Roadrager Rankings - ß4 - Tilted Forum Project Discussion Community

RoadRage · 10-08-2003, 08:10 PM

I know. I'm making changes about as fast as M$ puts out patches.

I finished the 2002 season Tuesday. http://members.cox.net/wd40/2002.htm

Problem: With ß3, I couldn't make Ohio State (14-0) #1 over Southern California (11-2). I fiddled with the W and L equations and the mu equation, and things only got WORSE. On a lark, I chunked the Omega (standardized output to a 12-game season, basically took the average result from each game and multiplied it by 12). Now Ohio State was it it's rightful place, the lone undefeated on top on the pack.

Applying this change to this season (http://members.cox.net/wd40/thisweek.htm), Army is #117, as any team who is shutout by TCU should be. In ß3, Army GAINED spots from being shutout, which was definitely unacceptable. TCU and Northern Illinois are down a bit though still in the Top 20. Minnesota leapt from #7 to #2, but that is solely due to being the only 6-win team so far. Averaging for a per game basis, they drop and Arkansas leaps back into #2, plus we're right back where we were for the 2002 season.

I don't foresee a ß5, at least until I get 2001 and 2000 historical data inputted, and maybe not until the full season is played out.

BTW, Texas is still #18.

The_Dude · 10-08-2003, 08:16 PM

Nice rankkings!(from last year that is)

I disagree about Notre Dame. They should be like 50 something.

other than that, 3 big 12 teams in top 10. nice. texas #6 is also cool.

RoadRage · 10-09-2003, 12:19 AM

Last year Norte Dame was 8-0 going into November, with wins against Maryland, Michigan, Pittsburgh, and Florida State. Before their loss to Boston College, they were BCS #3, right behind Oklahoma and Miami-FL. Everyone was talking about the "Return to Glory" and shit. Then Navy hung 23 on the vaunted defense and then everyone knew how to beat them, hence the current tailspin.

gov135 · 10-09-2003, 09:37 AM

This weeks rankings:
I'd hate to be one of the undefeated teams behind some of the one loss teams. But I understand that Va. Tech has played only four games. My question to you is: Will your current ranking system favor teams like Kansas State that play fifty games a year?

Debating where teams are is kinda stupid - they only way to measure the rating system you've come up with is to look at the results at the end of the season (before the bowls).

But, just for giggles: Let's see, from a MAC perspective Miami of Ohio (with one loss) is holding court over Northern Illinois, but not by a large margin. Then, there is a large drop (which there should be) to Bowling Green at number thirty-eight. Then Toledo. Everything looks pretty good. Game of the week #98 Ohio versus #94 Central Florida. Go Bobcats. If Ohio prevails, they should be ranked above Texas if the Longhorns fall. just my opinion.

RoadRage · 10-09-2003, 07:02 PM

There's gotta be a middle-ground between favoring those who play 14 games and those who settle for 11. I originally thought standardizing to 12 games would settle it, but it didn't get the correct #1 for last year. (Before the bowls, Southern California was ahead of Miami-Florida using the Omega and behind them using the Sigma.)

Maybe the problem is in the iterations. Too many? When to stop? Maybe it should be a logarithmic or exponential fit instead of linear. Maybe some kind aof sliding-scale factor to attempt to equalize is needed.

Maybe I should let the season play out and see what happens.

Ohio U above Texas?

Pass it around!

RoadRage · 10-10-2003, 05:36 AM

Two games last night, but for this purpose we'll analyze after the Kentucky-South Carolina game, in which South Carolina won by 6.

Top 10 gainers: South Carolina +3.77, UAB +0.64, Virginia +0.63, Georgia +0.53, UL-Lafayette +0.53, Tennessee +0.27; Duke, Oklahoma State, Clemson, and Minnesota all +0.12

Obviously South Carolina won, so they went up. The next five had all played South Carolina; so as South Carolina's fortunes improved, theirs did too, with the differences increases being due to margin of victory and the fact that Tennessee required overtime to settle.

On the next four, earlier this season: Duke was shutout by Virginia 27-0, Clemson had been shutout by Georgia 30-0, and UL-Lafayette had been beaten by Oklahoma State 56-3 and by Minnesota 48-14. There were other fractional gains, but these were the largest and most obvious of the second-order gains, teams who benefitted from the teams who themselves benefitted from the South Carolina win.

On the flip side: Kentucky -4.20, Indiana -0.66, Louisville -0.63, Alabama -0.53, Florida -0.47, South Florida -0.15, Washington -0.12, Connecticut -0.11, UTEP and Michigan State -0.10.

Without an overtime game, the difference between first-order and second-order gains is more apparent.

In light of this analysis, the overtime addends need to be changed back to the original idea, counted as 0-point wins or loses.

Full reversion back to ß2 will be in process today. All changes made in ß3 are being removed.

The moral of the story: tinkering with rules shouldn't be done mid-season.

Quick, someone call NASCAR and inform them of our finding.

