maryland
by 14, but Wake won by six.  So, if you throw out *those* two
games, then in games where both teams have played at least
two games and there is a clear margin, the algorithm is 8-3.  
Of course, that's just picking the winner, rather than against a
spread.

=--Donnie

--
Real men don't need macho posturing to bolster their egos.

George W. Harris  For actual email address, replace each 'u' with an 'i'.