Week 14 College Football Predictions - 11/29/2007

This is the last regular season set of predictions for the Modus Rankings. This week, there are only 18 games, but these games pack a wallop determining almost all of the BCS bowls. Here are the competitive statistics (note the Top 25 games is almost double):

• Average Rank Delta of Opponents: 36 (34)
• Median Rank Delta of Opponents: 32 (27)
• Games with Top 25 Opponents: 22% (10%)
• Games with Opponents within 10 Ranks: 33% (25%)
• Opponents more than 25 Ranks apart: 50% (56%)
• Div I-A vs. Div II-A Games: 0% (4%)

The teams highlighted green are the predicted winners for this weekend's games:

Home	Rank		Opponent	Rank
Hawaii	1		Washington	89
West Virginia	2		Pittsburgh	87
Missouri	5		Oklahoma	7
LSU	6		Tennessee	22
Arizona St.	9		Arizona	72
Southern California	10		UCLA	57
Boston College	11		Virginia Tech	14
Oregon	20		Oregon St.	41
Tulsa	25		UCF	24
Troy	29		Fla. Atlantic	62
Navy	46		Army	104
Central Mich.	59		Miami (Ohio)	68
Louisville	69		Rutgers	37
Nevada	73		Louisiana Tech	77
San Diego St.	91		BYU	13
Stanford	97		California	52
New Mexico St.	98		Fresno St.	43
Florida Int'l	119		North Texas	111

A smidgeon of history provided free for nothing to journalists and other interested parties

The Associated Press printed this story today about Al Gore going to the White House for the traditional congratulations by the President for winning the Nobel Prize. Of course, instead of reporting the story, the AP leads with this commentary in the first paragraph:

Talk about an inconvenient truth. Al Gore finally won his place in the Oval Office on Monday — right next to George W. Bush. Forever linked by the closest and craziest presidential race in history, the two men were reunited by, of all things, White House tradition.

I briefly mentioned this effect yesterday when I wrote about the Electoral College. In that post, I originally was going to include the outcomes of each vote since 1876 but then decided against it because I thought it was not relevant to the point I was making. After reading the AP story, I changed my mind. A lack of historical context is permeating our society and it is important for us to keep our history in perspective so that we can better meet the unexpected in the future.

Regarding the AP claim that 2000 was the closest election in history, this is not 100% accurate:

• In 1876, Rutherford B. Hayes lost the popular vote and only won the election by 1 (one) electoral vote. In 2000, George W. Bush did not win the popular vote and won the electoral college vote by 5 (five) electoral votes. Samuel Tilden's 1 vote loss is a much closer electoral loss in 1876 than Al Gore's 5 vote loss in 2000.
  
• In 1876, Tilden won the popular vote by 252k votes. Whereas in 2000, Gore won the popular vote by 544k votes.
  
• The Gore margin of victory was only 1% whereas the Tilden margin of victory was 3%. However, in 2000, the 3rd candidate (Nader) accumulated 3% of the vote while in 1876, the 3rd party candidate (Cooper) only gained 1%.
  
• The fact that the Florida vote in 2000 was so close has no bearing on the argument because the Florida margin only affected the closeness of the electoral college which was not as close as 1876.
  

The claim that 2000 was the "craziest" presidential race in history is harder to dispute because it is a subjective statement (subjectivity and journalism is, of course, now allowed). Here are some situations which I subjectively think could qualify as crazier:

• The election of 1800 in which political parties emerged for the first time, a sitting President was deposed, there was a tie in electoral votes, the House of Representatives chose the President, and the presidential and vice presidential candidates fought it out for 36 rounds of voting in the House of Representatives before Jefferson emerged the winner. This was a true crisis - with the country only 12 years into the new constitution, it could have easily fallen apart.
  
