The Pi-Rate Ratings

August 16, 2020

College Football Update

Welcome back to the PiRate Ship.  We hope we have now plugged all the leaks onboard, and our vessel is seaworthy for the potential 9-month long, two-part football season.

As of this morning, 76 FBS football teams plan to play football in the Fall, and 54 FBS football teams hope to play in the Spring.

After working many hours more than in a typical season, the Bucs onboard our PiRate Ship have finally calculated the ratings for all 130 teams.

There is a fly in the ointment here.  Our annual ratings are based on a par score of 100.0.  In other words, if you total the ratings for the 130 teams, the PiRate, Mean, and Bias ratings should each total to 13,000.  How can the ratings adjust to par if 54 teams play no games every week until at least February, while 76 teams do play?  Should we then adjust our ratings so that the 76 teams playing total to 7,600?  Doing so, would require changing all the ratings, so that one group totaled 7,600 and another group totaled 5,400.

Ah, but there’s more to this than meets the eyes.  Most of these teams will play a conference only schedule.  How can a computer algorithm properly compare a team in the Atlantic Coast Conference with a team in the Southeastern Conference if there are no inter-conference games?

For many years, Major League Baseball had no inter-league play.  The best team in the American League and the best team in the National League had no common opponents.  The 1954 New York Giants won 13 fewer games than the 1954 Cleveland Indians.  The baseball media believed this World Series could be over in the minimum four games.  They were 100% correct that it took just four games to decide the Series, but the wrong team won the four games.  Other than the Yankees and White Sox, the rest of the AL was about as weak as it has ever been in a non-World War season.   The NL was six-deep in quality that year, and the 6th place St. Louis Cardinals were talented enough to be a first division team in the AL.

With 16 total teams, it may be somewhat possible to compare any eight with any other eight when the two pairs of  eight never play.  That is not possible with 130 or 76 or 54 teams.  Our ratings may need to be adjusted so that each conference averages 100.0.  But, then where would that leave Brigham Young, Army, and Liberty?  These three Independents plan to play some number of games.  How can we possibly make these three teams total 300.0 and then play games against other teams in leagues that average 100.0?

That is where we are at the present time.  Our conclusion today, thanks to the evil Covid-19, Corona, is that we will release separate preseason ratings, three different ways.

Our first set of PiRate, Mean, and Bias ratings will show you all 130 teams rated just like Covid never existed.

Our second set of PiRate, Mean, and Bias ratings will show you the 76 teams playing this Fall.  If any of the remaining 54 teams actually play in the Spring, we will be here to rate those teams the best way we know how.

Our third set of PiRate, Mean, and Bias ratings will show you the rating of teams within their own conference sampling.

Here’s an example, using a rating for State U.

Their normal preseason ratings might be

PiRate: 109.7

Mean: 108.9

Bias: 110.1

Let’s say State U plays in one of the 76 teams that plan to play this Fall.  After removing the ratings of the 54 teams not playing this fall, State U’s preseason ratings will no doubt change, because the exact average of those 54 teams will not be 100.0.  Let’s say that the 54 teams opting out due to Corona average 97.4.  The average of the remaining 76 teams is now 101.8.  Reducing that average to 100.0 might now adjust State U’s ratings to:

PiRate: 107.7

Mean:  106.9

Bias: 108.1

Let’s say that State U plays in the ACC, which will now have Notre Dame playing a full conference schedule this year.  The 15-team league will need to sum up to 1500.0 in all three ratings.  Normally, the ACC might have begun the preseason with a conference average of 110.0, or basically 10-points stronger than an average team.  Now State U’s ratings against a closed league of 15 ACC teams would have to be adjusted to:

PiRate: 99.7

Mean: 99.0

Bias: 100.1

So what ratings would we go with in this unique scenario?  We can submit the regular ratings before the first game is ever played, because nothing would be different in this regard.  Nobody has played a game, so the preseason ratings would be the same.  Definitely, we realize that with limited or no Spring Practice and with assumed 2-deep Depth Charts considerably less accurate than in past years, these ratings cannot be as accurate as past seasons.

If we wait until all 76 teams’ schedules are in place and then adjust the ratings so that every team will be rated on the same scale, then what happens when on Tuesday of a game week, one team has to quarantine and schedule somebody not rated on the same scale?  For instance, let’s say Florida State is scheduled to play North Carolina State, and the Wolf Pack has to cancel due to 20 positive cases.  What if at the same time Florida’s scheduled game with Vanderbilt is also postponed, because the Commodores must cancel.  Might a Florida State versus Florida game be scheduled on the quick?

Our final conclusion is to go with a 76-team ratings’ release that sum as close to 7600 for each rating and then to go with a different 54-team ratings’ release in the Spring, assuming the Spring actually does feature those 54 teams.

Monday afternoon Eastern Time, we will reveal our raw preseason ratings for the entire 130 teams and then separate ratings for the 76-team field, as well as ratings for teams by each conference.  We will then go with the 76-team format for the Fall season.

There will be no conference previews.  We will release ratings and pointspreads for weekly games, hoping that we can keep up with whatever alterations are made to the schedule.  We are sailing into the Corona Triangle this year.  Let’s hope we can get through it to the other side.

Note: In our few hours of spare time, we have been working on a tabletop football game with NFL teams and with college teams.  If you remember the old Avalon Hill/Sports Illustrated games that once existed, our games will be similar but different.  If we ever determine them to be marketable, you will not need the specialized 10-39 dice used by AH/SI.  Regular, 6-sided dice will work.  Hopefully, we will have the first set available to be purchased at a ridiculously cheap price with the chance to purchase future sets at additionally ridiculously cheap prices.  By sending out the charts in PDF form for you to print, we can sell the game for less than $10.  That is if we ever get any of the sets finished.  We have been working on three different sets–a 32-team Best of NFL by franchise from 2000 to 2019; the same best of for each SEC team between 2000 & 2019; and the best of for each Big Ten team between 2000 & 2019.  In the works in the future will be best of teams from each Major Conference between 2000 & 2019, as well as the best of NFL (and AFL) from 1960 to 1979, and from 1940 to 1959.

Blog at WordPress.com.

%d bloggers like this: