As sports bettors, the more useful tools we have in our handicapping toolbox, the greater our potential for success.
Crafting a solid set of power ratings is a daunting task for most college football handicappers. A host of valid questions stop many wannabe numbers makers before they even put pencil to pad: Where do I start? What numerical values do I assign the teams? Once the season starts, how do I adjust ratings?
I have been doing college football power ratings for about 30 years and still don’t have all the answers. The development of ratings — especially in the preseason, when you have a blank canvas, no data points and 130 teams to mix and mingle — is a tricky proposition.
Developing power ratings represents the true definition of subjectivity. Beauty is indeed often in the eye of the beholder when you’re weighing the number of returning starters, disparate levels of competition and coaching changes.
First, it is critical to understand your objective in making power ratings. Why are you doing it? What are your goals? How will you use your ratings?
For me, the answer is clear and simple: Power ratings are the starting point in developing my own betting line for every game each week of the season.
I spend the latter part of Saturday and much of Sunday updating my ratings while making betting lines on the upcoming week’s games.
Once Circa Sports book in Las Vegas or some other outfit posts the opening lines in college football on Sunday, I compare them with mine and look for discrepancies.
Let’s get into the nuts and bolts of creating your own power ratings.
— Assigning numerical values to 130 college football teams. I use a system that typically puts the top team at or near 100 at some point in the season. On the bottom end, the worst team is typically somewhere in the 30s.
In my preseason 2019 ratings, for example, the top teams were Alabama at 97 and Clemson 96.5. No other team was within a touchdown. The lowest-ranked team, Massachusetts, checked in at 36.
Sixty-two teams — or about half, at 47.7% — had starting power ratings between 50 and 65, a relatively high concentration in that range.
A seven-point range from 62.5 to 69.5 encompassed 30 teams, or 23%.
While I sometimes start my process by ranking each team’s likely starting quarterback on a scale of 1 to 6, with 6 being the highest, the ultimate power rating is not computed by adding numerical values assigned to position groups, coaching staffs and so forth.
Through years of compiling my own ratings, I know what a team with a power rating of 70 looks like in my subjective analysis. You have to develop your own style and process and be prepared for some likely period of trial and error.
The first 30 teams will be easy for most knowledgeable college football handicappers. It’s a mixture of the sport’s blue bloods and up-and-comers. Information on these upper-echelon programs is readily available through various “way too early” preseason rankings and other sources.
The bottom 30 teams are also fairly easy to identify. In much the same way that many of the really strong teams are good year after year, many of the really weak teams are bad on an annual basis.
As one might imagine, the Sun Belt and Mid-American conferences are well represented in the bottom 30. Conference USA and the Mountain West also have multiple entries.
In my 2019 preseason power ratings, I did not have a single Power 5 school in the bottom 30.
My final preseason product is the culmination of weeks of research, primarily consuming numerous articles in the electronic editions of local newspapers and other sources. Phil Steele’s preview and the other preseason football annuals have not hit the newsstands during the compilation of most of my information, so I’m independently perusing multiple sources one at a time. It’s not neatly packaged in one place, so a lot of legwork is needed.
— Making a betting line through utilization of power ratings and other factors. Power ratings are not the be-all, end-all in forming your projected point spread. They are a starting point, a piece of the puzzle.
To initiate the linemaking process, a handicapper might take the home team and add three or four points to its power ranking to arrive at a numerical value for that team. If the home team’s numerical value, including home-field advantage, is higher, it would be a home favorite. If lower, it would be a home dog.
For example, let’s say the home team’s numerical value is 10 points higher than the visiting team’s. While you might decide to indeed make the line 10, remember, this is a starting point, not an absolute.
What other factors should you consider in arriving at your number?
First, are the opponents familiar to one another? If they compete in the same conference or traditionally meet in non-conference play, they might have a great deal of familiarity, especially if the coaching staffs have remained fairly constant.
In general, familiarity should tighten the number, since the teams are more in tune with the other’s personnel and tendencies.
If they are conference rivals or regular non-conference combatants, you should have some point-spread history to examine. This might be extensive if the teams have played in the same conference for many years. Again, this factor gains relevance if the coaching staffs have remained mostly intact.