• Or how about 1912 when Theodore Roosevelt, unhappy with his hand-selected successor William Taft, split the Republican party when he decided to run as a 3rd party candidate (the Bull Moose Progressive Party) after he did not win the Republican nomination. The incumbent, Taft, won only 8 electoral votes, Roosevelt 88 and Woodrow Wilson, the winner, garnered 435 (even though he won just 42% of the popular vote).
• What about 2008? There are 17 people running in the primaries of the two major parties. None of those involved has been either President of Vice President - the first time in a long time. In addition, this election cycle features several prominent minorities who have a realistic chance of winning it all. This election also features how powerful the internet is and will be in electing people. Previously the internet was used as a fundraising medium, this cycle is the first time it is being prominently integrated into all aspects of the campaign - fundraising, advertising, PR, debates, etc. This internet movement is empowering voters like never before which is probably one of the primary reasons that there are so many candidates running and that the polls are so evenly split amongst them.
  

Lately, I've used this forum to critique others writing. This is not the intention of this blog. This blog exists to capture my thoughts on a multitude of topics while also serving as a way for me to test new internet technologies. However, sometimes I read something that moves me to write my own thoughts down. This is not intended as an attack on the writer (well sometimes it is) but more as an outlet of expression on topics that interest me.

Thank you to infoplease for the historical election data.

I support the Electoral College

I support the electoral college. I think I am in the majority with this viewpoint but there are a lot of columns and letters to the editor that still to this day decry the electoral college failure in 2000 to elect the popular vote winner (Al Gore).

Coming from a large state (California), it would seem I should support a move away from the electoral college and toward electing the president based on the popular vote winner. In fact, because of the electoral college system, my vote in California only has 1/4 the power of a vote in Wyoming.

With 36M California residents, each electoral vote represents 657k people. While in Wyoming, with only 509k residents, each electoral vote represents 169k people. In other words, 169k people in Wyoming have the same voting power as 657k in California. The primary cause of this discrepancy is that each state (+ DC) is guaranteed at least 3 electoral votes and with a limited set of electoral votes (538), the larger states end up having less "per popular vote power" than the smaller states.

I support the electoral college because it embodies within its practice a fundamental element that makes the United States of America unique and special - Federalism. Without Federalism, my vote would only be 1/8th as strong as it is now on ALL issues - not just the presidential election.

California makes up only 12% of the total population. With a national government, the other 88% of the country would have a say in California affairs. So... If we in California want to enact higher fuel efficiency standards, not allowed without a law from the national government. If we in California want to enact environmental protections, not allowed without the national government. If we in California want to create universal healthcare, not allowed without the national government. If we in California want to pay less taxes, not allowed without the national government.

Without a federal system, there would be no autonomous California state government. In a national government, our Governator would report directly to George Bush (and would probably be appointed by the President) and there would be no California Assembly or California Senate - you do not need these institutions with a centralized, national government.

If we like Federalism, but think the electoral college is not a fair shake for California voters, should we re-evaluate the allocation of electoral college votes? Absolutely! California's 55 electoral votes represent only 10.2% of the electoral college vote while its population accounts for 12% - a discrepancy of 17.6%. Wyoming's population represents 0.17% of the total while their electoral votes represent 0.56% of the total - a discrepancy of 70%.

Of note, since 1876, 30% of our presidents have been elected by the electoral college after not winning the majority of the popular vote including:
1876 Rutherford B. Hayes
1880 James A. Garfield
1884 Grover Cleveland
1888 Benjamin Harrison
1892 Grover Cleveland
1912 Woodrow Wilson
1948 Harry S. Truman
1968 Richard M. Nixon
1992 William J. Clinton
2000 George W. Bush

Post-Script: In this post, I have not addressed the issue that the 50.1% of the voters who are in the majority in almost all cases determine which candidate gets 100% of the electoral votes. The other 49.9% of voters (regardless of whether they vote for the winner or loser), have in effect zero voting power in the race. For example, if, in California, the Democratic candidate receives 68% of the popular vote and the Republican candidate receives the other 32%, in effect, 27% of the Democratic voters and 100% of Republican voters have not voted for the President. I will reserve judgment on this because it is an issue that deserves its own post especially with the current debate in California.

