The Packard Power Ratings take into account previous Packard Power Ranking,
home field, game performance as seen by the program, previous oponents' game
performance as seen by the program, 
... previous oponents' previous oponents' previous oponents' previous oponents'...
game performace.

Basically if Team A has a "connection" to Team B and Team B does well then Team A
will go up although less than Team B.  Most any two teams have a "connection", 
however, this "connection" is usually so weak that the two teams affect each other
less than .01, which is how far I print out my power rankings.

Suppose you are a Texas Tech fan and want them to rise in the Packard Power
Rankings.  Then you should hope Texas Tech's oponents do well and their oponents
oponents do well etc.  But the biggest way for them to move up is to do better than 
the program expects them to do.

My rankings are based only on performance and not on expectation.  Although I like 
to think that my program will do a half way decent job of predicting. 

Teams move up if they do better than the computer expected and get credit for winning.
If a team is supposed to win by 7 but only wins by 4 may still move up since they won
and were put at a significant risk of loosing.  A team supposed to loose by 7 and wins
by 4 will move up alot.  A team supposed to win by 25 and wins by 40 will move up a
little.  I.e., there are dimishing returns, and you get rewarded for winning games
especially with oponents of near equal power and really rewarded when you pull an big
upset.  One more example if you are supposed to lose by 40 and lose by only 10 you will
go up a good amount.

Preseason rankings are simply last years final rankings.

----------------------------------------------------------------------------------------

Attached is a CHART that gives the change in power ranking based on different situations.

Examples of how the chart can be used:

Example 1: You see two teams that have a diffence in power rating of 15 and the higher 
rated team is at home.  Suppose this is a college basketball game.  Add 4.5 points to the 
15 to get 19.5.  (For HS Basketball it's also 4.5, for NBA it's 4, for College and HS 
football it's 3 and it's 2.5 for NBA)  Going to the table we see that the 19.5 difference
yeilds an expected margin of victory is 15 points.  Also from the table we see that for the
team to break even they need to win by about 9 points.

Example 2: Suppose your favorite team after adding any homefield advantage is 7 points 
lower in power rating than their oponent, yet they win by 3 points.  How much will they 
go up?  Approximating 7.6 for 7 and going to the row for loosing by 3 points we see the 
opponent will go down about .77, hence your team will go up by the same .77.

Example 3: Suppose your team is supposed to win by 3 points.  You can see there is not 
an approximate break-even point.  This is because if the team wins even by 1 they go up, 
and if they loose they will go down.

Example 4: Suppose your team is supposed to win by 10 points.  From the chart we see that
winning by 40 and loosing by 1 point are rated approximately opposite.  That means in 
this case that doing 11 points worse than expected and 30 points better are rated about
the same in opposite directions.