Total Div IA Teams:.....................119 Total Div IA Teams playing this week:....82
Total Games involving Div IA Teams:......41
Div IA Teams playing non-IA teams:........0 (One game played at Div IA home)
Div IA Teams losing to non-Div IA team....0
Average Points per game:.................57 Highest Total:..........................108 (Central Michigan at Eastern Michigan[w]) Lowest Total:............................15 (UAB[w] at UCF)
Average spread of victory:...............18 Highest Spread:..........................51 (Fresno State at Boise State[w])
Lowest Spread:............................1 (LSU at Arkansas[w])
Visiting Teams that won:.................15 Visiting Teams that won - Non-Div IA:.....0
Interesting Facts:
Eight of the BCS Top 25 did not play in Week 14
Number of Teams Undefeated Teams left after 14 weeks: 4
BCS Top 5 - FBS Opponents Cumulative Win/Loss Records Alabama....64-66...49% Oklahoma...78-52...60% Texas......85-58...59% Florida....75-56...57% USC........67-73...48%
It didn't help USC to have played any of the Washington teams as far as SoS goes.
Next Week, Week 15...The Last Week of Regular Season.
Total Div IA Teams:.....................119 Total Div IA Teams playing this week:....96
Total Games involving Div IA Teams:......49
Div IA Teams playing non-IA teams:........2 (One game played at Div IA home)
Div IA Teams losing to non-Div IA team....0
Average Points per game:.................55 Highest Total:..........................107 (Eastern Michigan at Temple[w]) Lowest Total:............................16 (Clemson[w] at Virginia)
Average spread of victory:...............19 Highest Spread:..........................55 (Iowa[w] at Minnesota)
Lowest Spread:............................1 (two games) (Syracuse[w] at Notre Dame & Cal Poly at Wisconsin[w]
Visiting Teams that won:.................20 Visiting Teams that won - Non-Div IA:.....0
Interesting Facts:
Seven of the BCS Top 25 did not play in Week 12
Number of Teams Undefeated Teams left after 13 weeks: 4
The Cal Poly Mustangs Number 3 in FCS Coaches Poll Record: 8-3, 1-1 against FBS Teams
Washinton Leads the Nation... ...in Return Defense. Lowest Total Return Yardage - Kickoffs and Punt Returns.
Next Week, Week 14...Two Weeks of Regular Season Left.
Total Div IA Teams:.....................119 Total Div IA Teams playing this week:....99
Total Games involving Div IA Teams:......50
Div IA Teams playing non-IA teams:........1 (One game played at Div IA home)
Div IA Teams losing to non-Div IA team....0
Average Points per game:.................54 Highest Total:..........................100 (two games) (Tulsa at Houston[w] & Arizona at Oregon[w]) Lowest Total:............................24 (East Carolina at Southern Miss[w])
Average spread of victory:...............17 Highest Spread:..........................59 (Louisiana-Monroe at Mississippi[w])
Lowest Spread:............................2 (two games) (Virginia Tech at Miami[w] & North Carolina at Maryland[w]
Visiting Teams that won:.................24 Visiting Teams that won - Non-Div IA:.....0
Interesting Facts:
Five of the BCS Top 25 did not play in Week 12
Number of Teams Undefeated Teams left after 12 weeks: 5
Big game this week: Texas Tech at Oklahoma Both Teams coming off a bye week.