Modus College Football Rankings - 11/25/2007

As expected, a solid weekend of college football. Upsets. Emerging powers. Reigning kings re-emerging. And a good amount of overtime! This week's Modus Rankings differ from the BCS by about the same margin as last week but there are two interesting items of note in the Modus Rankings:

1. The Modus algorithm has Kansas ahead of Missouri and Arizona State ahead of USC. This is something I'll be focused on solving in the algo tweaks this offseason.
2. Hawaii is number 1 because they are the only undefeated team left. The question - does the Modus Rankings overweight win percentage or is Hawaii being punished by the pollsters in the BCS? The BCS has Hawaii at #12 while the computers have Hawaii at #14. (Boise State appears to have the same effect)

The following are the statistics from this past weekend's Modus Rankings predictions and this weeks rankings (parentheses denotes statistic from last week).

• Modus Rankings Prediction Success Rate: 73% (83%)
• Modus Rankings vs. BCS Top 25: -0.44 (0.00)
• Modus Rankings vs. BCS Top 10: -2.40 (-2.20)

The following are this weeks Top 25 Modus College Football Rankings plus teams that fell out of the the top 25:

	MODUS Rank	Last	Change		BCS Rank	Diff
1	Hawaii	3	2		Missouri	-4
2	West Virginia	4	2		West Virginia	0
3	Ohio St.	5	2		Ohio St.	0
4	Kansas	1	-3		Georgia	-8
5	Missouri	8	3		Kansas	1
6	LSU	2	-4		Virginia Tech	-8
7	Oklahoma	9	2		LSU	1
8	Boise St.	7	-1		Southern California	-2
9	Arizona St.	6	-3		Oklahoma	2
10	Southern California	15	5		Florida	-6
11	Boston College	13	2		Boston College	0
12	Georgia	14	2		Hawaii	11
13	BYU	16	3		Arizona St.	4
14	Virginia Tech	17	3		Tennessee	-8
15	Texas	10	-5		Illinois	-6
16	Florida	21	5		Clemson	-3
17	South Fla.	20	3		Oregon	-3
18	Virginia	12	-6		Wisconsin	-8
19	Clemson	22	3		BYU	6
20	Oregon	11	-9		Texas	5
21	Illinois	19	-2		South Fla.	4
22	Tennessee	24	2		Virginia	4
23	Cincinnati	23	0		Cincinnati	0
24	UCF	29	5		Auburn	-10
25	Tulsa	28	3		Boise St.	17
26	Wisconsin	25	-1
27	Air Force	27	0
28	Connecticut	18	-10

Week 13 College Football Predictions - 11/21/2007

This is going to be a great week 13 for college football. Not only is it Thanksgiving (4-day weekend!), there are a number of BIG games that affect the national title picture and there are a good amount of games with evenly-matched opponents. The competitive stats for this week show the drastic jump in quality of the games vs Week 12 (denotes last week):

• Average Rank Delta of Opponents: 34 (40)
• Median Rank Delta of Opponents: 27 (36)
• Games with Top 25 Opponents: 10% (4%)
• Games with Opponents within 10 Ranks: 25% (19%)
• Opponents more than 25 Ranks apart: 56% (65%)
• Div I-A vs. Div II-A Games: 4% (2%)

Big Games (top ranked teams with potentially stiff competition - predicted winners in green - listed in no particular order):

Home	Rank		Opponent	Rank
Arizona St.	6		Southern California	15
Georgia Tech	41		Georgia	14
Hawaii	3		Boise St.	7
Kansas	1		Missouri	8
LSU	2		Arkansas	43
West Virginia	4		Connecticut	18

Most Competitive Games (opponents who rank within 10 of each other):

Home	Rank		Opponent	Rank
BYU	16		Utah	26
Colorado	65		Nebraska	69
Idaho	118		Utah St.	116
Kent St.	100		Buffalo	91
Marshall	109		UAB	111
North Carolina St.	75		Maryland	72
Virginia	12		Virginia Tech	17
Washington	85		Washington St.	86
Western Mich.	90		Temple	88

