Most predictive nfl stats

Most predictive nfl stats DEFAULT

NFL Betting Strategy: The Stats That Matter Most to Our Experts

There are a bevy of NFL statistics at our finger tips, but which are the most important to consider when betting?

A panel of our top experts reveal the stats they use to inform their analysis and picks, from turnover margin to special teams and everything in-between.

Skip To:Matthew Freedman | Sean Koerner | Chris Raybon | Stuckey

Matthew Freedman

It&#;s far too simple to compare teams by looking at points scored and points against. While past points aren&#;t meaningless, they can be misleading. It’s far more instructive to analyze teams on the basis of yards accumulated and allowed on a per-play basis.

So after making some adjustments for strength of schedule, I look at yards per play much more than total points or points per play.

I also give more weight to passing offense and defense, since in today&#;s NFL most games are won or lost based on the ability to move the ball and prevent the other team from moving the ball aerially.

Finally, I give some extra (albeit subjective) weight to two factors: Explosive plays and unit-level strengths and weaknesses.

If an offense that tends to get yards in big chunks is facing a defense prone to big plays, I might like that matchup for the offense more than I otherwise would if it were facing a consistent defense that allowed the exact same yards per play.

Similarly, if the strength of an offense &#; let’s say its wide receivers &#; matches up directly with the weakness of a defense &#; in this case, its cornerbacks &#; then I am inclined to like the offense more than I would if it were facing a defense with better corners and an identical yardage average allowed.

Sean Koerner

It&#;s very easy to get lost in the weeds when it comes to the wealth of data we get week-to-week in the NFL. It&#;s important to never put too much weight into any one stat and it&#;s equally important to always take into account any context when it comes to season to date metrics.

That’s why for the purpose of this piece, I would say I take a backwards approach:

Turnover margin.

Turnovers, in general, are much more random and therefore luck-driven. They can have a huge impact on the game and, more importantly, have a huge impact on how the public perceives teams the following week. A team with a turnover margin of +2 or more could be a team to fade while a team with a turnover margin of -2 or worse could be a team to back the following week.

Again, no single stat should act as the be-all-end-all, but turnover margin is typically the first stat I use to explain the difference between my projected spreads and the market.

In New Jersey? Check out

Another angle I&#;m taking for each matchup is the strength and weakness of each team heading into a matchup. I spend most of my time setting up projections for each offensive player, and even defensive players. So by the end of each week, I have a good handle on how teams stack up against each other in terms of pass vs. rush.

Football Outsiders&#; offensive/defensive efficiency ratings are a good way to quickly compare how teams&#; offenses and defenses stack up against each other. If one team seems more equipped to take advantage of a specific matchup, it can usually uncover some hidden value on that team. More importantly, I think that this particular angle can have even more value when thinking about the matchup in terms of in-game betting. The dynamics of the matchup can change based on who is leading, shifting the expected rest-of-game spread and the total.

Chris Raybon

There are three stats I want to get into, so let&#;s take it one-by-one.

Early-Down Pass Success Rate

The reason early-down pass success rate tops the list for me is really a combination of the inherent predictive advantages of early (first/second) down performance vs. third/fourth-down performance, passing vs. rushing, and success rate vs. other per-play efficiency metrics like yards per play, Football Outsiders&#; DVOA, expected points, etc.

Success rate measures how effective a team is at getting first downs. On first and second down, a successful play is defined as picking up 40% of yardage for a first down, while on third and fourth down, only a conversion will be deemed a successful play.

Success rate essentially filters out the &#;noise&#; from per-play efficiency metrics that can be driven by down/distance/situation/outlier big plays. It shows us how confident we can be in a team&#;s ability to consistency put itself in position to prevent/allow points.

First and second down are a lot more predictive of future performance than third and fourth for two reasons:

  1. The offense&#;s play-calling options aren&#;t limited by distance needed for a first down, and thus the defense has to account for both the run and pass.
  2. The odds of moving the chains on third and fourth down are worse than 50% at any distance of more than 2 yards to go, and are cumulatively 33% when 3-plus yards are needed.

Ultimately, performance on first and second downs measure how efficient an offense is at either getting to third-and-short or avoiding third down entirely &#; in other words, how well an offense can keep its odds of maintaining possession long enough to score.

