Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.1 6/24/83; site msdc.UUCP
Path: utzoo!watmath!clyde!cbosgd!gatech!gt-stratus!msdc!gpb
From: gpb@msdc.UUCP (George P. Burdell)
Newsgroups: net.sport.football
Subject: Re: George part of AP poll?
Message-ID: <144@msdc.UUCP>
Date: Fri, 4-Oct-85 10:06:45 EDT
Article-I.D.: msdc.144
Posted: Fri Oct  4 10:06:45 1985
Date-Received: Sat, 5-Oct-85 07:00:44 EDT
References: <140100001@uiucdcsb>
Distribution: net
Organization: George's Football Picks - Atlanta, GA
Lines: 63

>>                               But then I saw the AP poll this
>> week, and the SEVENTEENTH-ranked Terps got one FIRST place vote.
>>
>> So what gives?  Is George on the AP panel of sportswriters?
>> Is George THAT influential?  I'd like George to respond.
>>
>>                                 Stephen L. Borodkin

Stephen (and the rest of the net):
    No, George is not THAT influenetial.  We only wish that we get
paid the way thoses sportswriters do for doing this stuff.  Although,
we'd like to use this as a way of getting rid of most of our local
sportscasters and writers.  But that's another story.

I imagine that the sportswriter you mentioned in your article
uses the same algorithms for determining his college rankings that 
George uses.  The basic algorithm is: given two power ratings of
some teams, predict the outcome.  Now take the real outcome of
the game and adjust the power rating of the two teams.
Ex: Team A has a power rating of 113; Team B has one of 108.
    Teams A plays at Team B:  113 - (108 + 3 for home field)
             Team A is favored by two points.
    Now take some hypothetical results:
     1. A beats B by 10.  Real margin of victory is 8.  
            Add one point to A's power rating, subtract 1 from B's
     2. A beats B by 40.  Real margin is 38.
            Adjust each team by 4 points.
     3. A beats B by 1.  A actually loses due to the spread
            but because the gap is close, no adjustment is made
     4. B beats A by 10.  Actually margin is 12 in B's favor
            Add two points to B; subtract two from A

This is why we post George's spread results in each week's article.
0-4 point margin causes no change in the power ratings; 5-9 causes
one point difference; 10-14 = 2 points; 15-19 = 3 points; and 20-up
causes a 4 point difference.  Also, late in the season you do not
adjust the power rankings as much.  i.e. use 3 point adjustments
instead of 4.

It seems that the Vegas oddsmakers do something similar because
George is remarkably close with their lines.  Last year we did
predictions using George on college basketball games (but not on
the net) and he did very well.  This year we will post George for
basketball games (more on that in net.sport.hoops in the next
couple of weeks).

I'm glad to see some people out there are actually reading George's
picks and enjoying them.  We have fun working them out and we hope
you have some fun gleaming over his prognastications.  BTW, we
apologize for not getting George's picks to the net in time for
this week's games.  George's owners have had several projects due
this week and haven't had time to put the picks together.  However,
the next article will have his picks. For the record, they are
being posted Friday morning (not that any of them will be perfect
anyway).

Now to quote another expert football picker,
"Get me out of here, Percy"

George P. Burdell
George's Football Picks
Atlanta, Georgia
{gatech,ihnp4,mcnc,ncsu,akgua}!msdc!gpb