Josh Kirshner writes:
>   How much has it "hurt" BG by picking up two games against OSU.  It would
>be shame if BG somehow loses a bid based on these rankings beacause of
>the low Opponent Win%.  Also, would it be bnetter in the eyes of the NC$$
>if BG would have left open spots in their season?
 
This is an interesting question.  First, some background for those who
don't know what is being discussed:
 
After last season, Kent dropped out of the CCHA, leaving each CCHA
team with 3 extra available games (CCHA league schedule dropped from
30 games last year to 27 this year; it will be 30 again next year when
UAF joins).  BG ended up adding 2 nonconference games with fellow CCHA
member Ohio State, giving them a total of 5 games with OSU this year.
Note that they could have chosen to play fewer than the limit this
season; they did not have to fill the extra games with anyone.
 
In those two games, BG and OSU split.
 
I decided to remove the two games and run the RPI to see what would
happen to BG.  The results are below.  Since one of the things that is
important about the RPI is what teams are within the .01 (or below,
1.00) threshold of you, because those teams can be compared with you
in other factors, I included all teams within that threshold of BG too.
 
Including BG-OSU NC games
RPI                                             Opp         OppOpp
RK                     G  W  L  T   Win% RK     Win% RK     Win% RK      RPI
10 Northeastern       33 15-13- 5  53.03 19    54.59 2     50.76 13    53.24
11 Denver             34 21-12- 1  63.24 9     48.84 33    52.01 1     53.23
12 BOWLING GREEN      32 20-10- 2  65.62 6     46.97 41    49.83 19    52.35
13 Princeton          25 13- 9- 3  58.00 12    50.53 15    49.24 35    52.08
14 Lake Superior      32 16-10- 6  59.38 11    49.13 28    49.95 17    51.89
15 Brown              25 13- 9- 3  58.00 13    49.41 27    48.92 39    51.44
16 Vermont            30 16-12- 2  56.67 14    49.65 21    49.57 29    51.38
 
Not including BG-OSU NC games
RPI                                             Opp         OppOpp
RK                     G  W  L  T   Win% RK     Win% RK     Win% RK      RPI
7  Michigan State     32 19-10- 3  64.06 8     50.96 14    49.71 22    53.93
8  Clarkson           29 17- 8- 4  65.52 7     49.50 23    49.66 26    53.54
9  BOWLING GREEN      30 19- 9- 2  66.67 6     48.82 32    49.80 19    53.52
10 Wisconsin          34 17-13- 4  55.88 15    53.27 7     51.47 3     53.47
11 Northeastern       33 15-13- 5  53.03 19    54.62 2     50.81 13    53.27
12 Denver             34 21-12- 1  63.24 9     48.78 35    51.98 1     53.19
 
It certainly hurt BG a bit that their Win% dropped by losing one of
the two NC games to OSU.  But as you can see above, their Win% rank
remained 6th even without the games.
 
The big difference is in their OppWin%.  When the games are removed,
BG's OppWin% rank changes from 41st to 32nd, helping them move from
12th to 9th in the RPI.
 
But what really matters is BG's possible range.  I determine this by
figuring out all of the teams that are within the threshold of BG.
When the games are included, BG's range is 10-16.  Without the games,
it is 7-12.
 
So to me, it appears that BG's chances at a bid would have been helped
considerably if they had chosen not to play any games rather than to
play OSU twice.  I did not run the RPI to see where BG would be if
they had won both games with OSU, but I am guessing that the change
would not have been as significant as if they had simply not played at
all.
---                                                                   ---
Mike Machnik                                            [log in to unmask]
Cabletron Systems, Inc.                                    *HMM* 11/13/93