Passing is far more predictive of point differential than rushing because the average pass pay nets yards while the average rush nets , and there will be various points throughout the game where some combination of down, distance, time remaining and score will necessitate a pass. In fact, passing efficiency is the most predictive metric of score differential other than turnovers, which we know are almost impossible to predict in a vacuum, but l are much likely to occur on passes &#; where either an interception or fumble can occur &#; than on runs, where only a fumble is possible.

Add it all together and you get early-down pass success rate as the most important metric I look at in a matchup of two teams.

Pressure Rate

The odds of a successful passing play go way down under pressure, and the odds of a successful drive go way down with a sack, and pressure rate is better than sack rate at predicting future sacks.

Since certain quarterbacks are better than others at overcoming pressure, some offensive lines are better than others at preventing it, and some defenses are better than others at causing it, pressure rate is essentially my way of schedule-adjusting pass success rate.

Explosive Play Rate

The rate of explosive plays &#; generally defined as plus yard passing plays and plus yard rushing plays (though the cut-offs are arbitrary) &#; is essentially the high-variance big-play component of efficiency we’re trying to filter out by using success rate, etc. It shouldn&#;t be weighted as heavily as pure success rate or the like, but it still provides important context as far as range of outcomes scoring-wise, and I always want to be aware of it when betting totals.

Of course, this is factored in by some extent to the total, so many times you&#;re looking for value. In these cases, you&#;re essentially trying to find situations in which success rate and explosive play rate don’t match up due to past strength of schedule in the context of the current matchup, which can lead to incorrect perceptions by the market.

For example, offense generally factors in more heavily than defense in terms of predicting how a matchup will unfold, so sometimes the public incorrectly assumes that a defense that’s been allowing a low amount of points will have a substantial impact on a high-scoring offense. But if that defense has been getting by more on not allowing big plays against below-average offenses and has a middling success rate, they could have minimal impact on an offense that excels at both play-to-play success and at producing explosive plays.

By the same token, a high-scoring offense that&#;s middling in success rate but good at producing explosive plays could be in for a bigger downturn than usual when going against a defense that&#;s strong in both areas.


When I first sat down to think about this, I first thought about adjusted pace (invaluable when betting totals) as well as the importance of handicapping the matchups on both sides of the ball in the trenches with a variety of metrics such as adjusted sack rate, which also takes into account the impact a quarterback has in that department.

I also started to think about how important scheme is to me &#; types of man and zone defense, etc.

However, after seeing the very smart things my colleagues have already mentioned, it finally dawned on me that I have to mention special teams.

Field position matters, field goals swing covers and that hidden yardage can so often decide the outcome of a football game.

I personally set my own special teams power ratings for each team by looking at all of the obvious factors:

  • Field goal kicking
  • Net punting
  • Kickoffs
  • Coverage units
  • Return teams

A lot of times when a team is under- or over-performing relative to their talent and offensive/defensive statistics, special teams may be part of the explanation, along with penalties, coaching and/or turnover luck.

Not everyone wants to make their own special teams power ratings nor feels comfortable doing so, which is perfectly fine as there are stats you can reference online such as special teams DVOA. Or you can just manually compare special teams statistics such as net punting average, punts inside the 20 vs. touchbacks, field goal splits, etc.

At least giving it some kind of consideration is important &#; a complete mismatch in special teams may be the difference in making a play or passing.

I also use it for handicapping totals. For example, if I&#;m leaning toward an under and know I have a matchup of two teams with good punters and bad field goal kickers, that might just be the push I needed. I also think special teams performance serves as a proxy for measuring the effectiveness of a coaching staff. There&#;s a reason teams like the Ravens and Patriots usually have the top overall special teams units in the league.

It&#;s the small things that are so important in the game of football, and special teams is a great way to encapsulate how well a team values those things. Plus, field position and field goals matter.

How would you rate this article?


NFL Numbers and Statistics That Matter Most When Placing a Bet

NFL Betting Academy

NFL Betting Strategy Course

The National Football League is the most popular sport to bet on in the United States. There are all kinds of tips, strategies, and information available to bettors to aid in gaining the greatest edge possible. Just like the handicapping process for other sports and leagues, there are several key statistics and numbers to be aware of when betting on the NFL. This article will provide several examples of each and discuss why they are of significance to the NFL betting process.

