With the official BCS standings scheduled to be released on Sunday for the first time this season, it seems like an appropriate time to take a closer look at the major polls included in that system—providing both a context to make sense of them and a framework for interpreting them. Then, we will see that the apparent lack of love for the Cougars from the pollsters in recent weeks is not what it seems.
The Context
Rankings, by definition, are relative. A team’s ranking then, merely says that a team is supposedly better than all teams ranked below it and not as good as those ranked above it. A ranking does not quantify the difference between teams. For that, the total votes relative to other teams is the key, which provides a much more accurate assessment of how a team is viewed by voters in any given poll. The BCS standings, to their credit, only use the votes and not the ranking, in its calculations. Fans, then, should not be paying as much attention to a team’s rank when the polls come out on Sunday afternoon, as to the number of votes received and the change over the prior week.
The Framework
Since the votes are the more accurate assessment of how a team is perceived by voters in a given poll, here is an easy way to quickly assess the impact of votes:
1. Magic number is 4%—Each voter ranks 25 teams, and then “votes” are awarded based on that ranking with 25 points given to the top team, 24 to the second and so on. Not news to you. But what that means is that each rank is given 4% more points than the one after it (1/25=4%). Therefore, if a team gains 4% more votes than the previous week it has effectively moved up on average by one spot on voters’ ballots.
2. Total voters equals 4%—The number of first place votes given (60 for AP, 59 for USA Today, 114 for Harris, etc) will tell you how many ballots are submitted in each poll and represents how many votes are needed to get to 4%, or effectively “move up” one ranking spot. So if a team gains 90 votes in the AP poll (on 60 ballots), it effectively “moves up” 1.5 spots regardless of whether its actual ranking moves up or down, and is thus listed on the average ballot 1.5 spots higher than the previous week.
Application
With this context and framework in mind below is how the Effective AP Poll (as I am calling it) would look after week 6. The effective rank can also be thought of as the average rank given on voter’s ballots. When there are several teams clumped closely together in the final vote tally, it leads to the team in the front being ranked higher than its votes justify and the team in the back, being ranked lower than its votes justify. For example, TCU is ranked 12th, despite an average ballot placement of 10.72. And conversely, Penn State is ranked 14th, despite an average ballot placement of 16.05. To use an accounting analogy, TCU is sitting on some rankings receivables (an asset, having done the work without receiving payment yet) and Penn State is sitting on some unearned rankings (a liability, having received the payment and not yet done the work).
Effective
Rank Team Votes AP
1.17 Florida 1490 1
2.17 Alabama 1430 2
2.75 Texas 1395 3
4.62 Virginia Tech 1283 4
6.02 Boise State 1199 5
6.65 USC 1161 6
8.53 Ohio State 1048 7
8.70 Cincinnati 1038 8
9.83 Miami (FL) 970 9
10.22 LSU 947 10
10.68 Iowa 919 11
10.72 TCU 917 12
13.07 Oregon 776 13
16.05 Penn State 597 14
16.40 Nebraska 576 15
16.68 Oklahoma St 559 16
16.82 Kansas 551 17
17.83 BYU 490 18
18.45 Georgia Tech 453 19
18.80 Oklahoma 432 20
20.50 South Florida 330 21
20.68 So Carolina 319 22
22.80 Houston 192 23
24.73 Utah 76 24
24.75 Notre Dame 75 25
Impact on BYU
With that in mind, it is interesting to note how BYU has fared in the effective rankings. Despite seemingly going nowhere in the rankings that last several weeks, BYU has in fact been moving up the effective rankings and is sitting on some “rankings receivables” in the AP and USA Today polls, and a rankings liability in the Harris poll.
AP
Wk Opponent Rec Rk votes %votes gain effective
P ---- 0-0 20 267 17.8% --- 21.55
1 Oklahoma 1-0 9 984 65.6% 47.8% 9.60
2 Tulane 2-0 7 1122 74.8% 9.2% 7.30
3 Florida St 2-1 19 349 23.3% -51.5% 20.18
4 Colorado St 3-1 20 349 23.3% 0.0% 20.18
5 Utah St. 4-1 18 403 26.9% 3.6% 19.28
6 UNLV 5-1 18 490 32.7% 5.8% 17.83
AP—The rankings slip from #19 to #20 in week four after the CSU game, was relative only, as the team held steady at 20.18 in the effective rankings. This last week, BYU did not move up at all in the relative rankings, remaining at #18, however that is misleading since the team experienced a large jump in the votes (effective rankings) moving up 6% or roughly 1.5 spots.
USA Today/Coaches
Wk Opponent Record rk votes %votes gain effective
P ------ 0-0 24 178 12.1% --- 22.98
1 Oklahoma 1-0 12 755 51.2% 39.1% 13.20
2 Tulane 2-0 9 941 63.8% 12.6% 10.05
3 Florida St. 2-1 20 279 18.9% -44.9% 21.27
4 Colorado St 3-1 21 298 20.2% 1.3% 20.95
5 Utah St. 4-1 20 353 23.9% 3.7% 20.02
6 UNLV 5-1 19 441 29.9% 6.0% 18.53
USA Today—Despite dropping from #20 to #21 in week 4, the team actually moved up about a half spot from 21.27 to 20.95. And what appeared to only be a jump of one spot this last week, was actually 1.5 spots (6%).
Harris Poll
Wk Opponent Rec rk votes %votes gain effective
4 Colorado St 3-1 19 796 27.9% 19.02
5 Utah St. 4-1 17 847 29.7% 1.8% 18.57
6 UNLV 5-1 17 963 33.8% 4.1% 17.55
Harris—BYU was sitting on some unearned ranking in the Harris poll last week, and so despite gaining more than one spot in the effective rankings (4.1%, 18.57 to 17.55), they teams official rank of 17 remained unchanged.
No comments:
Post a Comment