The Pi-Rate Ratings


About The PiRate Ratings

The PiRate Ratings are not 100% computer-tabulated.  Unlike some of the famous ratings that predict games through the use of complicated statistical computations, the PiRates are part statistical and part human judgment.  A standard statistical formula is used but the actual ratings are adjusted by interpreting game results.

Every year, I start with the final PiRate ratings from the year before.  Based on changes to each team, I adjust that rating and end up with an initial rating for the new season.  I break down several factors (such as blocking for the run and pass, run defense, pass rush, secondary coverage, defending at the goal line) on each team and attempt to determine how much better or worse that unit will be compared to the end of the previous season.  Once all 128 Division I-A teams have been computed, I then look at the number 64-ranked team for every factor (usually each factor from a unique team) and compare how many points better or worse each team’s exact opposite factor is to team number 64.  Say Kansas State has the 64th best pass rush.  I compare every other team‘s (including Kansas State) pass protection against the Wildcats’ pass rush.  Then, I assign each team a positive or negative number (or zero) based on how much better or worse that match up would be.  This is done for every factor.  Once every team and every factor has been computed, the sum of each team’s factors is added to or subtracted from 100, which is used as a par number.  If my math and spreadsheet are correct, the average rating for the 128 D1-A teams should always be 100.  Each team’s rating should show how many points better or worse they are than the average team.

When applying the ratings on a game-by-game basis, several more variables go into this process.  While you can take the difference in the ratings and factor in home field advantage (*) to get a close approximation to what my game prediction will be, I include several subtle factors that can alter the predicted outcome by up to 10 points.  Variables include, but are not limited to:

1. A team from a mild climate having to play on the road in a hot and humid climate in September.

2. A team that has something to prove or revenge for something that happened (like Michigan against Ohio State in 1969).

3. A team playing for a coach they love who’s on the hot seat.

4. A team quitting on a coach they do not care for.

5. A really good team that specializes in one factor that faces a mediocre team that just happens to excel in the one factor that can stop the one dimensional team or vice versa (this applies mostly to teams that run triple option or unorthodox schemes).

(*) There is no standard home field advantage.  See below for an explanation.

The major difference in the PiRates from most other ratings services comes with the weekly in-season updates.  The outcome of each game must be examined carefully to determine how to change each team’s ratings.  Here is how some of the many variables affect my ratings.  You may be surprised to learn how some of my “out-of-the-box” opinions differ from mainstream and accepted beliefs.

Schedule Difficulty: This year’s schedule means very little when determining my weekly ratings.  As an example, let’s say Southern Cal successfully recruited every future NFL First Round Draft Pick for four years.  It would be a given that USC would be several points better than even the nation’s number two team when comparing all the factors.  What if the rest of the Pac-12 was unusually weak, Notre Dame was in a major rebuilding project, and the other two out-of-conference opponents were headed to 1-11 seasons?  Would you say USC was not the best team because their schedule was so weak?  That’s pure garbage.  The best team is the best team regardless of schedule and not because they lucked into picking strong teams to play five years ago when the contracts were signed.  What if in 2001 Notre Dame had scheduled Illinois, Colorado, Iowa State, Miami of Florida, North Carolina St., Washington, Stanford, and Syracuse for the 2006 season?  Those teams were coming off 10-1, 10-2, 7-4, 11-0, 7-4, 8-3, 9-2, and 9-3 regular seasons.  That type of schedule would have been brutal for 2001 when the games would have been scheduled, but by 2006 those teams went a combined 27-69.  The so-called experts would discount the Irish for playing such a lousy schedule, when they actually attempted to have the toughest one.

The prior year’s schedule actually plays an important role in determining the beginning rating for the following year.  When I determine the ranking of each factor, the prior year’s schedule greatly affects how each team is ranked.  A secondary that gave up 60% completions and 200 passing yards per game may rate several spots above another secondary that surrendered 53% completions and 150 passing yards per game depending on which opponents each team played.  Compare facing Hawaii’s passing attack against facing Navy’s passing attack.  If a defense had given up 60% completions and 200 yards to Hawaii, they would have been celebrating a nice victory, but if they had surrendered that same amount to Navy, it would have been a sad Saturday night for them.

Weather: This has a definite effect on the ratings.  If a game is played in a monsoon and a 21-point favorite wins 10-6, the winning team’s rating won’t suffer like it would have if the game had been played in beautiful weather.

The Final Score: State U. beats Tech 41-20, but just how did they beat them?  What if State U led 41-0 in the second quarter and called off the dogs?  What if State U led 27-20 with two minutes to go, with Tech driving for the tying score when the QB threw weak into the flat leading to a quick interception for a touchdown followed by another just two plays later?  41-20 in the first instance could have easily been 76-0.  Tech could have actually won 28-27 in the second example.  Each game is studied to determine how the final score occurred.

Defensive Statistics:  So, you say your favorite team has the league’s top three leaders in tackles, hmm.  Is that really a good thing?  No way!!!  That probably means they have had to play defense too much.  It’s not the number of tackles that matters; it’s where those tackles are made.  A linebacker who makes 150 tackles may look like a star, but what if those tackles came five to seven yards downfield?  Wouldn’t 75 tackles be better if they all came at or near the line of scrimmage?  Thus, tackles for loss and quarterback sacks are much more important than total tackles.  I look at the box scores of the college games to determine where each tackler made his tackles.

Rushing Statistics: State U averages 4.5 yards per rush, while Tech averages just 4.0 yards per carry.  The two play in the same conference against basically the same opponents and have comparable offensive lines.  Is State’s ground attack automatically better than Tech’s?  There are variables that may show Tech to be the better rushing team.  Let’s say for every six rushes, State runs for 11, 9, 6, 1, 0, and 0 yards with no fumbles for their 4.5 average.  Tech runs for four yards with no fumble on every attempt.   You cannot stop a team that gains four yards on every play, but you can stop a team that gains 11, 9, 6, 1, 0, and 0 yards on six consecutive plays for after garnering two quick first downs, State will have to punt on fourth and nine.

Additionally, let’s say State gains a lot of yards rushing on draw plays that occur on third and 15.  Defenses will surrender eight yards willingly in that situation.  What if Tech finds itself in several third and two situations, frequently runs inside blast plays, and always gets the first down?  It’s not rushing average that counts, but whether the rush accomplishes the goal on that play.  Once again, I look for these outcomes in the box score play-by-play charts.

Passing Statistics: Similar to the rushing statistics, passing stats can be misleading.  One quarterback can complete 75% of his passes and be much less effective than another who completes just 45% of his passes.  Completing 15 of 20 passes for 180 yards may not be as beneficial as completing nine of 20 for 180 yards.  In the old American Football League of the 1960’s Oakland’s Daryle Lamonica and New York’s Joe Namath frequently put up stats that look pathetic compared to Eli Manning and Tom Brady today.  Yet, the old guys’ teams scored many more points per game than today’s stars.  A typical Lamonica box score might have been 15 of 32 for 345 yards and four touchdowns.  If Manning goes 22 of 32 for 230 yards and two touchdowns, is that superior to Lamonica’s effort?  I think not.  Not only did Lamonica produce more passing offense, his constant deep threat forced defenses to play much looser.  Thus, the running game had more seams.  Oakland could run off-tackle to the weak side and destroy teams that were forced to stay back in 4-deep coverage.  The same holds true in the college game today.  Baylor spreads the field horizontally and vertically, and it makes their backs much more effective than a team that throws three and five yard passes all day, hoping to get significant yardage after the completion.  Today, there are teams averaging eight yards per completion and less than five yards per attempt.  That’s not much of a passing attack, and it doesn’t open holes for the running game.

Special Teams: When there is a wide discrepancy in these factors, special teams can completely turn around the outcome of a game.  Let’s say team A is 14 points better than team B in all other aspects but one.  Now, let’s say team A has one weakness-defending kickoffs.  What if team B has just one strong factor-their kick return game?  That one factor can negate the entire 14-point advantage team A has in all other factors!  Two long kick returns that set up short field position can make this game a tossup.

Home Field Advantage:  Most other ratings you see on the Internet apply the same home field advantage for every college team.  This is ludicrous!  Nobody will ever convince me, and I doubt you as well, that LSU and Slippery Rock enjoy the same home field advantage.  With my ratings, it’s not just every team that enjoys a unique home field advantage, every game does.  When LSU hosts Alabama, the Tigers don’t receive the same advantage they would if they were hosting Boston College.  The Tide will have a couple dozen players who have already experienced this game.  Likewise, when Oregon hosts Oregon State, the home field advantage is entirely different from when Oregon hosts Syracuse.  Aside from probably never playing in Autzen Stadium before, the nearly 3,000 mile road trip across three time zones is more of a factor than when the Beavers bus down to Eugene from Corvallis.  As I have witnessed before with Vanderbilt University, the visiting team (Alabama, Kentucky, Tennessee, and Auburn) may enjoy more of an advantage than the home team because in some years their fans outnumber the home team’s fans.