10-08-2003, 08:10 PM	#1 (permalink)
RoadRage Stay off the sidewalk! Location: Oklahoma City, OK	Roadrager Rankings - ß4 I know. I'm making changes about as fast as M$ puts out patches. I finished the 2002 season Tuesday. http://members.cox.net/wd40/2002.htm Problem: With ß3, I couldn't make Ohio State (14-0) #1 over Southern California (11-2). I fiddled with the W and L equations and the mu equation, and things only got WORSE. On a lark, I chunked the Omega (standardized output to a 12-game season, basically took the average result from each game and multiplied it by 12). Now Ohio State was it it's rightful place, the lone undefeated on top on the pack. Applying this change to this season (http://members.cox.net/wd40/thisweek.htm), Army is #117, as any team who is shutout by TCU should be. In ß3, Army GAINED spots from being shutout, which was definitely unacceptable. TCU and Northern Illinois are down a bit though still in the Top 20. Minnesota leapt from #7 to #2, but that is solely due to being the only 6-win team so far. Averaging for a per game basis, they drop and Arkansas leaps back into #2, plus we're right back where we were for the 2002 season. I don't foresee a ß5, at least until I get 2001 and 2000 historical data inputted, and maybe not until the full season is played out. BTW, Texas is still #18. __________________ Join TFP Team SETI 43K workunits complete, 34 members, more of each needed.

10-08-2003, 08:16 PM	#2 (permalink)
The_Dude The GrandDaddy of them all! Location: Austin, TX	Nice rankkings!(from last year that is) I disagree about Notre Dame. They should be like 50 something. other than that, 3 big 12 teams in top 10. nice. texas #6 is also cool. __________________ "Luck is what happens when preparation meets opportunity." - Darrel K Royal

10-09-2003, 12:19 AM	#3 (permalink)
RoadRage Stay off the sidewalk! Location: Oklahoma City, OK	Last year Norte Dame was 8-0 going into November, with wins against Maryland, Michigan, Pittsburgh, and Florida State. Before their loss to Boston College, they were BCS #3, right behind Oklahoma and Miami-FL. Everyone was talking about the "Return to Glory" and shit. Then Navy hung 23 on the vaunted defense and then everyone knew how to beat them, hence the current tailspin. __________________ Join TFP Team SETI 43K workunits complete, 34 members, more of each needed.

10-09-2003, 07:02 PM	#5 (permalink)
RoadRage Stay off the sidewalk! Location: Oklahoma City, OK	There's gotta be a middle-ground between favoring those who play 14 games and those who settle for 11. I originally thought standardizing to 12 games would settle it, but it didn't get the correct #1 for last year. (Before the bowls, Southern California was ahead of Miami-Florida using the Omega and behind them using the Sigma.) Maybe the problem is in the iterations. Too many? When to stop? Maybe it should be a logarithmic or exponential fit instead of linear. Maybe some kind aof sliding-scale factor to attempt to equalize is needed. Maybe I should let the season play out and see what happens. Ohio U above Texas? Pass it around! __________________ Join TFP Team SETI 43K workunits complete, 34 members, more of each needed.

10-10-2003, 05:36 AM	#6 (permalink)
RoadRage Stay off the sidewalk! Location: Oklahoma City, OK	Two games last night, but for this purpose we'll analyze after the Kentucky-South Carolina game, in which South Carolina won by 6. Top 10 gainers: South Carolina +3.77, UAB +0.64, Virginia +0.63, Georgia +0.53, UL-Lafayette +0.53, Tennessee +0.27; Duke, Oklahoma State, Clemson, and Minnesota all +0.12 Obviously South Carolina won, so they went up. The next five had all played South Carolina; so as South Carolina's fortunes improved, theirs did too, with the differences increases being due to margin of victory and the fact that Tennessee required overtime to settle. On the next four, earlier this season: Duke was shutout by Virginia 27-0, Clemson had been shutout by Georgia 30-0, and UL-Lafayette had been beaten by Oklahoma State 56-3 and by Minnesota 48-14. There were other fractional gains, but these were the largest and most obvious of the second-order gains, teams who benefitted from the teams who themselves benefitted from the South Carolina win. On the flip side: Kentucky -4.20, Indiana -0.66, Louisville -0.63, Alabama -0.53, Florida -0.47, South Florida -0.15, Washington -0.12, Connecticut -0.11, UTEP and Michigan State -0.10. Without an overtime game, the difference between first-order and second-order gains is more apparent. In light of this analysis, the overtime addends need to be changed back to the original idea, counted as 0-point wins or loses. Full reversion back to ß2 will be in process today. All changes made in ß3 are being removed. The moral of the story: tinkering with rules shouldn't be done mid-season. Quick, someone call NASCAR and inform them of our finding. __________________ Join TFP Team SETI 43K workunits complete, 34 members, more of each needed.

10-09-2003, 09:37 AM	#4 (permalink)
gov135 Junkie Location: Midwest	This weeks rankings: I'd hate to be one of the undefeated teams behind some of the one loss teams. But I understand that Va. Tech has played only four games. My question to you is: Will your current ranking system favor teams like Kansas State that play fifty games a year? Debating where teams are is kinda stupid - they only way to measure the rating system you've come up with is to look at the results at the end of the season (before the bowls). But, just for giggles: Let's see, from a MAC perspective Miami of Ohio (with one loss) is holding court over Northern Illinois, but not by a large margin. Then, there is a large drop (which there should be) to Bowling Green at number thirty-eight. Then Toledo. Everything looks pretty good. Game of the week #98 Ohio versus #94 Central Florida. Go Bobcats. If Ohio prevails, they should be ranked above Texas if the Longhorns fall. just my opinion.