In this 10-point example, if the underdog has dominated the series recently from a point-spread perspective and the head coach has been the same for both programs during that span, your hypothetical point spread might be closer to seven, or perhaps even lower, than 10.
On the flip side, if the favored team has averaged more than 40 points a season the last three years and the underdog has a new, untested quarterback piloting an offense that has been sputtering in the early going, your number might creep upward to 13 or so.
— Adjusting power ratings, especially early in the season. Once games are being played and we have results, another tricky proposition arises: How do I adjust each team’s power ratings based on those outcomes?
In one 2019 season opener, I made Texas a 24.5-point home favorite over Louisiana Tech. As a rough overview, I arrived at this number because the Longhorns’ power rating was 21 points higher than the Bulldogs’ and 3.5 points were added since the game was in Austin.
Ultimately, Texas, actually favored by 20.5 points in Vegas at the close, easily handled Louisiana Tech 45-14. The Longhorns dominated from the get-go and were never challenged, leading 38-0 after three quarters.
Still, Texas benefited from three Louisiana Tech turnovers — the Longhorns were plus-2 in that category. And they outgained the Bulldogs by only 41 yards, although that was skewed by Louisiana Tech’s two meaningless fourth-quarter scoring drives.
I had Texas favored by 24.5 points, and it won by 31, a difference of 6.5. So how did I react?
After a careful examination of the box score and the flow of the game, I raised Texas’ power rating by one point and lowered Louisiana Tech’s by the same amount.
Some well-respected persons will argue that you should not adjust a team’s power rating by more than two or three points in any given week, even early when many unknowns remain. This opinion stems primarily from a belief that a handicapper should not overreact to a single result.
While I generally agree with this approach, I disagree in this specific instance.
Although I am confident in the depth and intensity of my spring and summer preparation, I’m careful not to fall in love with my preseason ratings. With the amount of personnel turnover in college football from year to year, many uncertainties cloud the beginning of a season. Even as a so-called expert, you cannot predict exactly what type of product each team will trot onto the field in late August or early September.
So once we start getting games and outcomes — data points, if you will — I believe this information should be weighted heavily. The first-time quarterback might actually be an upgrade over the previous multiyear starter, the team chemistry could have improved after the exodus of some entitled veteran players and the changeover in the coaching staff — especially among head coaches and coordinators — might have a positive or negative impact on performance.
Many of my approaches to the handicapping puzzle would be considered conservative, but this is not one of them. I will freely adjust a team’s power rating by four or five points — in some cases even more — in the first few weeks of the season. Preseason power ratings are merely an educated guess, but games provide more objective information to digest and apply.
In the case of Texas’ victory over Louisiana Tech, I made only minor adjustments, primarily because the teams are in different spheres of college football’s pecking order and I didn’t want to overreact to a single result when the principals were punching at different weights.
— The beat goes on: Maintaining your power ratings through the dog days of November and beyond. Using my approach, a team’s power rating can be moved several points based on a single outcome early in the season. But once the page turns to October and most teams have played a handful of games, there’s not as much volatility besides a key injury or some other significant event.
While mostly operating as an army of one, I believe it is still important to have some system of checks and balances, especially if you’re a newcomer to the art form.
To do this, I suggest using one or two other reputable, widely available sets of power ratings to judge your work and expose potential mistakes. I recommend this carefully, however, since most power ratings on the market have little value, in my opinion.
Jeff Sagarin, an MIT graduate who has been publishing power rankings for multiple college and professional sports for more than three decades, is probably the most recognizable figure in this field and puts out good numbers.
Another person who produces college football power ratings, Sonny Moore, is not nearly as well known but also puts out useful numbers.
Their power ratings are easy to find and available at no cost through a simple internet search. Both also feature strength-of-schedule ratings for all teams, which has some value after teams have played multiple games.
You might also look at their past power ratings to gain a better understanding of the numerical value they assessed to all 130 teams.
Developing power ratings is like most things in life: The harder you work, the better your result. Ultimately, you have to outwork the opposition (the house) to have any chance at overcoming the standard 11-to-10 price in betting sports.
It’s not easy, but it is certainly doable.