Total Div IA Teams:.....................119 Total Div IA Teams playing this week:...100
Total Games involving Div IA Teams:......50
Div IA Teams playing non-IA teams:........1 (One game played at Div IA home)
Div IA Teams losing to non-Div IA team....0
Average Points per game:.................55 Highest Total:...........................93 (Rice[w] at UTEP) Lowest Total:............................23 (three games) (Air Force[w] at Army & Utah[w] at New Mexico & East Carolina[w] at UCF)
Average spread of victory:...............16 Highest Spread:..........................58 (Stanford[w] at Washington State)
Lowest Spread:............................1(three games) (Wisconsin at Michigan State[w] & Troy at Louisiana-Monroe[w] & Kentucky[w] at Mississippi State)
Visiting Teams that won:.................19 Visiting Teams that won - Non-Div IA:.....0
Interesting Facts:
Four of the AP Top 25 did not play in Week 10
Number of Teams Undefeated Teams left after 10 weeks: 6
Undefeated teams Away (11) ---- Air Force Alabama Ball State Boise State Florida Georgia Oklahoma Penn State Pittsburgh Texas Tech Utah
At Home (26) ---- Alabama Ball State Boise State BYU California Central Michigan Cincinnati Louisiana Tech Louisiana-Lafayette Maryland Northern Illinois Oklahoma State Oregon State Penn State Rice Stanford TCU Texas Texas Tech Troy Tulsa USC Utah Virginia Tech West Virginia Western Michigan
Next Week, Week 11. Yeah, not much time to do much with all the numbers I've been accumulating, but I'll be better prepared next year. That's right, don't forget to vote. If not, the curse of the Fatman will follow you...
Total Div IA Teams:.....................119 Total Div IA Teams playing this week:...100
Total Games involving Div IA Teams:......50
Div IA Teams playing non-IA teams:........0 (One game played at Div IA home)
Div IA Teams losing to non-Div IA team....0
Average Points per game:.................53 Highest Total:...........................93 (Oklahoma[w] at Kansas State) Lowest Total:............................17 (Duke[w] at Vanderbilt)
Average spread of victory:...............16 Highest Spread:..........................58 (Two games) (Colorado at Missouri[w] & Kentucky atFlorida[w])
Lowest Spread:............................1 (Two games) (Florida Atlantic[w] at Louisiana-Monroe & Central Michigan[w] atToledo)
Visiting Teams that won:.................23 Visiting Teams that won - Non-Div IA:.....0
Interesting Facts:
Two of the AP Top 25 did not play in Week 9
Number of Teams Undefeated Teams left after 9 weeks: 8
BCS 1 thru 5 in Ranked Stats Rnk.Team......W/L..TYds..RYds..Pyds..TPt s..TDef 1. Texas......8-0...5th..30th..11th...6th..28th 2. Alabama....8-0..35th..16th..98th..27th..10th 3. Penn State.9-0...8th...8th..18th...5th...3rd 4. Oklahoma...7-1...2nd..33rd...2nd...2nd..51st 5. USC........6-1..14th..41st..37th..22nd...1st
So, Penn State best balance between offense and defense. Does defense win championships?
Next week, Week 10.
DF
Ps. Vista sucks. Took all morning to run updates on my new computer...
While this article from the NY Times talks about boosters and how the current financial problems may change their ability to give, it also talks about the amount money tied up by the six members of the BCS bowl series and related institutions.
Part Three – The BCS Computer Component and BCS Problems and Fixes
The BCS Computer Component
The 2008 BCS computer component consists of six ranking; Fox has the list of the participants and links to their sites here.
How the rankings in each "poll" are calculated is too complex for this blog. If you are truly interested, you can follow the links on the page listed above to the participant home page and look for the detail there. Here is how the BCS component is calculated:
For each team, their rank in each poll is added together, the final score for each team being that sum divided by six (the average);
The scores can be sorted to rank the computer component.
The final score for each team is the sum of individual scores given by each component of the BCS, the Harris Poll, the Coaches Poll, and the computer-ranked component, divided by three (the average). The scores are sorted to give the BCS ranking for that week. The first BCS rankings come out two Sundays from now (October 19th)
Flaws in the Human Polls
If you've been following along, you probably noticed that the mathematics of the three components tend to feed on itself. Tell me if this statement makes sense: The average of the sums of the averages of the sum of the ranks creates a meaningful relationship between the true strength of teams. Nonsense, right? If you remember my first blog on this topic, performance on the field is the only true determinant of strength.
What makes the human polls mostly useless are, of course, the humans. Believe it or not, I am not talking about a human failing like corruptibility, but the more positive aspects of preference and differentiation.
Fixes for the Human Polls
In the human polls, how many points are there between each ranked team? One. Only one. So the difference in strength between the first place team and second place team on any one ballot is one point. The difference between first and third is two points. And, so on.