Key Numbers in NFL Betting

Every sport has its own set of key numbers to pay specific attention to when shopping the lines and searching for potential wagering options. Key numbers refer to the most common margin(s) of victory. As the scoring systems for different sports differ, so do the key numbers.

By knowing the most common margins of victory in a sport, you can better predict outcomes and ultimately place bets around these numbers. It is worth noting that while they help predict individual game spread and totals bets, key numbers are not useful for other forms of sports betting, such as futures.

The sport of football’s unique scoring system affords a handful of key numbers to use when betting both NFL point spreads and totals.

NFL ATS Betting Numbers to Know

The following shows a list of the probabilities for each margin of victory between 0 and 21 points in NFL games dating back to according to Wizard of Odds. The major key numbers are expounded upon below.

NFL Scoring Margins Probabilities

Key Number: 3

A team earns 3 points for kicking a field goal, thus making it a number of prime importance for NFL bettors. How many games do we see where it comes down to a last-second field goal to decide the outcome? With the amount of relative parity in the NFL compared to other sports leagues, you are likely to see several close games in a single week and many more across the lifespan of an entire season.

Key Number: 7

The key number of 7 comes into being because it is the number of points a team earns when they score a touchdown (6 points) and then kick an extra point. While 7 is the second most common margin of victory in the NFL, the drop-off from the first key number of 3 is significant.

Key Number: 10

When you add 3 and 7 together, you get This is the third key number to note when betting NFL spreads. Many games see a team already up a touchdown tack on a field goal late to increase their lead to While 10 is the smallest and most common combination number to be aware of, note the slightly higher probabilities for margins that reflect other combination numbers such as 14 (a two-touchdown margin), 17 (two touchdowns plus a field goal) and 21 (three touchdowns).

NFL Totals Betting Numbers to Know

Given that totals wagering deals with higher numbers than ATS bets, it makes sense that the range of key numbers for NFL totals would be greater than those for the spread. According to the same Wizard of Odds research, eight different numbers have proven significant with a lot of NFL game results. The big four are 43, 44, 41, and The second level, if you will, are the numbers 51, 47, 40, and It should be noted that the final scores that might generate any one of these key totals numbers are not limited to any single combination.

NFL Key Numbers Betting Strategies

Being familiar with the key numbers in both NFL ATS and totals betting is helpful, but as a bettor, it is important to know how to translate this knowledge to your handicapping process. The following sections outline strategies for using these key numbers to your advantage.

Look for Key Numbers Amongst the Lines

It is important to always keep an eye out for key numbers showing up amongst NFL betting lines. They can show up as exact key numbers themselves or in the form of a small half-point variation up or down. For example, a common NFL point spread is , falling just under the key number of 3. Depending on the team you are looking to back ATS in a given game, this small half-point discrepancy, commonly referred to as a “hook”, can make all the difference in whether or not you place a bet, or on the size of the wager. Since nearly 25% of all NFL games end with 3 or 7-point margins, the hooks for these numbers are of particular importance.

Shop for Key Number Betting Opportunities

It is important to remember that bookmakers are aware of the key numbers for the NFL as well. As a result, they are often hesitant to move lines off of a key number. If you notice a point spread sitting firmly on an NFL key number of 3 or 7 at one book, it is worth your while to shop around at other sportsbooks. Doing so can lead to situations like the following. The odds shown are for the same game between the Green Bay Packers and Minnesota Vikings. Note the difference just off of the key number of three at one sportsbook compared to the other.

Bovada Packers-Vikings Line

DraftKings Packers-Vikings Line

Discrepancies in liability or the analysis of oddsmakers could lead to finding a hook on that same game elsewhere, as in the example above. If the hook is in favor of the side you wish to back, you will have found a much better opportunity than simply accepting the original line.

