Alliance of American Football (AAF) Sports Betting Picks: Week 3
Sports betting is a love affair with numbers; “key numbers,†“best of the number,†“my numbers,†“the opening numbers,†“pros bet numbers,†“rouge numbers,†etc., etc. Nobody is betting on whether or not an NFL team’s “Color Rush†jerseys will be difficult for color blind people to view. In England, and at offshore sites, one can place a myriad of non ‘field of play’ types of wagers (the U.S. Election, Brexit, the length of a Kardashian marriage). However, under Nevada regulation, (and presumably the rest of the United States, as sports betting expands) wagers must be placed on events that are settled ‘on the field.’ Therefore, here at BettingPros, we will always be about numbers — more specifically, power ratings, and market numbers.
In this piec,e we’ll examine how power ratings are constructed (for the AAF, specifically) and provide Week 3 picks for the AAF. In subsequent pieces, we will apply those ratings to that week’s AAF betting slate, in a search for value from those numbers.
My power ratings are purely analytical, and the model in its current iteration has been developed and tested over nine NFL seasons. As you can see in the BettingPros expert rankings, I hit 56% Against The Spread (ATS) this season. I’m pretty confident in my approach to building new power ratings from the ground up, in this new league.
The Foundation of the Power Ratings: Play-by-play Efficiency
You may be wondering how Power Ratings can be developed with only two games played per team. It’s a fair point, if we’re looking to derive a lot of insight from the results of those games — but we are deriving our insight from the 1,013 plays executed thus far in the season — an average of 63 plays per game for each franchise.
The average yards per play through the air and on the ground are easily computed with some basic boxscore math. However, that data alone, is not very informative or predictive. We can derive per-punt data as well, but does that have any predictive value? Does that tell us who the best teams are? No, not really.
We need to adjust the per-pass and per-run data for its relative importance. We know from Stanford PhD, Ed Feng’s regression analysis at The Power Rank, that about 70% of win predictability can be derived from passing efficiency, while running success is only 30% predictive of a team’s success.
With this in mind, we need to then weight our per-play data on offense and defense according to its relative importance. We then have a single weighted offensive number, and a single weighted defensive number. We can combine those two numbers to give us a singular weighted team per-play number, but the most important step comes next.
Weighted, Net Yards Per Plays (Offense minus Defense)
| Team | Net Yards Per Play |
| Arizona Hotshots | 3.90 |
| Atlanta Legends | -1.19 |
| Birmingham Iron | 0.99 |
| Memphis Express | -3.11 |
| Orlando Apollos | 4.16 |
| Salt Lake Stallions | -1.58 |
| San Antonio Commanders | -2.22 |
| San Diego Fleet | -0.58 |
Normalize the Number and Adjust for Schedule Strength
The numbers above are not yet actionable from a betting perspective. They must be normalized to a scale that works for power ratings/betting.
So using the following formula, we normalize on a scale of -8.0 (Xmin) to +8.0 (Xmax): Xnew = (X – Xmin)/(Xmax – Xmin).
The reason I chose an Xmin and Xmax of -8.0 to 8.0 is somewhat subjective, but it’s based on my nine years of doing power ratings. +/-8.0 are about 20% from the highest power rating and the lowest power rating I’ve ever seen in nine seasons. With this being a startup league, and there having been one single window for player acquisition, I am confident that the teams are more closely bound in talent, and the variance is going to be largely execution and coaching.
That does not mean, however, that all of the opponents are the same. This is the last step in creating the power ratings. I then adjust for schedule strength based upon the rankings of each of the offensive and defensive units that the teams have faced thus far.
That number is then also normalized, and applied as a percentage weighted factor to the schedule agnostic number to create the final power rating for each team. For this week, anyway.
| Team | Power Ratings |
| Orlando Apollos | 6.83 |
| Arizona Hotshots | 6.52 |
| Birmingham Iron | 0.94 |
| San Diego Fleet | -2.02 |
| Atlanta Legends | -2.85 |
| Salt Lake Stallions | -3.40 |
| San Antonio Commanders | -4.59 |
| Memphis Express | -5.91 |
Apply the Power Ratings to the Week 3 Schedule
The week three schedule is below, courtesy of NoExtraPoints.com:
Each team will be assigned a Home Field Advantage Number of 2.0, except for the Alamo Dome, home to the Commanders in San Antonio, where their fans do give them an edge at key points in the game. They will get 2.5 points.
Arizona at Salt Lake (+4.5, 44)
LIKE: Arizona -4.5. I really like Arizona here (up to Arizona -6), and my numbers make the gap between these two teams much larger, at Arizona -7.9.
Memphis at Orlando (-15, 45)
No opinion: In a matchup of the team that I rank as the best and the team I rank as the worst, the market and I are exactly aligned. I make this game Orlando -14.7 to be precise.
Birmingham at Atlanta (+7, 38.5)
LIKE: Atlanta +7. Even though I have Birmingham as the third-rated team in the league, I do not have the difference between these two teams nearly as large. This line translates to Birmingham -9 on a neutral field? There are 2.2 net yards per play difference between these two teams. this is priced like there’s a 6.2 difference. Atlanta is sixth offensively, and sixth defensively, and the Iron are fifth and second. However, Atlanta’s numbers came against the number one, and number four teams in the league, where Birmingham’s came against teams ranked sixth and eighth. Take the seven points here at home vs. a team that’s averaging a paltry 2.3 yards per carry.
San Antonio at San Diego (-2.5, 43.5)
LIKE: San Diego -2.5. This game is priced as if these teams are nearly equal. My numbers don’t show that at all. San Antonio has a negative yards per play differential of more than two yards (-2.2), derived from giving up 8.7 yards per pass attempt, easily the worst in the league. While San Diego hasn’t exactly been the return of Dan Fouts and Don ‘Air†Coryell, they will find less resistance in the Commanders secondary (unless Mike Riley makes a massive overhaul this week). I like the Fleet to win by five points or so this week (my number is 4.6).
Right now I have no opinion on any of the game totals, but check back early Saturday, as we might make late plays, depending on where the lines go from here.
Jared Still in a featured writer for BettingPros. For more from Jared, check out his archive and follow him @jaredstill.