Season
Résumé Rankings After Week 7
1. Alabama 6-0 (3rd)
2. Florida 6-0 (5th)
3. South Carolina 6-1
(1st)
4. Notre Dame 6-0 (2nd)
5. Ohio State 7-0 (4th)
6. Oregon State 5-0 (NR)
7. Kansas State 6-0 (10th)
8. Oregon 6-0 (6th)
9. USC 5-1 (NR)
10. Texas Tech 5-1 (NR)
Out:
West Virginia (7th); Louisiana Tech (8th); Rutgers (9th).
Comments:
Although 12 teams remained unbeaten through the weekend, I stuck with my plan
of making 1-loss teams eligible for these rankings this week. The 23 teams with
only 1 loss increased the number of eligible teams to 35.
While the requirements
for eligibility changed, I kept the formula the same this week. It took a while
but eventually I came up with this week’s rankings. It was interesting to see
how the 1-loss teams would fair against the unbeatens. But by the time I
finished I realized that there were still some fairly obvious flaws in my
formula which I will have to tweak once again next week.
The first step this
week was to place the opponents of the 35 eligible teams on my strength of
opponent scale. Once again, I realized that there is a need for yet another
level on the scale. It’s not that hard to figure out the teams that are at the
two extremes (poor teams and great teams). It’s in between those two extremes
that it gets tricky trying to determine whether a team should be labeled as “decent”
or “average.” The same goes for trying to decide whether a team should be
considered “average” or “good.” Since there is a 1-point difference between
each level it ends up being rather important.
Eventually I settled on
a grade for each opponent. Losses were also being graded this week, and with
each part of my formula (strength of opponent, location, and margin) I simply
applied the opposite of what it would be for a win. For example, a win over a
poor team would be worth zero points, while a win over a great team would be
worth 4 points. Therefore, a loss to a poor team would be -4 points, while a
loss to a great team would be zero points. I did the same thing with margin and
location.
While creating this
week’s rankings I decided that I would also need to do some tweaking to the
other two parts of the formula as well. While playing on the road is certainly
a major factor in college football that needs to be taken into account in any
sort of ranking system, my formula at the moment doesn’t recognize that not all
road wins are the same. My formula cannot differentiate between winning at LSU
and winning at FAU.
Obviously the strength
of opponent part of the formula will recognize the difference in a way, as a
team would receive 4 points for beating LSU and zero points for beating FAU.
The problem is that an automatic 2-points for winning anywhere on the road is
too much. I realize now that my way of coming up with the location scale was
flawed to begin with.
It seems logical that a
win on a neutral field would fall exactly in between a win at home and a win on
the road, since you are assuming that generally speaking you would have all the
fan support at home, no fan support on the road, and a 50-50 split at a neutral
site. However, my intervals probably need to be smaller.
And to be honest, a win
on the road should maybe even be more than 50% more than a neutral field win. I
may even have to introduce something that negates the road win bonus when it
comes against a very weak team. In other words, I’ll have to consider strength
of opponent when factoring in how much credit I dole out for location.
I came to a similar
conclusion this week regarding the margin of victory/loss scale. I had already
downsized this scale in previous weeks, going up by a half point instead of a
full point. Also, I had already put a cap on the scale at 25+ points so that
teams calling off the dogs when the game was out of reach wouldn’t be punished,
and so that a team with a 70-point win over one hapless opponent wouldn’t be overly
rewarded. But again, a 40-point win over Army probably shouldn’t give out a
bonus of 1.5 points more than a 7-point win over Oklahoma. It may be that I’ll
even have to factor in strength of opponent when assigning scores for margin of
victory/loss.
The real problem
actually came about whenever there was a combination of weak opponent, on the road,
and big margin of victory. That was when it became clear that I had to fix
something. With the current formula, a 7-point home win over Alabama would be
worth 4 points, while a 27-point road victory over UMass would be worth 3.5
points. That simply doesn’t make sense. So unfortunately, I think I’m much
farther away from creating a truly relevant formula than I felt I was heading
into this past week.
While the system isn’t
perfect, I still think this week’s rankings are interesting. Amazingly, all 10
spots in my rankings experienced change this week. I switched out 3 teams. 4
teams moved down in my rankings while still remaining in the top 10, with 3 of
those teams moving down at least 2 spots. 3 teams also moved up in the rankings
and all 3 moved up at least 2 spots.
West Virginia,
Louisiana Tech, and Rutgers dropped out of the rankings this week from #7, #8,
and #9 respectively. Oregon State, USC, and Texas Tech all jumped into the top
10 at #6, #9, and #10 respectively.
Florida and Kansas
State made the biggest moves up the rankings this week, both climbing 3 spots.
Florida moved up 3 spots from #5 to #2, while KSU moved up 3 places from #10 to
#7.
No team fell more than 2 spots while remaining
in the top 10 this week. South Carolina dropped out of the top spot to #3,
while Notre Dame moved down 2 spots from #2 to #4, and Oregon fell 2 spots from
#6 to #8.
Alabama moved up 2
spots from #3 to retake the #1 spot in these rankings. Ohio State was the only
7-0 team in the rankings, coming in at #5. South Carolina was the highest
ranked team with a loss, earning the #3 spot. In total, 7 of the 12 undefeated
teams made the top 10, while 3 of the 23 teams with 1 loss made the rankings.
No comments:
Post a Comment