Imagine if we gave each voter a thousand (1,000) points to distribute along with the ranking on their ballot. This would show a difference between each of the ranks and create a preference differential that can be included in a larger calculations.
Flaws in the Computer Polls
I don't have any simple fixes for the problems that the computer polls have, though I'm willing to list them:
Strength of Schedule - is based on last's years win/loss of teams your team will play against this year
Margin of win and quality of opposition - supposed to remove the 'cupcake' factor
Calculations to remove Home field advantage
Calculations and recalculations to give higher weight to a team's performance in their more recent game
Unknown portions of systems
Well, I had to know for myself And, of course, you know how I feel about it all: performance on the field...
And, you see how much is involved in creating promoting the system...it feeds upon itself. That's why there won't be playoff system...for a long time.
The BCS computation comes in three parts, two of which are the human polls, the Harris Interactive and the Coaches Poll, and one part that is made of an amalgam of computer-generated rankings.
The Harris Poll is composed of former players, coaches, and persons associated with various communications media. The 2008 pool consists of 114 voters. The poll is managed by Harris Interactive.
The coaches poll is composed of Division I-A coaches, all members of the American Football Coaches Association. The 2008 pool consists of 61 coaches. The poll is now managed by USA Today.
For these human polls, each member in the pool picks his top 25 teams and ranks them from 1 (top) to 25 (bottom). The key is that each coach doesn't necessarily start out with the same 25 teams or ranks them in the same order. In the case of the Coaches Poll, there have been questions of bias as the general public does not know what was on the individual ballots as they are kept secret.
In the published list of both polls, there is a points column on which the list is sorted. Where do the points come from? The formula is simple: Each team on each ballot is given points in inverse of their position on the ballot. That is, the first place team on a ballot receives 25 points and the last place team receives 1 point. The points are accumulated for each team across all ballots in that poll.
Lets look at an example: In week three of the Coaches Poll, where USC is ranked number one with 56 first place votes and a total 1517 points. 1517 minus 1400 (56 first place votes times 25 points) leaves 117 points and five voters. 117 points in five parts is 4 second place votes and a fifth place vote at best or two second and three thirds at worst.
As a component of the BCS calculation, these points and divided by the total number of possible point if everyone voted a particular team. In the case of the Coaches Poll, the divisor is (61 coaches times 25 first place points = ) 1525. So, if the BCS were to come out this week, USC would have a component value of .995 (1517/1525).
Enough math, lets look at how the members of the Coaches poll are distributed:
Distribution of Coaches by Conference
CONF......................COUNT Atlantic Coast Conference....6 Big 12 Conference............7 Big East Conference..........4 Big Ten Conference...........7 Conference USA...............6 FBS Independents.............1 Mid-American Conference......6 Mountain West Conference.....4 Pacific-10 Conference........5 Southeastern Conference......7 Sun Belt Conference..........4 Western Athletic Conference..4
South East 22 AL.....2 FL.....6 GA.....1 KY.....1 LA.....2 MS.....1 NC.....3 SC.....2 TN.....2 VA.....1 WV.....2
South West 9 OK.....1 NM.....2 TX.....6
West 11 CA.....4 CO.....1 ID.....1 OR.....2 UT.....1 WA.....1
The names of the 114 members of the 2008 Harris Interactive Poll can be found here. Someone before the season is out will look-up each member and their affiliation. I will put the link to that document in this blog when it happens.
Total Div IA Teams:.....................119 Total Div IA Teams playing this week:...101
Total Games involving Div IA Teams:......56
Div IA Teams playing non-IA teams:.......11 (All games played at Div IA home)
Div IA Teams losing to non-Div IA team....0
Average Points per game:.................51 Highest Total:..........................109 (Fresno State[w] at Toledo)
Lowest Total:............................12 (Wake Forest[w] at Florida State)
Average spread of victory:...............19 Highest Spread:..........................54 (South Carolina State at Clemson[w])
Lowest Spread:............................1 (three games) (Fresno State[w] at Toledo & New Mexico State[w] at UTEP & Iowa at Pittsburgh[w])
Visiting Teams that won:.................18 Visiting Teams that won - Non-Div IA:.....0
Conf.......W.....L ACC.......28....13 Big 12....33.....8 Big East..16....11 Big Ten...31.....8 C-USA.....19....25 MAC.......20....29 MWC.......23....11 Pac-10....17....17 SEC.......32....11 Sun Belt...9....19 WAC.......14....15 Ind........4.....6
Interesting Facts:
Four of the AP Top 25 did not play in Week 4
Number of Teams still undefeated: 27
Distribution of Undefeated Teams across Conferences
ACC........1 Big 12.....7 Big East...2 Big Ten....4 C-USA......1 MAC........1 MWC........3 Pac-10.....1 SEC........6 Sun Belt...0 WAC........1 Ind........0
Current Undefeated Team to Face a Current Undefeated Team:
Statistics is about gathering data for the purpose of making some sort of statement about a collection of objects. The data collected about the objects is quantitative which means that the data consists of counts or measures. What the designer of the collection has in mind is to create a model that emulates a real-world entity, so the first goal is TO DESCRIBE. Any other goal is secondary, including PREDICTION. One quantitative tool is the AVERAGE or the MEAN which is used to describe central tendencies. In other words, the AVERAGE places a point in the mathematical middle of the collected data which acts as a signpost from which individual objects can measure how close or how far they are from the middle. This tool works best when comparing subsets of the collection to the whole.
Surveys also gather data, but the data is mostly qualitative which means the studied set of objects has a quality that can be used to group together or distinguish between objects. Surveys are marked by their subjective nature such as LIST CREATION or the EVALUATION of each member of a selected group. The most common qualitative tool is the QUESTIONAIRE. The questionnaire is given to a set of individuals who are asked a series of questions, the set of questioned designed around a subject or related set of subjects in order to elicit an OPINION or PREFERENCE. The predictive value of a survey is small unless the survey is restricted to responders who have the characteristics about which the questions are designed or the responder has unique knowledge about the subject of the survey. Poll is just another name for a survey.
The Best Team in College Sports?
Most college sports are organized into divisions based on college size and geographical location, the largest being the NCAA. Of the major sports "managed" by the NCAA, only Division I football does not have a championship playoff. This, at times, has caused much furor when, at the end of a season, more than one team has been able to claim it is the best. The NCAA attempted a solution to the championship issue by creating the Bowl Championship Series (BCS), but the selection process has come under question and the series has ended with undefeated teams with rightful claims to being #1. The only thing that the winner of the BCS championship game can lay claim is to being the BCS champion. Nothing assures the BCS champion of being the national champion or the best college team in the US.
The Logic behind a Claim
Claims are opinions. Opinions are subjective. Any team with one win has a claim. Any team without a win has no claim. Any team without a loss has a claim to being the best team. At the end of a season without a championship playoff, ALL teams without a loss have claims to the championship. How to resolve this conundrum?
There are 119 Division IA teams, soon to be 120. Ideally each team would play the other 118 teams and the team with the best record would be called the best, ties being resolved by head-to-head results or a head-to-head playoff for the time the transitivity of head-to-head fails to make one team stand out. Of course, this is fiscally, physically, and politically impossible although this case is the only time where the who-beat-who argument has any sort of validity.
The next best is solution would be to play one game against every team in your division and the champions of each division enter into a playoff. Any sort of invitational tournament is subject to politics or being twisted for financial reasons. In fact, this is why there is not a playoff system. The traditional bowls are afraid to lose status and the financial benefits that go with that status.
In Part 2: .The Logic behind a Poll and How BCS bracketing scheme works and how it can be improved.
