RTT draft BlogPoll ballot: the raw computer rankings

by Joel on Sep 22, 2008 9:37 PM EDT in College Football Comment 12 comments

Time for yet another iteration of the RTT Ready, Fire, Aim BlogPoll ballotty thing. For a complete refresher on what we're doing, go here, here, here, and here.

Adjustments for this week:

None. The thing appears to be stablizing somewhat, and any tweaks, I think, can simply be ad hoc human interventions, for which I'm soliciting assistance below.
The categories again: Win/Loss, Strength of Schedule, Pass Efficiency Defense, Rush Defense, 3rd Down Offense, Total Defense, Pass Efficiency Offense, Experience, Opponents Points Per Game, Total Offense, and 3rd Down Defense. The AVG R column is Average Ranking, and the last column is Weighted Average.
The current weighting of the categories: 30, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1.
Current spreadsheet here.

All right, then. The Top 35, according to the computer:

Rank	Team	WL	SOS	PED	RD	3DO	TD	PEO	EXP	OPPG	TO	3DD	AVG R	W AVG
1	Southern California	1	26	9	6	29	2	10	72	1	18	28	18.36	94.45
2	Utah	1	25	33	10	32	3	25	14	42	40	22	22.45	118.91
3	Penn St.	1	102	8	7	5	7	18	38	9	6	8	19.00	140.36
4	TCU	1	85	3	1	6	1	79	26	6	38	6	22.91	142.00
5	Oklahoma	1	91	15	11	11	10	3	59	20	4	18	22.09	148.82
6	Alabama	1	38	13	8	15	14	69	95	8	57	16	30.36	150.82
7	Florida	1	52	10	15	9	5	34	116	2	86	25	32.27	153.64
8	Nebraska	1	10	44	18	55	63	17	32	22	20	63	31.36	165.00
9	Georgia	1	52	54	3	14	24	15	90	18	22	34	29.73	169.91
10	LSU	1	38	16	9	64	18	61	77	14	32	14	31.27	174.09
11	Wisconsin	1	6	47	21	43	44	31	77	17	37	93	37.91	175.18
12	South Fla.	1	79	39	14	25	12	29	66	40	19	29	32.09	193.27
13	BYU	1	91	35	37	1	35	8	49	11	13	68	31.73	193.73
14	Missouri	1	38	60	30	20	77	4	32	55	1	45	33.00	195.64
15	Kansas	28	38	18	27	3	41	19	32	27	15	47	26.82	204.18
16	Kentucky	1	52	11	5	100	17	92	38	2	81	12	37.36	207.55
17	Ball St.	1	26	17	104	7	86	5	53	50	12	84	40.45	212.27
18	Boise St.	1	10	57	75	29	52	6	114	28	24	18	37.64	217.64
19	Wake Forest	1	52	6	33	105	25	35	49	23	85	43	41.55	218.73
20	Florida St.	45	1	45	4	13	4	55	53	2	34	2	23.45	225.00
21	Texas	1	85	52	13	28	58	2	116	11	9	37	37.45	232.55
22	Oklahoma St.	1	117	36	59	10	29	9	45	56	5	3	33.64	240.55
23	Texas Tech	1	117	37	32	17	62	22	68	30	3	3	35.64	251.00
24	Louisville	45	1	40	2	57	13	77	18	61	42	13	33.55	268.18
25	Connecticut	1	91	30	22	53	21	100	109	15	39	57	48.91	272.64
26	Colorado	1	52	42	71	41	45	40	114	39	68	59	52.00	274.64
27	Northwestern	1	112	26	28	56	32	93	59	13	72	20	46.55	281.45
28	Arizona	28	109	5	49	27	6	26	77	29	36	11	36.64	290.91
29	Minnesota	1	91	38	42	68	65	11	112	35	45	47	50.45	293.55
30	Iowa	28	91	2	30	89	19	51	53	5	52	16	39.64	304.27
31	Georgia Tech	28	52	12	64	38	30	53	109	21	48	90	49.55	306.91
32	Clemson	28	10	51	44	102	48	38	59	25	56	94	50.45	314.55
33	Vanderbilt	1	70	32	54	75	72	67	90	32	110	66	60.82	316.73
34	Ohio St.	28	79	22	29	72	15	60	59	24	92	71	50.09	319.73
35	East Carolina	28	73	56	53	48	31	32	6	47	66	58	45.27	321.36

I'd like to do things a little bit different this week. Rather than me issue-spotting the computer results and implementing the human interventions myself, I'd like to get your feedback. No limit on interventions this time around. We're just looking for something approaching a consensus. I won't actuallly submit the ballot MGoBlog's site until we're somewhat comfortable with it.