Key Numbers and Buying Points

One popular way to bet point spreads and totals involves buying points. This option affords bettors the ability to bet an adjusted version of the line on a game with increased juice. Take the following spread for a game between the Cleveland Browns and Pittsburgh Steelers for example:

Buying Points

Suppose a bettor likes the Steelers, but noticing the listed spread is on the north side of a field goal margin, he/she is hesitant to back them in a divisional rivalry game. By buying a point to adjust the line in the Steelers’ favor, they can now place a bet on Pittsburgh with a new spread of The Steelers could now win by a field goal margin of 3 points exactly and you would cash your ticket.


Buying points also comes into play with teaser bets. Different books offer different variations, but the most common type of NFL teaser is the 6-point teaser. This wager enables bettors to adjust the lines a full six points in the direction of their choosing. The catch is that they must tease two lines together and have both results come to fruition to win the bet. Looking for lines that can be teased through the NFL key numbers will help to maximize the teaser feature to the fullest. For more information on teaser bets, CLICK HERE.

NFL Betting Statistics to Know

Statistics are under the microscope in sports now more so than ever. There are stats for seemingly every aspect of every sport, and there are even adjusted versions of these stats to give a more accurate picture of a team or player’s performance. Thanks to this increased availability and accuracy, sports bettors have an ample volume of statistics at their disposal to assist in the handicapping process.

Stats are kept in the game of football at various levels, both for collective teams and individual players. The primary unit battles (offense vs. defense), as well as individual player tussles (wide receivers vs. cornerbacks, etc.), can all be measured, gauged, and predicted thanks to statistics and analytics. The following sections discuss several statistics to monitor closely when it comes to betting the NFL.

Yardage Statistics

When it comes to predicting future points scored and allowed by a certain team, it turns out that yardage statistics prove more reliable than previous scoring stats themselves. There are a whole host of yardage statistics that NFL bettors should track diligently, including but not limited to:

Offensive Yards Per Carry

Offensive Yards Per Pass Attempt

Offensive Starting Field Position

Defensive Yards Per Carry Allowed

Defensive Yards Per Pass Attempt Allowed

Defensive Starting Field Position

One other specific statistical area that deals with a team’s yardage gained or allowed is Explosive Play Rate. Looking for offenses that routinely strike for big gains and defenses that are prone to allowing chunk plays can prove particularly advantageous when these units are set to be matched up head to head in a game.

Defensive Points Allowed per Yards

The strength of a team’s defense has proven quite valuable in determining outcomes for both moneyline and spread betting over time. The defensive points allowed per yards statistic is a great indicator of just how tough it is to score on a given defense.

The stat is calculated by first dividing the total number of yards a defense has allowed by The answer is then divided into the total number of points a team has allowed. The resulting single-digit value provides a great statistic by which to compare several teams’ abilities to prevent opponents from scoring. The majority of teams will have DPA/ ratios between and A good defense will fall below and a bad one will be above

Turnover Margin

While there is an undeniable element of luck involved in turnovers, knowing how a team’s turnover margin stacks up relative to their opponent in a given week can help bettors rationalize inflated market lines that differ noticeably from their projections. Because turnovers can propel a team to victory or spell their doom in a given game, they often correlate directly with how the betting public perceives a team the following week.

Pressure Rate

While the battle in the trenches garners little spotlight or fanfare, the front lines are without a doubt the heart and soul of a football game. The pressure rate statistic is the best way to predict whether a future drive will be shut down by a crushing sack, even more so than sack rate. A defense with a strong pressure rate will prove especially effective on pass plays. Knowing which defenses are best at generating pressure, which offensive lines are best at preventing it, and which quarterbacks perform well or poorly under it are all helpful for NFL bettors.

Time of Possession

Time of possession is an important statistic to track on a couple of fronts. On the surface, it speaks to an offense’s ability to control a game. The underlying consequence stemming from this is how rested a team’s defense will be late in games. If the offense is controlling the time of possession, a defense will have a better chance of making plays late in a close game if they haven’t been on the field for extended periods of time. While it is a great way to draw comparisons between teams, making betting decisions based on time of possession alone is ill-advised.

Bettors should also be aware of how the time of possession data can become skewed. For instance, if a team plays several games where they are up by 20 points or more in the second half, they will tend to have a greater time of possession, thanks to both their efforts to bleed the clock and the opposing team’s attempts at preserving time to try and mount a comeback.