New For 2015–Retrodictive Rankings: Beginning in September of 2015, the PiRate Ratings debuted retrodictive rankings.  This is our own version of how the teams have performed to date based on who they beat or lost to.  It ranks only on wins and losses and strength of schedule (road games count more than home games), sort of like the RPI in college basketball.  It is an experimental ranking at the moment and could be tweaked after the season when we backtest it.


  1. 🙂

    Comment by — March 8, 2008 @ 3:45 pm

  2. I followed your predictions to a tee in my March Madness bracket this year & I’m dead fuck’n last with no chance whatsoever….embarrassingly so. Not even fuck’n close! You might want to reconsider everything you know about college hoops…. The only reason why I’m not pissing on your page right now is that you won me $250 on Xavier because of the “35+ win by potential” (even though it was a nail biter till the end). So I guess I begrudgingly say thanks…

    Comment by Rod Valetine — March 18, 2016 @ 7:55 pm

  3. You are the man for cfb ratings. I used your ratings to crush my picks league last year.

    What do bias, mean and average mean in the context of your rankings? I know what these mean in general, just trying to see how I should interpret them here.


    Comment by Marc — September 20, 2016 @ 6:23 am

    • Thank you for your patronage, Marc. Please remember, that we beg you to never wager real cash on our picks. We worry about our E-friends not having enough left to pay the electric bill.

      The short answer about what our three ratings mean is:
      We use the same data to calculate our initial ratings of the season for college and NFL football. The regular PiRate Rating uses the data the same way it has been used for 20+ years and stays very similar to the data we first used more than 40 years ago. Only the algorithm has varied a little as passing efficiency has become more important than running efficiency on both sides of the ball.

      The Mean Rating takes this same information and gives every facet the same amount of contributing factors. It does not mean that every facet is equal–just that we give equal weight to each facet after we have applied our normal grading.This rating will deviate from the other two, which we like.

      The Bias rating uses the same exact data as the PiRate Rating, but the algorithm is biased with five criteria counting 2.5x the other criteria. It should always be similar to the PiRate Rating, but the deviation will be a bit more liberal.

      Comment by piratings — September 20, 2016 @ 7:14 am

  4. Do injuries factor into your formula?

    Comment by Jeff Mrytlebank — November 7, 2018 @ 2:12 pm

    • Injuries do factor into our ratings, but basically in order to move a game by as much as a point this late in the season, the injury has to be to a star quarterback like Trevor Lawrence or Gardner Minshew in college or Tom Brady or Aaron Rodgers in the NFL. At the beginning of a season, injuries to certain players will even flip games.

      We do not apply new injuries to the actual rating, but rather the home team advantage/visiting team disadvantage. An actual game must be played before the injury is reflected in the actual ratings.

      Comment by piratings — November 8, 2018 @ 6:22 am

  5. Did you post the predicted wins for Conference USA for 2019-2020? I don’t see them anywhere.

    Comment by William Bartoli — August 20, 2019 @ 2:34 am

  6. so a positive number is how much you think 1st team listed will win by and a negative number the opposite?

    Comment by Billy McDade — September 5, 2019 @ 6:15 am

    • Yes, Mr. McDade, your assumption is correct. If you see something like: Utah USC 2.4 -0.6 1.2, this means that the regular PiRate Rating favors Utah by 2.4, the Mean Rating favors USC by 0.6, and the Bias Rating favors Utah by 1.2

      Thanks for asking this question that we get asked on an annual basis, so that all new visitors that aren’t 100% sure can know for sure. Most importantly, realize that we NEVER use our ratings to wager real money with Nevada or offshore. We are just math nerds that used to be athletes and coaches.

      Comment by piratings — September 5, 2019 @ 6:44 am

      • Thanks so much, so for example if baltimore is -6.5 but Pi-rate predicts -1.5, one would lean miami

        Comment by Billy McDade — September 5, 2019 @ 7:41 am

  7. I predicted the death of disco years before it happened. As soon as FM radio started playing The Scorpions and Judas Priest, you knew it was over for disco. I guess you could compare disco to an expansion franchise that sells out for some quick short lived success. Or like the 1998 Tulane football team. They went 12-0 by not playing anybody. Disco filled the void between the end of the heyday of Motown R&B and the start of the hard rock and alternative rock craze and various sell-out disco influenced pop like Michael Jackson and Prince. Some of the less puke inducing disco can now be heard all these years later in TV commercials such as “Flashlight” or “Atomic Dog” by Parliament Funkadelics. Or “Love Rollercoaster” by the Ohio Players.

    Comment by Smith Jones — October 7, 2019 @ 4:16 pm

    • Mr. Smith Jones, I bet you were eating Smyrna Figs and watching old episodes of Lavergne and Shirley when you made this comment. Hey Stewart, the way we’ve been picking losers this year, we are up the Creek without a paddle.

