Welcome back to the PiRate Ratings. Many of you have read some of our preseason previews, but by looking at the increased volume to this site in the last 24 hours, we can tell that we have a lot of new readers today.
Yesterday, we posted the spreads for FBS teams only for our three ratings–PiRate, Mean, and Bias. For those new to this site, a brief explanation follows.
Our ratings are unique in that we do not rely on scores alone to update our ratings. Most of us here are sports metric statistics lunatics. Our head man actually works in professional baseball as a “Moneyball” statistician/scout.
We use advanced statistics for each game to come up with the “theoretical score” of the game rather than the actual score, and then we update on our theoretical final score. For example, if State beats Tech 42-21, this 21-point spread tells us very little. What if State led 35-0 midway into the second quarter, and they pulled their starters after going five for five in touchdown drives? What if Tech then scored twice in the final 7 minutes of a 42-7 game? On the other hand, what if State led 28-21 with 7 minutes to go in the game, and Tech had driven 70 yards to the State 2 yard line, before fumbling at the goal with State returning the ball 100 yards for a TD, and then State added a second TD on an interception return with Tech driving again?
The 42-21 score is the only thing these two examples have in common. In the first instance, State might have won 63-0 if they had continued to use their starters and top backups; State would win 100 out of 100 times against Tech. In the second instance, there is a good chance that Tech might beat State 5 times out of 10. We carefully peruse the play-by-play and statistics of every college football game among FBS teams.
Our three ratings use the same type of data, but we have three different algorithms to come to the actual number. The PiRate Rating is the same algorithm in use for the last 30+ years. The Mean Rating is just that; it takes the mean of all our variables with no bias. Of course, the Bias Rating puts a bias on some data at the expense of other data. Because it is similar to the PiRate Rating, these two will have a much higher correlation than they do to the Mean Rating.
Okay, now for something completely different. Yesterday, we revealed our spreads for FBS vs. FBS Week One games. Today, we show you our PiRate Spreads for FBS vs. FCS teams for Week One. For reasons that involve how our ratings are calculated, we cannot supply Mean or Bias spreads with FCS teams, as it would take maybe 20 additional people to train and work with us. The FBS vs. FCS PiRate Ratings are purely mechanical, so they are just an approximation of our actual PiRate Ratings.
We have repeated our FBS vs. FBS games so you will not have to look at yesterday’s entry.
|This Week’s Games|
|Thursday, September 3|
|North Carolina (N)||South Carolina||3.6||6.4||4.0|
|Central Florida||Florida Int’l||8.4||11.2||8.4|
|Central Michigan||Oklahoma St.||-30.3||-23.5||-29.0|
|Minnesota||T C U||-17.5||-5.4||-19.1|
|Arizona||U T S A||49.3||31.3||48.6|
|Friday, September 4||PiRate||Mean||Bias|
|Western Michigan||Michigan St.||-23.5||-15.5||-22.9|
|S M U||Baylor||-40.2||-29.9||-41.8|
|Saturday, September 5||PiRate||Mean||Bias|
|Eastern Michigan||Old Dominion||-6.9||-2.8||-6.2|
|Nebraska||B Y U||8.2||5.8||6.2|
|Arkansas||U T E P||42.3||30.9||42.0|
|U C L A||Virginia||27.2||22.4||26.4|
|Tennessee (N)||Bowling Green||32.2||20.4||29.8|
|N. Carolina St.||Troy||40.5||44.0||39.0|
|Texas A&M (N)||Arizona St.||-0.8||2.6||-2.3|
|West Virginia||Georgia Southern||33.9||25.3||32.9|
|Florida||N. Mexico St.||39.6||33.3||38.6|
|Northern Illinois||U N L V||18.2||18.6||18.8|
|Florida St.||Texas St.||33.4||28.1||31.6|
|Southern Miss.||Mississippi St.||-25.4||-20.7||-27.1|
|U S C||Arkansas St.||36.8||31.0||40.9|
|Sunday, September 6||PiRate||Mean||Bias|
|Monday, September 7||PiRate||Mean||Bias|
|Virginia Tech||Ohio St.||-15.0||-6.5||-16.1|
|FBS vs. FCS||Week 1||PiRate|
|Utah St.||S. Utah||37.0|
|Ball St.||V M I||28.0|
|Georgia Tech||Alcorn St.||44.0|
|San Jose St.||New Hampshire||1.0|
|Fresno St.||Abilene Christian||17.0|
|Oregon St.||Weber St.||31.0|
|Kansas||S. Dakota St.||-1.0|
|Washington St.||Portland St.||22.0|
|Texas Tech||Sam Houston||15.0|
|Air Force||Morgan St.||34.0|
|Colorado St.||Savannah St.||51.0|
|Missouri||S E M O||40.0|
|N. Mexico||Mississippi Valley||32.0|
|S. Florida||Florida A&M||21.0|
|Middle Tennessee||Jackson St.||28.0|
|Kansas St.||S. Dakota||36.0|
|L S U||McNeese St.||34.0|
|San Diego St.||San Diego||28.0|
|Iowa St.||N. Iowa||7.0|
Please see our sister site: http://www.piratings.webs.com for complete rankings of all 128 FBS and all 32 NFL teams.
Edit: Special Thanks to Reader Charles for catching mistakes that allowed us to correct on Thursday morning.
Selections Against The Spread
Your voices/emails have been heard. At our sister site, http://www.piratings.webs.com, we have received exactly 137 requests from you to bring back our selections against the spread. We are happy that some of you sports fans remember that great 2011-12 season in which our ratings finished at the top of the Prediction Tracker ratings against the spread for the NFL, while our college ratings have had multiple top 10 finishes.
Before reading further, please make sure you read the following paragraph in bold: He or She that uses these ratings as their lone source before wagering their house on this data might as well just sell their house now and at least have funds to move elsewhere. We NEVER, EVER use this data to wager on games. As analytics’ specialists, we understand that the one way to make money wagering in football is to be the book. Yes, there are the Billy Walters of the world, but if you were he, you wouldn’t be reading this blog. So, since you are not, be advised that wagering any amount of your hard-earned money on football must be considered in the same light as paying for something, because that is exactly what you will be doing: paying somebody else.
Okay, now here is how we will make our selections this year. First, we will isolate those games in which our Mean rating differs enough from the official line to make it a possibility. Second, of those possibilities in step one, the six of us will pick 5-10 games that we personally like against the spread. If 4 of us have the same game and nobody has the opposite pick in that game, then we go with that as one of our selections. Because we believe that it is best to pick an odd number of games to prevent a .500 record and a loss, we will eliminate a game if we have an even amount of games.
Here are our 5 selections for Week 1
|Central Michigan||Oklahoma St.||-24.5||Central Michigan|
|Florida||New Mexico St.||33.0||New Mexico St.|
|Texas A&M (N)||Arizona St.||3.5||Arizona St.|