Total Div IA Teams:.....................119 Total Div IA Teams playing this week:....97 (One game involving two Div IA teams cancelled because of Hurricane Ike) Total Games involving Div IA Teams:......54 (Minus the one cancelled game)
Div IA Teams playing non-IA teams:........8 (All games played at Div IA home)
Div IA Teams losing to non-Div IA team....0
Average Points per game:.................50 Highest Total:...........................86 (Nevada at Missouri[w] Lowest Total:.............................5 (Auburn[w] at Mississippi State)
Average spread of victory:...............21 Highest Spread:..........................65 (Alcorn State at Troy[w])
Lowest Spread:............................1 (two games) (Auburn[w] at Mississippi State & Memphis at Marshall[w])
Visiting Teams that won:.................16 Visiting Teams that won - Non-Div IA:.....0
Conf......W.....L ACC......20....11 Big 12...28.....4 Big East.10.....9 Big Ten..25.....6 C-USA....14....18 MAC......16....21 MWC......18.....8 Pac-10...14....14 SEC......26.....6 Sun Belt..7....14 WAC.......9....13 Ind.......3.....4
Interesting Facts:
Two of the AP Top 25 did not play in Week 3.
Coaches AP Poll Ranks 25 teams each week.
Number of Games involving Ranked Teams in the first 3 weeks: 61 (Minus cancelled games)
In those 61 games:
Average Points per game:.................54 Highest Total:...........................94 (Illinois(2) at Missouri(6)[w](Week 1) Lowest Total:.............................5 (Auburn(9)[w] at Mississippi State)(Week 3)
Average spread of victory:...............25 Highest Spread:..........................69 (Kansas(14)[w] at Louisiana Tech)(Week 2)
Lowest Spread:............................1 (Auburn(9)[w] at Mississippi State)(Week 3)
Ranked Teams that have lost: Week 1: Pittsburgh(25), Texas A&M(12), Clemson(9) to ranked Alabama(24), Virginia Tech(17), Tennessee(18), Illinois(20) to ranked Missouri(6) Week 2: South Carolina(24), West Virginia(8) Week 3: Arizona State(15), California(23), Ohio State(5) to ranked USC(1), Kansas(13) to ranked South Florida(19), Fresno State(21) to ranked Wisconsin(10)
Next Week, Week 4 and Top 25 analysis. Or, you could pose a question that could be answered in a statistical fashion...and I'll give it a try.
10. Your grocery check list should include the following: flank steaks, hamburger, Louisiana hot links, Italian sausage, jumbo shrimp, and chilled caviar for the visiting Swedish Bikini Team.
9. Hit Bev-Mo: Budweiser keg for the lowbrows, Pyramid Pale Ale for the nouveau riche-techno hippies, fizzy water for the guys who are not allowed to drink by court order, and a couple of bottles of the decent dry champagne for the Swedish Bikini Team. If Husker comes by, then hide the good dope and breakout the Santa Cruz ragweed. Vodka. Lots of vodka. Cheap tequila for the backyard boys. And, movie tickets for the neighbors if they don't want to come over and party.
8. Don’t worry about the wife or girlfriend as you have gotten her wasted the night before and she won’t wake up until after the games are over and you move the party to the local sports bar or to the hotel room of the Swedish Bikini Team.
7. A basket full of nerf footballs to throw at the screen whenever something comes on that you don’t like. Make sure that there's nothing in the basket other than nerf footballs.
6. Your trusty old football to caress in your hands to keep you focused? Aw, screw it, I can’t take my mind off of the Swedish Bikini Team.
5. Clean underwear. Spare no expense. They have to last you the season. Novelty boxers are out, unless they contain the team logo instead of hearts or polka dots.
4. Adorable cheerleaders outfits for the Swedish Bikini Team Who cares what team. Or, outfits from the Catholic girls school up the road.
3. Bookmark all the CFB web sites on the browser of your internet-ready cellular device. Make sure you sign up for all the text messaging services for by-quarter scoring. Don’t forget the NFL injury reports for the fantasy leagues. Get a separate line for your bookie so you can look at the scores on one phone while making bets on the other.
2. A workable alarm clock? Tons of uppers to carry you over from Friday to the Sunday evening NFL game. You can sleep at half time and a work on Monday. Oh, don’t forget to bookmark the site where you can put your name up for a liver transplant.
1. A bank of voice-activated computers with touch screens that can translate your slurred speech into coherent braggadocio and insults about the unmanly nature of your opponents who have hopefully lost.
