The Algorithm Predicts the NFL Playoffs and SB 50

January 16, 2016 1:08 pm Published by

As a fun offseason exercise this season, we are crunching the numbers on the playoff matchups to get an idea of what to expect from the game and to see if fantasy ratings can help predict game winners and winners against the spread.

After correctly predicting the winners of the last two Super Bowls, this is our first look at the playoff matchups.

Specifically, individual offensive player matchup ratings and defense ratings for each team. Although, we are running fantasy football ratings, we are basically saying: “Who
btn_donateCC_LGwould win if one fantasy team had all the Seattle offense and the Seattle DST and the other had all Patriots and their DST?”


Last year, this lead to a prediction that the Patriots would win 28-19 against the Seahawks, after even Vegas called the matchup as close to a dead heat as any Super Bowl in recent history. Actual score: 28-24.

The approach here takes the QB/RB/WR/WR/TE/DST matchups for each team and totals up the “Weekly” rating produced by the algorithm for each.

Then, the system integrates “Expected” points for each team, telling us what Vegas expects to be the outcome of the game. These values are generated by taking the Vegas “Over/Under” and the point spread and running this formula for each team:

Vegas Expected Points: OU/2+SPREAD/2

Then, we run the algorithm to determine it’s own “Expected Points”.

Instead of using Vegas against the O/U, we tally the Weekly Ratings to give us the team expected to have the highest ratings, then apply the % higher to the O/U Vegas value.


…this gives us “Algorithm Expected Points Against O/U” and we compare those to the opponent to predict final score and “Algorithm Expected Spread“.


Let’s see what the fantasy numbers tell us for the playoffs and leading into the Super Bowl.



The algorithm was getting warmed up in the Wildcard Round. The scores above were produced by the algorithm as predicted “Expected Points” for each team.



Picking winners: 3 out of 4 winners predicted

Against the Spread: 2 out of 4 winners predicted





Picking winners: 4 out of 4 winners predicted

Against the Spread: 3 out of 4 winners predicted





Picking winners: 1 out of 2 winners predicted

Against the Spread: 0 out of 2 winners predicted



Here are the details on how the omatic arrived at it’s prediction:


CAR enters the game as consensus 5.5 point favorites in a game that Vegas has as a 45.5 over/under.

This translates to 24.75 expected points for CAR and 19.75 points for DEN.

The oMatic compares Player Ratings at every position, with CAR posting higher ratings for every position, in clouding kicker and Defense. This gives CAR a 41.33% “Rating Advantage” while only enjoying a 25.32% Vegas line advantage.

Adjusting the score for this advantage projects an 18 point CAR margin of victory, a significant increase over the 5.5 point projection.

Thus oMatic predicts a CAR win and for CAR to cover the spread.

One interesting note is the difference between the quarterback Player Ratings:


Cam Newton finishes the season as the fantasyomatic algorithm’s #1 highest rated quarterback. Peyton Manning finished as the #38 rated quarterback…second worst all season with only Nick Foles finishing lower.


Categorised in: Fantasy News, Features

This post was written by Fantasy oMatic