the 1998 Season

Versions / 1998 / 1999 / 2000 / 2001 / 2002 / 2003 / 2004 / 2005 / 2006 / 2007 / Conclusions

Version A was used in 1998 - below is what the final rankings looked like.

Version A: 1998
Rank	Team	AP	Coach	Poll Avg	A&H	JS	NYT	Comp.Avg	SoS	SoS Rank	Loss	TOTAL
1	Tennessee	1	1	1	1	2	2	1.67	20	0.80	0	3.47
2	Florida State	2	2	2	2	2.25	1	1.75	4	0.16	1	4.91
3	Kansas State	4	4	4	4	1	4	3	49	1.96	1	9.96
4	Ohio State	3	3	3	6.75	6	3	5.25	28	1.12	1	10.37
5	UCLA	6	5	5.5	3	4	5.25	4.08	8	0.32	1	10.90
6	Texas A&M	8	9	8.5	6	5	4	5	5	0.20	2	15.70
7	Arizona	5	6	5.5	5	9	9	7.67	58	2.32	1	16.49
8	Florida	7	7	7	10	8	11	9.67	32	1.28	2	19.95
9	Wisconsin	9	8	8.5	9	10	10	9.67	61	2.44	1	21.61
10	Tulane	10	10	10	8	14	16.5	12.83	96	3.84	0	26.67
11	Nebraska	14	16	15	11	7	13.5	10.5	14	0.56	3	29.06
12	Virginia	13	12	12.5	13	18	17	16	43	1.72	2	32.22
13	Arkansas	11	11	11	17	12	21.75	16.92	59	2.36	2	32.28
14	Georgia Tech	12	14	13	16	20	12	16	44	1.76	2	32.76
15	Syracuse	18	17	17.5	17.25	16	7	13.42	22	0.88	3	34.80

Tennessee, Florida State. By a big margin. So let's look at what the rankings would be if Version B had been used. (Remember, the only change between Versions A & B was that B added in a lot more computers...) *A quick note: all of the rankings in the first column are the teams' official BCS rankings from that year, not the rankings from each version's results.

Version B: 1998
Rank	Team	AP	Coach	Poll Avg	Comp.Avg	SoS	SoS Rank	Loss	TOTAL
1	Tennessee	1	1	1	1.71	20	0.8	0	3.51
2	Florida State	2	2	2	1.43	4	0.16	1	4.59
4	Ohio State	3	3	3	4.57	28	1.12	1	9.69
3	Kansas State	4	4	4	3.43	49	1.96	1	10.39
5	UCLA	6	5	5.5	4.43	8	0.32	1	11.25
6	Texas A&M	8	9	8.5	5.43	5	0.2	2	16.13
7	Arizona	5	6	5.5	7.71	58	2.32	1	16.53
8	Florida	7	7	7	7.43	32	1.28	2	17.71
9	Wisconsin	9	8	8.5	9.57	61	2.44	1	21.51
11	Nebraska	14	16	15	10.14	14	0.56	3	28.70
13	Arkansas	11	11	11	14	59	2.36	2	29.36
10	Tulane	10	10	10	15.57	96	3.84	0	29.41
14	Georgia Tech	12	14	13	14.29	44	1.76	2	31.05
12	Virginia	13	12	12.5	14.86	43	1.72	2	31.08
15	Syracuse	18	17	17.5	15.29	22	0.88	3	36.67

No change in #1 or #2, and they're still way ahead of #3, which switches to Ohio State in this Version.

Let's take a look at the first ranking I had to project, Version C in 1998. Columns in light green are ranges of mathematical possibilities, in this case the lowest & highest computer averages as well as the highest and lowest quality win components.

Version C: 1998
Rank	Team	Poll Avg	LoComp	HiComp	SoS Pts	Loss	Lo-Sub	Hi-Sub	Hi-QW	Lo-QW	Lo-Tot	Hi-Tot
1	Tennessee	1	1.50	1.67	0.80	0	3.3	3.47	-1.4	-1.3	1.90	2.17
2	Florida State	2	1.67	2	0.16	1	4.83	5.16	-2.4	-2.2	2.43	2.96
5	UCLA	5.5	3.33	4.17	0.32	1	10.15	10.99	-1.0	-0.9	9.15	10.09
3	Kansas State	4	3.33	4	1.96	1	10.29	10.96	-0.5	-0.3	9.79	10.66
4	Ohio State	3	5.33	6.17	1.12	1	10.45	11.29			10.45	11.29
6	Texas A&M	8.5	5.17	5.67	0.20	2	15.87	16.37	-1.8	-1.4	14.07	14.97
7	Arizona	5.5	7.17	7.83	2.32	1	15.99	16.65			15.99	16.65
8	Florida	7	7.83	8.5	1.28	2	18.11	18.78			18.11	18.78
9	Wisconsin	8.5	8.83	10	2.44	1	20.77	21.94			20.77	21.94
10	Tulane	10	13.33	15.33	3.84	0	27.17	29.17			27.17	29.17
13	Arkansas	11	13.33	14.33	2.36	2	28.69	29.69			28.69	29.69
11	Nebraska	15	10.5	12.5	0.56	3	29.06	31.06			29.06	31.06
12	Virginia	12.5	13.83	15.33	1.72	2	30.05	31.55			30.05	31.55
14	Georgia Tech	13	15.33	17.17	1.76	2	32.09	33.93	-0.4	-0.3	31.69	33.63

As you can see, Tennessee & Florida State are still way ahead, but now they're probably followed by UCLA at #3 because of the Bruins high quality win deduction.