Football Outsiders Offensive and Defensive Efficiency

Combing through statistics can be a long and tedious process if you don’t know where to look. One particularly beneficial resource to utilize is the Football Outsiders efficiency rankings. The rankings tables are broken down in an easy-to-follow manner and offer a great way to quickly compare teams’ units against each other. They are adjusted with the company’s own DVOA (defense-adjusted Value Over Average) system.

As you can see from the following image of the defenses table, the overall rankings are broken down further into passing and rushing categories. Strength of schedule is also accounted for at the far-right hand side of the table.

Football Outsiders

Final Thoughts

Key numbers and statistics are both great tools for NFL bettors to use to take their handicapping to the next level. There are strong key numbers to use for both ATS and totals bets. Looking for these numbers and their hooks amongst posted lines can ensure that you get the best value on the bet you intend to make. Buying points is another way to take advantage of the key NFL betting numbers. In the vast sea of NFL statistics, several have proven particularly effective in assisting bettors. Focusing on the specific stats of yardage, team efficiency ratings, turnover margin, pressure rate and time of possession has proven particularly important to NFL handicapping. It is important to remember that key numbers and statistics should not be the only things driving a betting decision. But when working together with other tools and strategies, they can help to make a successful NFL betting venture that much more attainable.

NFL Betting Academy

Henry John
Henry JohnHankTimeSports
Henry is pursuing a Communication Studies degree and a Psychology minor at SUNY University at Buffalo. He has been a passionate sports fan from a young age and got hisstart writing about Fantasy Football. In an effort to combine an aspect of specialization along with his enjoyment of any and all sports, Henry expanded to cover other Fantasy Sports while also foraying into the Sports Betting sector. He continues to relish every opportunity to learn, grow and network within the industry. He counts the New York Jets, Toronto Raptors, Vegas Golden Knights and San Diego Padres among his team loyalties. In addition to school and sports, Henry is passionate about health and fitness and is currently studying to earn certification as a personal trainer.
  1. Fallout 3 fanart
  2. Exmark mower oil filter
  3. 25mm inner diameter bearing
  4. Dashbot dmo
  5. 1 dollar singapore coin

Improving ’s Famous NFL Prediction Model

You should now have a higher quality understanding of how the FiveThirtyEight model works. It is a simple but effective model to predict the outcome of games, win probabilities, and their point spreads.

However, the model can be improved even further, and improve it I did! I have found other adjustments that the people at FiveThirtyEight did not and ended up improving the model greatly.

When trying to figure out how to improve FiveThirtyEight’s model, I had to think about what aspects of the game did they not adjust for. As we discussed before they already have adjustments for: home field, bye weeks, playoff favorites, travel distance, and changes at starting quarterback. But diving deeper into the game of football I was able to find 3 new adjustments that heavily improved the model.

1. What Day a Team Plays On

Kirk Cousins, the quarterback of the Minnesota Vikings is the inspiration for this adjustment. Cousins is famous for playing terribly in prime time games (Monday night. Thursday night, and Sunday night games). In fact, his record in these games is 6–13 despite playing on an above average team.

This got me thinking, in prime time games on national television when the whole world is watching, do teams play better or worse then when playing on a normal Sunday afternoon. Well when diving into the stats over the last 10 years this is what I found…

As you can see from the table above in Thursday and Sunday Night games, the home team has a greater chance of winning by nearly 5% more. These teams also increase their average margin of victory in these games by around 1 point on Thursday night and about points on Sunday night, thus proving at home, teams typically play better in prime time games. Therefore, the adjustment that is added to our improved model is to add 25 Elo points to the home team on Thursday night games, and add 50 Elo points to the home team on Sunday night games.

2. Divisional Games

In each division in the NFL there are 4 teams. These teams play each other twice in the regular season. Since teams in the same division see each other more often than other teams, they get the advantage of playing them once, adjusting for them, then playing them again. This would ensue that divisional games should be much closer than non divisional games since these teams have already seen how the other plays, and they can prepare for it.

The table below shows the difference between divisional games and non-divisional games over the last 10 years.

As seen from the table above, in divisional games the home team has a lower chance to win than non-divisional games by nearly 3%. Their average margin of victory also decreases by a point when playing a team in their division, proving that divisional games are much closer and harder to predict then non-divisional games. To adjust for the closer divisional games, our improved model reduces the team with the higher amount of Elo points by 25 to minimize the overall difference in Elo between the two division teams.

3. Vegas Point Spreads

As I discussed before, FiveThirtyEight’s model is famous because it competes with the ones in Vegas. In fact when comparing their model’s predicted point spreads with Vegas’s predicted point spreads, in about 80% of games the difference between the two is a mere field goal (3 points or less). But the disadvantage that the Elo model has is that it’s not up to date like the Vegas model.

When Vegas rolls out their point spreads they are imputing every factor they can find; injuries, weather, last second starting change, etc. The Elo model doesn’t do this. Most of its adjustments you can find the info a week before. The only up to date adjustment is the quarterback adjustment that applies when there is a quarterback change, but for every other position an adjustment does not exist.

If in about 80% of games the difference between the two point spreads is about 3 points, then what happened with the other 20%? In my conclusion I believe there may have been an injury besides the quarterback position, or an unseen circumstance that the Elo model couldn’t adjust for, that may have occurred to make the difference between the two point spreads greater than 3. This is where our 3rd adjustment takes place.

Since our model doesn’t adjust for these scenarios, I decided to get a little help from the Vegas model. Therefore, if the difference between our model’s predicted point spread and Vegas’s predicted point spread is greater than 3, then we adjust our predicted point spread to match Vegas’s. For example, if our model predicts the Cowboys to win by 10, but Vegas predicts they win by 6, then we adjust the Elo’s for the two teams so our predicted point spread is then 6.

So like in the example above, Dallas’s Elo was and New York’s was The difference in Elo is and /25 = Our model predicts the point spread to be 10, but we want it to match Vegas at 6. Therefore, we reduce Dallas’s Elo from to

– = , /25 = 6, our now current predicted point spread (This also reduces Dallas’s predicted win probability from 80% to 70%).

was a bit of an unpredictable NFL season. This is backed by the fact that in both of FiveThirtyEight’s newest and older Elo model versions, both accuracy metrics (Brier Scores, Mean Squared Error) took a dip from its previous years. But how well did our improved model do in this unpredictable NFL season? By comparing all 3 models in both accuracy metrics, we can determine which model performed the best in the season.

Brier Scores:

If you remember the lower the Brier Score the better, and our improved model was able to lower the Brier Score significantly. In New Elo had an average Brier Score of , and our improved model had an average of , showing in our model was able to predict the win probabilities of team’s much better than both Elo versions. But it wasn’t just If you look at the graph above, over the last 5 years our improved model performed better than both Elo versions every year.

Using Brier Scores, let’s take a look at how each model’s accuracy evolved over time in the regular season. As you can see from the graph below in weeks 3, 4, and 9 our model did worse than the Elo versions. But for the rest, our improved model was able to outperform. This means in 14/17 weeks our improved model performed the best. This resulted in the decrease of the overall average Brier Score for the season.

Mean Squared Error:

Similarly to Brier Scores, our improved model out performed both Elo versions every year in reference to Mean Squared Error. In New Elo had an average MSE of , while our improved model had an average of , showing in our model was able to predict the point spreads of games much better than both Elo versions.

Greatest Hits and Misses

You’ve already seen how our improved model has done better than the Elo versions on a season and week to week basis. In theory, the next part is to show what specific games did our improved model increase in accuracy the most.

Down below are two charts. The first displays our improved models best picks relative to the newest Elo version, the second displays the worst.

(OLD column refers to FiveThirtyEight’s newest Elo models predicted winning percentage for the team that won. NEW column is our improved models predicted winning percentage for that same team).

Every year FiveThirtyEight host an NFL forecast competition to see if anyone can beat their Elo models in forecasting NFL games. Yearly, thousands sign up to prove that they know more about football than FiveThirtyEight and anyone else.

The game works by using Brier Scores. Every week of the NFL season (including playoffs) players decide which team will win by giving that team a winning percentage. After each NFL game finishes, players will either gain or lose points based on whether they picked the winning team and how confident they were that they would prevail. The higher a win probability you assign a team, the more points you can earn — but also the more you can lose if you were over confident.