All The Rest:

Home	Rank		Opponent	Rank
Akron	89		Central Mich.	64
Auburn	36		Alabama	51
Boston College	13		Miami (Fla.)	78
Bowling Green	46		Toledo	79
Colorado St.	107		Wyoming	71
East Caro.	56		Tulane	93
Florida	21		Florida St.	35
Florida Int'l	119		Fla. Atlantic	70
Fresno St.	48		Kansas St.	73
Houston	42		Texas Southern	NR
Kentucky	37		Tennessee	24
La.-Lafayette	103		La.-Monroe	80
Memphis	61		So. Methodist	117
Mississippi St.	58		Mississippi	99
New Mexico	44		UNLV	105
North Carolina	97		Duke	113
North Texas	115		Western Ky.	NR
Northern Ill.	112		Ball St.	57
Ohio	83		Miami (Ohio)	60
Oklahoma	9		Oklahoma St.	53
Pittsburgh	84		South Fla.	20
Rice	104		Tulsa	28
San Diego St.	87		TCU	49
San Jose St.	94		Nevada	67
South Carolina	54		Clemson	22
Southern Miss.	59		Arkansas St.	77
Stanford	95		Notre Dame	108
Syracuse	110		Cincinnati	23
Texas A&M	55		Texas	10
Troy	31		Middle Tenn. St.	82
UCF	29		UTEP	92
UCLA	62		Oregon	11
Vanderbilt	74		Wake Forest	38

Modus College Football Rankings - 11/18/2007

I am going to brag for a moment. This past weekend, the Modus rankings picked 83% of the outright winners correctly. EIGHTY-THREE PERCENT! That is up 14% from the average of 69%. This is a good sign, but even still, at the end of the season, I am going to evaluate the algorithm to determine if it needs tweaks. I am also going to analyze its performance against the spread, if I have time.

The following are the statistics from this past weekend's Modus Rankings predictions and this weeks rankings (parentheses denotes statistic from last week).

• Modus Rankings Prediction Success Rate: 83% (69%)
• Modus Rankings vs. BCS Top 25: 0.00 (-1.00)
• Modus Rankings vs. BCS Top 10: -2.20 (-1.10)

The following are this weeks Top 25 Modus College Football Rankings plus teams that fell out of the the top 25:

RANK	MODUS	Last	Change	BCS	Modus v. BCS Diff
1	Kansas	1	0	LSU	-1
2	LSU	2	0	Kansas	1
3	Hawaii	8	5	West Virginia	-1
4	West Virginia	6	2	Missouri	-4
5	Ohio St.	3	-2	Ohio St.	0
6	Arizona St.	4	-2	Arizona St.	0
7	Boise St.	10	3	Georgia	-7
8	Missouri	9	1	Virginia Tech	-9
9	Oklahoma	7	-2	Oregon	-2
10	Texas	12	2	Oklahoma	1
11	Oregon	5	-6	Southern California	-4
12	Virginia	11	-1	Florida	-9
13	Boston College	17	4	Texas	3
14	Georgia	19	5	Boston College	1
15	Southern California	13	-2	Hawaii	12
16	BYU	21	5	Virginia	4
17	Virginia Tech	14	-3	Illinois	-2
18	Connecticut	15	-3	Tennessee	-6
19	Illinois	22	3	Boise St.	12
20	South Fla.	24	4	Connecticut	2
21	Florida	30	9	Wisconsin	-4
22	Clemson	18	-4	Clemson	0
23	Cincinnati	16	-7	South Fla.	3
24	Tennessee	28	4	Cincinnati	1
25	Wisconsin	25	0	BYU	9
26	Utah	27	1
27	Air Force	26	-1
28	Tulsa	35	7
29	UCF	31	2
30	Penn St.	20	-10
31	Troy	33	2
32	Michigan	23	-9