Be Flexible, and SMART With Power Ratings
Be Flexible, and SMART With Power Ratings
Many handicappers use Power Ratings or Computer Ratings to get a quick assessment of teams when evaluating college football. It’s important not to be blinded by those ratings when contemplating erratic teams who rarely play to their projections.
A great example of this phenomenon is the Baylor Bears of 2013. Under head coach Art Briles, Baylor is a high octane team that loves to run up the score when things are going well. The perceptions (and stats!) created when they’re running up the score can lead to false reads when Baylor is playing teams who are less likely to be bullied.
Reality in Numbers
Here’s the best way to visualize this reality in numbers. I grabbed the final 2013 computer ratings from Jeff Sagarin at USA Today, and will use those to create game-by-game “level of play” assessments for Baylor. The Bears finished the year at 89 in Sagarin’s numbers (rounding to the nearest number to make things simple). Yet, they rarely played anything like “an 89.”
Quick example to set up the full listing. Baylor opened the season against Wofford, who finished at 45 in Sagarin’s numbers. Baylor won that game by 66 points. Factoring in a basic 3 points for home field advantage, the Bears “graded” out at 108 for that game. On a neutral field, they would have been 63 points better than a 45…which is a 108-level performance.
Here’s Baylor’s full 13-game slate in schedule order…
[box] Baylor beat Wofford (45) by a score of 69-3. Level of Play: 108Baylor beat Buffalo (65) by a score of 70-13. Level of Play: 119
Baylor beat La. Monroe (57) by a score of 70-7. Level of Play: 117
Baylor beat W. Virginia (67) by a score of 73-42. Level of Play: 95
Baylor won at Kansas State (83) by a score of 35-25. Level of Play: 96
Baylor beat Iowa State (67) by a score of 71-7. Level of Play: 128
Baylor won at Kansas (58) by a score of 59-14. Level of Play: 106
Baylor beat Oklahoma (88) by a score of 41-12. Level of Play: 114
Baylor beat Texas Tech* (80) by a score of 63-34. Level of Play: 109
Baylor lost at Oklahoma State (88) by a score of 17-49. Level of Play: 59
Baylor won at TCU (74) by a score of 41-38. Level of Play: 80
Baylor beat Texas (79) by a score of 30-10. Level of Play: 96
Baylor lost to Central Florida* (82) by a score of 42-52. Level of Play: 72
(* represents neutral field games)[/box]
Baylor may have been an 89 in Sagarin…but you don’t’ see a lot of 89’s up there! In fact, there are only two games where Baylor was within a touchdown of 89. Only four games were within 10 points.
You see…Baylor wasn’t “an 89” when facing cupcakes. They ran up the score much more than teams of their own quality class normally do. Imagine two exactly even teams talent-wise. The head coach of Team A calls off the dogs at 52-7 in blowouts, but the head coach of Team B calls off the dogs at 70-7. Team B isn’t 18 points better. We said they were EVEN teams. They’ll play a toss-up against each other. Don’t let coaching preferences in blowouts warp your thinking about team quality.
A Closer Look
Let’s trim the list just to see what Baylor did vs. teams at 67 or worse in their Sagarin rating. Run your finger down the italics on the right side.
[box] Baylor beat Wofford (45) by a score of 69-3. Level of Play: 108Baylor beat Buffalo (65) by a score of 70-13. Level of Play: 119
Baylor beat La. Monroe (57) by a score of 70-7. Level of Play: 117
Baylor beat W. Virginia (67) by a score of 73-42. Level of Play: 95
Baylor beat Iowa State (67) by a score of 71-7. Level of Play: 128
Baylor won at Kansas (58) by a score of 59-14. Level of Play: 106 [/box]
In six games vs. lesser lights, Baylor bullied its way to a zillion points. Their average “level of play” was 112.2! An 89 should beat the teams in the mid 60’s by about four touchdowns when hosting. The Bears flew way past that.
Numbers Against Better Competition
Now, against teams who knew what they were doing…
[box] Baylor won at Kansas State (83) by a score of 35-25. Level of Play: 96Baylor beat Oklahoma (88) by a score of 41-12. Level of Play: 114
Baylor beat Texas Tech* (80) by a score of 63-34. Level of Play: 109
Baylor lost at Oklahoma State (88) by a score of 17-49. Level of Play: 59
Baylor won at TCU (74) by a score of 41-38. Level of Play: 80
Baylor beat Texas (79) by a score of 30-10. Level of Play: 96
Baylor lost to Central Florida* (82) by a score of 42-52. Level of Play: 72 [/box]
There were some stellar efforts in there, particularly the blowout of Oklahoma. But, the average “level of play” in the seven games is 89.4. “They were what they were” in composite against quality. Baylor was an 89 vs. quality, but would play to a 112 against cupcakes.
In pointspread terms, Baylor was 6-0 ATS against low-rated opponents who couldn’t stop them from running up the score. They were only 3-4 ATS against teams at 74 or better in Sagarin’s ratings (failing to cover vs. Kansas State, Oklahoma State, TCU, and Central Florida).
Baylor on the Road in 2013
Some of you may be wondering about Baylor’s vulnerability away from home. Here’s the short list of “games vs. quality away from Waco.”
[box] Baylor won at Kansas State (83) by a score of 35-25. Level of Play: 96Baylor beat Texas Tech* (80) by a score of 63-34. Level of Play: 109
Baylor lost at Oklahoma State (88) by a score of 17-49. Level of Play: 59
Baylor won at TCU (74) by a score of 41-38. Level of Play: 80
Baylor lost to Central Florida* (82) by a score of 42-52. Level of Play: 72 [/box]
That averages out to 83.2. The Bears were only 1-4 ATS in those five games (they were -17 at Kansas State).
The lesson is…if you’re going to keep your own Power Ratings by hand, or use math magic to “calculate” your own ratings…you need to account for the realities that come into play in college football.
Consider…
[box] Some powers run up the score vs. cupcakes, but others call off the dogs in much friendlier fashion. Don’t overrate bullies and underrate class.Pass-heavy teams in up-tempo attacks will play MUCH better vs. undermanned opponents than they will vs. opponents who can pressure the QB and play ball control on offense.
Run-based teams will play MUCH better vs. poor run defenses than they will against strong run defenses who can force a mediocre (or worse) quarterback to try and score in the air.
Some inexperienced teams will play MUCH better at home than on the road, more so than is typically accounted for with home field adjustments. [/box]
If you ONLY know a team’s Power Rating heading into a game, you’re still flying blind in a lot of ways if you don’t know the true skill sets of that team and their opponents. Knowing an average won’t help you pick pointspread winners for teams who rarely play to their average. Knowing the real-world on-field factors that cause teams to fluctuate wildly around their average will put you in much better position to beat the market.
Jeff Fogle is a freelance writer living in Austin, Texas. He writes about college and pro football, college and pro basketball, and MLB on his blog StatIntelligence. You can follow Jeff on Twitter @JeffFogle.
13 thoughts on “Be Flexible, and SMART With Power Ratings”
Leave a Reply
MAXIMIZE YOUR EDGE
Baylor never intentionally ran up the score. Baylor scored three points in the 4th quarter against Wofford; only 14 points in the second half against Buffalo; 0 points against ULM in the fourth quarter; and 7 in the fourth quarter against West Virginia.
Baylor was TRYING not to score late in these games. Had they tried to actually run up the score, they would’ve scored 100 points in all four of those blowouts.
For some reason Jeff couldn’t get his comment to show up. I tried to find it in WordPress, but it wasn’t there and when I tried to enter it manually it said that comment had already been posted.
Jack, here is Jeff’s response:
Jack, can we agree that Baylor is more likely than most other quality teams to run their point total up to around 70 before calling off the dogs vs. cupcakes…and that tendency can create illusions for handicappers who are compiling Power Ratings/Computer Ratings or using game stats?
– Jeff
TESTING!
Great article. I use power ratings just to find discrepancy from the line and then dig deeper into the matchups. I love hearing these nimrods from sites like Pregame talk about how “We don’t bet teams, we bet numbers” I cannot see how you can effectively handicap without knowing the teams
Here’s my question? Is Jeff Fogle a handicapper? Does he wager real money on college & Pro football and basketball games? If so how does he do?
I ask this question because when I read articles such as this article, which is well written, which provides a lot of good information and stats, but for me personally articles similar to this one carries a lot more weight coming from a writer that actually handicapps and wagers real money on games and has
history and a record of handicapping, which for me does not neccessarily have to be 54% or higher of success.
Glenjean, I’m a freelance writer living in Austin, Texas; where it isn’t legal to bet on sports. Article material will have to pass the merit/relevance test on its own.
How much weight does the pre season AP Power Ratings have? I am looking at Texas A&M at #8 (89.31) and South Carolina #13 (86.78) and to me that seems odd especially considering that South Carolina is now a -10.5 point favorite at home. I understand other statistics must be considered than rather simply looking at these two numbers.
So with that power rating in mind, these teams are almost evenly matched in a sense? Just curious, if someone can please elaborate on this, I would appreciate it.
Jeffrey, not familiar with the poll you’re referencing. Couldn’t find it in a Google search. I think you should dismiss ANY preseason poll that would show a 10-point market underdog as being the superior team in a matchup.
Fogle – Here is the link (you may have to copy and paste) let me know if I am looking at something incorrect. The headline is titled
“College Football 2014 Starting Ratings”
https://www.usatoday.com/sports/ncaaf/sagarin/
Jeffrey, I think you have to take some of these power ratings with a grain of salt this early in the season because there are so many unknowns. I have seen 3 different sets of power ratings that have A&M and S Carolina virtually as equals (ESPN FPI, Statfox, Sagarin) so its interesting that the line is double digits. I think you have to look at the situation here as A&M will be breaking in a new QB in his first start on the road, prime-time ESPN game in a very hostile environment. Nobody knows how good this QB will be and I think the line is saying that people are anticipating a drop off in performance at QB w/o Johnny Football. If you take a look at Phil Steele’s +/- power ratings from his magazine the line would be around 11/11.5 which is more in tune with the actual line. Steele has always been big on returning starters.
Thanks Jeffrey. Those are Sagarin,’s rather than “AP.” I would dismiss Sagarin’s preseason numbers. He has three teams better than Florida State…two of them almost five points better than FSU. And, as you point out, he has A&M rated higher than SC. Once he has some game results factored into the mix, they’ll better reflect team realities. OU wouldn’t be favored over FSU if they played this week on a neutral field.
Jack, I think the key word is “intentionally” when it comes to Briles running up the score. He may or may not intentionally do it, but the spread is always covered by the time he gives the hook to his first units. Much the same way that Mike Leach used to do for Tech’s home games. Last season Baylor had all of their home game spreads covered by halftime or a maybe few minutes into the 3rd quarter at worst. When it comes to the trends of a coach, this is the only thing that I’m concerned about when wagering these games. It doesn’t matter to me if they intentionally run up the score or not. The fact is they still covered the spread.
Sense this first week is strung out from thursday through monday, when will the Best Bets be posted? Numbers are starting to move a little in some spots. Central Florida seems to have attracted some sizeable money.
This move is surely not John Q. Public..