Just to help you get a feel for the math, let's look at an example. #1 Tennessee's lowest possible computer average was 1.50, and their highest was 1.67. The low number is calculated by assuming that Wolfe had them ranked #1, in which case their computer rankings would have been 1,1,1,1,2,2,2,2. Version C drops the highest and lowest, (a 1 and a 2) and averages the other 6 (1+1+1+2+2+2=9, 9/6) to get 1.5.

To calculate the high average, we don't need to plug in a number - all we have to do is assume that Wolfe's ranking was the highest one they received, whether it was a 3, a 5 or a 19. We can assume that because we're just going to eliminate that high number as well as their low, in this case a 1, as well. So after eliminating their 1, we're left with 1,1,2,2,2,2, which added together is 10, divided by 6 is 1.67. Get it? (I hope so, because these projected rankings require a lot of it.)

So because Tennessee's highest possible total is 2.17, while Florida State's lowest possible total is 2.43, (remember, the lower the number, the higher the ranking), there's no mathematical way that Florida State could be ranked higher than Tennessee, no matter what the missing rankings might have been. At the same time, Florida State's high is 2.96 and UCLA's low is 9.15, meaning that UCLA could not ever have a higher total than Florida State, giving the #2 spot to Florida State. So Version C would have definitely pitted Tennessee vs Florida State in the National Championship game.

Let's look at Version D, which is also a projection.

Version D: 1998
Rank	Team	Poll Avg	Lo Comp	Hi Comp	SoS Pts	Loss	Lo-Sub	Hi-Sub	Hi-QW	Lo-QW	Lo-Tot	Hi-Tot
1	Tennessee	1	1.5	1.67	0.80	0	3.3	3.47	-0.4	-0.3	2.90	3.17
2	Florida State	2	1.33	1.67	0.16	1	4.49	4.83	-0.9	-0.7	3.59	4.13
4	Ohio State	3	4.33	5.5	1.12	1	9.45	10.62			9.45	10.62
5	UCLA	5.5	3.33	4.5	0.32	1	10.15	11.32	-0.5	-0.4	9.65	10.92
3	Kansas State	4	3.33	4	1.96	1	10.29	10.96	-0.1	0.0	10.19	10.96
6	Texas A&M	8.5	4	5	0.20	2	14.70	15.70	-0.9	-0.6	13.80	15.10
7	Arizona	5.5	6.33	7.67	2.32	1	15.15	16.49			15.15	16.49
8	Florida	7	6.83	8.5	1.28	2	17.11	18.78			17.11	18.78
9	Wisconsin	8.5	7.17	9	2.44	1	19.11	20.94			19.11	20.94
10	Tulane	10	11.83	15.5	3.84	0	25.67	29.34			25.67	29.34
12	Virginia	12.5	10.83	13.67	1.72	2	27.05	29.89			27.05	29.89
14	Georgia Tech	13	10.67	13.83	1.76	2	27.43	30.59	-0.1	0.0	27.33	30.59
13	Arkansas	11	12.17	15.67	2.36	2	27.53	31.03			27.53	31.03
11	Nebraska	15	10.33	13.33	0.56	3	28.89	31.89			28.89	31.89

Remember, one of the minor changes from Version C to D was that one computer was dropped, resulting in a formula that only drops the lowest ranking instead of the highest & lowest. It hasn't resulted in anything big here - still Tennessee & Florida State.

Finally, let's look at Version E, the latest incarnation.

Version E: 1998
Rank	Team	AP	Votes	AP%	Coach	Votes	Coach%	HiComp%	LoComp%	Hi-Tot	Lo-Tot
1	Tennessee	1	1750	1.000	1	1548	0.9987	0.980	0.970	0.9929	0.9896
2	Florida State	2	1671	0.9549	2	1464	0.9445	0.980	0.960	0.9598	0.9531
4	Ohio State	3	1476	0.8434	4	1337	0.9065	0.820	0.770	0.8806	0.8640
3	Kansas State	4	1602	0.9154	3	1405	0.8626	0.900	0.860	0.8687	0.8553
5	UCLA	6	1398	0.7989	5	1253	0.8084	0.910	0.860	0.8391	0.8224
7	Arizona	5	1310	0.7486	9	1092	0.8026	0.760	0.720	0.7898	0.7765
6	Texas A&M	8	1412	0.8069	6	1244	0.7045	0.850	0.820	0.7677	0.7577
8	Florida	7	1337	0.7640	7	1156	0.7458	0.740	0.700	0.7499	0.7366
9	Wisconsin	9	1176	0.6720	8	1103	0.7116	0.720	0.650	0.7012	0.6779
10	Tulane	10	1067	0.6097	10	964	0.6219	0.540	0.420	0.5905	0.5505
13	Arkansas	11	815	0.4657	16	590	0.5290	0.490	0.430	0.5225	0.5025
12	Virginia	13	817	0.4669	12	818	0.5277	0.570	0.480	0.5215	0.4915
14	Georgia Tech	12	960	0.5486	11	820	0.4832	0.530	0.430	0.5042	0.4709
11	Nebraska	14	874	0.4994	14	749	0.3806	0.580	0.460	0.4755	0.4355
15	Syracuse	18	610	0.3486	17	538	0.3471	0.210	0.110	0.3019	0.2686

In Version E, the rankings switched from "Low Total Wins" to "Highest Percentage Wins", which is why we see the big visual change here. With 6 computers providing rankings, the formula drops highest & lowest, then adds the other 4 together. Dropping the highest and lowest serves as a buffer, making the range of possibile computer averages quite low for each team.

As was the case with all the other Versions, Tennessee & Florida State are at the top. There wasn't much controversy that year, and there isn't between the different versions either.

Onto 1999!

Versions home < Top > 1999