      To all the people wondering what the hey this is all about, Mr. Smith Jones is one of our secret field operatives sleuthing for information of value for us to use. He has a special helper, his Dog Friday.

      Comment by piratings — October 7, 2019 @ 5:12 pm

  8. Just finding out about this site and I find it very interesting. Thanks for your hard work on the NCAA tourney. Can you explain the reason for Iowa over USC in the Sweet 16? Also WVU over SDSU in Round of 32? Thanks!

    Comment by hokieharry — March 16, 2021 @ 2:02 pm

    • The answer to those two games as well as all the picks in this publication were made by analyzing the Criteria for each team and then looking to see which team has a better overall resume in head-to-head games. Iowa’s offensive efficiency will be too much for USC’s very good but not great defense, while USC will have to win this game from the perimeter. Iowa’s perimeter defense is good enough to make that a less than 50% proposition.

      West Virginia gets the nod over SDSU in the Round of 32 due to superior schedule strength with comparable stats. WVU is more than 6 points better in schedule strength, which counters the Aztecs’ R+T advantage. WVU has the #11 offense, while SDSU has the #11 defense. Offense is worth a little more than defense, much like On-base percentage is worth a little more than slugging percentage in baseball. In the end, the Mountainers have an ever so slight inside advantage and would be expected to win by 4 or 5 points. It’s only like a 53-47% advantage.

      Comment by piratings — March 16, 2021 @ 3:04 pm

  9. Just insufferable.

    Comment by Jason — April 21, 2022 @ 5:01 am

  10. I noticed that your 2022 ratings for NCAA college basketball teams pre-tourney had some teams that were tied because they had exactly the same rating. Yet one had to be better than the other. What do you do with ties if anything? Could you perhaps use a few more decimal points so that we could determine which of two now tied teams is better than the other?

    Comment by Rick — October 16, 2022 @ 4:57 pm

    • Interesting observation, but my ratings only have one significant digit. Basically, the second digit would always be “0”. The algorithm used to make the preseason ratings of the teams makes one decimal the maximum certain digit. The update process is made in tenths of a point. In essence, there is no way to break a tie when two teams rate the same to tenths of a point.

      Comment by piratings — October 16, 2022 @ 5:23 pm

  11. Hello, do you have a prediction for the upcoming Montana v Montana state game?

    Comment by Peter — November 14, 2022 @ 8:13 pm

    • Hello Peter. The PiRate Ratings are used strictly for FBS teams. I estimate FBS vs. FCS games using my old formula from the 20th century. I don’t rate FCS teams for games against other FCS teams. All I can do is provide a rough estimate, which would be Montana by 4 1/2 points. This is only a rough estimate based on factors that I wouldn’t rely on for handicapping purposes. After all, I stopped using this method in 1998. I only use it for FBS vs. FCS games, because I don’t have access to the numbers I need to rate FCS teams with the current 3-rating system.

      Comment by piratings — November 14, 2022 @ 8:27 pm

  12. Do you have a twitter or anyway to contact you directly?

    Comment by Pete — November 21, 2022 @ 9:07 pm

  13. Hello, any prediction for the Jacksonville state bowl game?

    Comment by Pete — December 13, 2022 @ 4:35 pm

  14. good afternoon. i hope you are doing well. i wanted to ask if syracuse being heavily favored by your system was intentional. every other system has minnesota winning, so i wanted to make sure there wasn’t an error or anything. thank you.

    Comment by Stanford — December 19, 2022 @ 1:24 pm

    • Oh no! Minnesota should be favored by that amount. Thanks for catching the error. You are correct.

      Comment by piratings — December 19, 2022 @ 9:21 pm

  15. THANK YOU! I was looking for an email to send a sincere thank you to the Pi-Rates. Long overdue, but after the UCONN win this year in the tournament I was able to take home first prize in about 5 contests, and finished 2nd, 3rd, 5th, 6th, and 7th in my large pool (720+ entries). Just wanted to extend a sincere thank you for everything – I wouldnt be nearly as successful if it wasn’t for your data and analytics (and love of the game)!

    Thank you Pi-Rates.

    Comment by CP — April 18, 2023 @ 2:17 pm

    • Oh, that makes us so happy. We hope to continue our success with our bracketnomics system in 2024. We learned our lesson about the Big Ten. Until there is a paradigm shift in the league as a whole, and the teams learn the need to create instant offense to become “spurtable”, they are going to continue to disappoint in March.

      Comment by piratings — May 22, 2023 @ 8:14 am

RSS feed for comments on this post. TrackBack URI

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

Blog at

%d bloggers like this: