Predicting the NFL

We have heard, “Common sense is not that common.” In the National Football League (NFL), it is logical that if a team scores a lot of points, they probably will win. Conversely, if they score few points, they most likely will lose. Numerically, what is a lot? A few?

Most NFL fans, even avid ones, probably really do not know these numbers.

My extensive 2013-2019 NFL season analysis has defined the loss and win point thresholds for selected teams. It is pertinent to many, if not all, NFL teams. These statistics are fun to use when games are live, too!

In other words, what is an NFL team’s ability to win any football game? For instance, if they score X amount of points, will they win? Or lose? Calculating point thresholds is virtually guaranteed to predict whether they will win or lose.

The model’s accuracy is a whopping 99.4 percent ! It is statistically sustainable and predictable.

The amount of points is the only factor in this extremely strong model. The analysis was done for each team’s 2013-2019 (partial) seasons. No consideration is made for home-field advantage, opponents’ season records, day/night game, injured reserve lists, etc. These factors are statistically insignificant over the long term.

In the somewhat randomly –selected seven teams, the analysis used their 750 NFL games, running correlation to the win or loss outcome. (The selected teams were made from friends’ requests to run the numbers.)

A qualifier for this model. For the 2019-2020 NFL season, about 60 percent of all NFL teams’ scores are in the range of 14-29 points. This model is applicable for points in about 40 percent of games — less than 10 to 16 points scored (10 percent) and more than 22 to 33 points scored (30 percent) per game.

Here are the accuracy results along with the thresholds for the 2017 through 2019/2020 (partial) seasons. This is for about 45 regulation games per team and runs through December 20, 2019.

You can see some variation when comparing these seven teams’ low points (guaranteeing a loss) and high points (guaranteeing a win).

To interpret the table, the Buffalo Bills will lose 100 percent of the time if they score 12 or fewer points. Also, they will win 100 percent of the time if they score 32 points or more.

Using the table, the Bengals’ and 49ers’ accuracy is 97.8 percent. In just one game in their 2017 through the present season, they scored more points than their point win threshold. However, each team lost — 49ers (39 points) and Bengals (27 points). These were the only exceptions in the 7-team study. More about the 49ers later.

The lower the low point number, the better chance for a team to win. Moreover, the lower the high number, the higher the probability of winning any given game.

Let’s use this information as a game is progressing. If the Bills are ahead of their opponent 32 to 27 at the five-minute mark in the fourth quarter, statistically, they will win — based on their past performance. They just squeaked into the defined, high-score threshold. Regardless, the Bills will win!

If the Cowboys just scored 15 total points, late in any of their games’ fourth quarters, they will lose 100 percent of the time. Much to owner Jerry Jones’ disgust, interference, and angst. In the Sunday, December 15 game again the visiting Rams, they scored 44 points. Their model high score threshold is 32 points, showing they would win. And they did, making the model correct!

I did not showcase the Seattle Seahawks since their model is 100 percent. Also, previous Nugget articles discuss the Seahawks’ performance.

Just once, the 49ers scored more than the threshold 33 points needed for a guaranteed win — but lost. In the graph, this was in their third game of the 2017-2018 season, playing the visiting Rams. Since then, the model has been 100 percent correct for the

49ers.

The 49ers’ 2017-2018 season snapshot shows them losing 100 percent of the games when scoring 10 points or less. That means when they score a maximum of a field goal (3 points) and a touchdown (7 points) in a game, they lose. In common terms, their scoring performance was lackluster and their defense could not bail them out to win games.

This terrific NFL model is exact and robust! Fun to chat with other fans and use during games to predict win/loss outcome.

In any of these named seven teams’ games, this commonsense model works. This tool logically applies to all NFL teams.

It strongly, simply, and uncommonly defines score-thresholds – all but guaranteeing each game’s outcome – win or loss.

Common sense numerically defined.

Send feedback/comments to [email protected]