So, what changes do y'all think we should make to the top 25?

0 recs | Comment 12 comments | Share

Story-email Email Printer Print

Comments

Here's a thought.

In the spreadsheet itself, I tried using a ‘square root of the sum of squares’ trick. Basically, in the “BlogPoll” sheet, I changed the W Avg column (column O) by squaring all of the individual terms in the numerator (e.g. (C2*Weighting!$A$2) becomes (C2*Weighting!$A$2)^2 ). Then you take the square root of the entire numerator before dividing by 11.

That will help emphasize categories that score really, really high. Their squares will dominate for a particular team, but the square root will bring the values back down to a more sensible level. (Take a look at Penn State’s score before and after, and you’ll see what I mean. Their Strength of Schedule has a lot of points, and this trick reduces them from #3 to #17.)

(I saved the file here.) Note that in the file, I re-sorted by the W Avg already.

If you don’t want to wait to open the file, here’s the top 30:

1. Southern Cal
2. Utah
3. Alabama
4. Wisconsin
5. Nebraska
6. Florida
7. LSU
8. Georgia
9. Missouri
10. TCU
11. South Florida
12. Oklahoma
13. Boise State
14. Wake Forest
15. Kansas
16. Kentucky
17. Penn State
18. BYU
19. Ball State
20. Colorado
21. Texas
22. Vanderbilt
23. Connecticut
24. Georgia Tech
25. Minnesota
26. Clemson
27. Texas Tech
28. Oklahoma State
29. Northwestern
30. Ohio State

by Hooper on Sep 22, 2008 10:37 PM EDT reply actions 0 recs

Of course,

you might want to dial down the weighting on wins a touch with this. Multiplying 80 and 30, then squaring will tend to swamp the other factors. You might not, though; a strong weighting makes it a lot easier to favor any undefeateds in the last week. Up to you.

by Hooper on Sep 22, 2008 10:53 PM EDT up reply actions 0 recs

Homerism aside

Ball State is a quality football program, Joel. Hope you leave them in.

by OverThePylon on Sep 23, 2008 8:39 AM EDT reply actions 0 recs

A tough loss to LSU knocks Auburn all the way out of the top 35?

Huh?

I would have moved us down further after a 1-point win over MSU than a 5-point loss to LSU

War Damn Eagle!

by PowerOfDixieland on Sep 23, 2008 10:25 AM EDT reply actions 0 recs

And how does Florida St. have the #1 SOS?

War Damn Eagle!

by PowerOfDixieland on Sep 23, 2008 10:26 AM EDT up reply actions 0 recs

At this point . . .

. . . all undefeated teams are tied for the No. 1-ranked SOS.

Go Vols!

by Joel on Sep 23, 2008 10:53 AM EDT up reply actions 0 recs

I thought they were all tied for #1 W/L rather than #1 SOS

War Damn Eagle!

by PowerOfDixieland on Sep 23, 2008 12:46 PM EDT up reply actions 0 recs

Uh, yeah

Meant that. But there are still several tied at No. 1 for SOS as well. Don’t have time to look it up, but have a look at the spreadsheet.

Go Vols!

by Joel on Sep 23, 2008 2:22 PM EDT up reply actions 0 recs

Yeah

I think that’s wacky. Where do you think they should go?

Go Vols!

by Joel on Sep 23, 2008 10:53 AM EDT up reply actions 0 recs

How about immediately after Tennessee? :-D

They’re getting killed by the loss (W = 28, then multiplied by 30) and SoS (85, then multiplied by 10), which are basically dwarfing everything else. The 3DO isn’t helping either (112, then multiplied by 7). the only comparable team ahead of them is Arizona (W = 28, SoS = 109), but they have great numbers multiplied by 9,8,7,6, and 5 to compensate.

by Hooper on Sep 23, 2008 11:06 AM EDT up reply actions 0 recs

In other words,

the MSU game killed Auburn in the SoS and offensive stats (and boy were they offensive! har, har, har [snort] har, har … [sigh] I kill me) and the LSU game killed them in the W/L column.

by Hooper on Sep 23, 2008 11:14 AM EDT up reply actions 0 recs