To further demonstrate that our improved model can forecast NFL games better than FiveThirtyEight, I decided to participate in their game to see how many points my improved model would have accumulated if we indeed played in the season.

In over 15, players signed up to play FiveThirtyEight’s NFL forecast game. At the end of the season FiveThirtyEight’s model had accumulated points, good enough to be in the 97th percentile and in th place. This means their model was able to predict NFL games better than 97% of those that played. In comparison, Griffin Colaizzi who came in 1st place finished with 1, points and was in the 99th percentile.

Question is, was our improved model able to beat FiveThirtyEight’s score, and did it come even close to Griffin’s?

Down below is a weekly chart of the net points our improved model gained each week in the NFL forecast game.

In all but 2 weeks our improved model stayed in the positive for net points, and was able to finish the season with a total score of 1, points. Not only did we nearly double FiveThirtyEight’s points, but we blew Griffin’s total out of the water by gaining almost more points than he did. This means if our improved model had competed in the forecast game during the season, not only would we have come in 1st place out of nearly 15, competitors, but it wouldn’t even have been close.

The reason the Elo model is so great is because most of the info needed to predict games can be found months before when the schedule for next season comes out. Since we are days away from the start of the season, and we now understand how the Elo ratings update as the schedule plays out, we have the info we need and can predict what will happen over the entire season.

The NFL season isn’t always definite, teams that should typically blow out others teams sometimes lose. This infers there’s randomness in an NFL season. To integrate this randomness into our model, we use the Monte Carlo Method and simulate the NFL season , times, tracking how often each simulated universe yields a given outcome for each team.

In the end the simulations will show a team’s expected Elo points, point differential, full-season record and it’s odds of winning it’s division, making the playoffs, getting a first round bye, and even winning the Super Bowl.

Find you favorite team down below, and see what our model predicts will happen for your team next season.


Numbers that matter for predicting NFL win totals

To help figure out good value for potential NFL win totals in when placing a bet, it is important to look back at what really occurred in Several underlying metrics have historically been effective in projecting whether NFL teams are likely to improve or decline in the upcoming season.

It’s important to note that it is not advised to blindly rely on one or more of these statistics to affect a team’s win total in Instead, these stats, along with qualitative factors like roster turnover, schedule, etc. should help give guidance into which teams have a good likelihood of improving or declining.

Numbers that matter for predicting NFL win totals:


So many games are decided by turnovers, field goals, and one play here or there at the end of games. Statistics say that over a course of a season, teams should win about half of the games decided by eight points or fewer. Any extreme win/loss records in those games tend to revert near 50% the following season.

78 teams since have posted a winning percentage of 20% or below in one-score games. The following season, their record in one-score games more than doubled to 45%.

Since , 18 teams posted a win percentage of 30% or worse in one-score games. All 18 of those teams improved their close game record the following season for an aggregate record of 46%. Those 18 teams on average improved their overall win total by nearly three games the following season as well.

The four teams last season who won fewer than 30% of their one-score games include Atlanta, Carolina, Cincinnati, Jacksonville, and Houston.

In the last two seasons, 14 teams produced a win percentage of at least 70% in one-score games. In aggregate, those 14 teams had a close game win percentage of 80 percent and those teams saw their close game win percentages decline to 38% the following season. Those 14 teams had their win totals decline by an average of games the following season from wins to

The seven teams who produced a close game win percentage last season of more than 70% include Buffalo, Cleveland, Indianapolis, Kansas City, Pittsburgh, Seattle, and Tennessee.


Turnovers play such a huge role in determining games in the NFL that extreme margins one way or another could greatly impact a team’s season. Within turnover rates, fumble recoveries specifically are very random. The argument can be made that forcing fumbles is a skill but recovering the football is a purely random proposition.

Very often, teams that rank at the top or bottom five in the NFL in recovery rate will see their rankings move towards the middle of the pack the following season.

Looking back at the last seven seasons, teams that ranked in the bottom five in recovery rate have improved by an average of 15 spots the following season. Looking at the other side of the coin, teams in the top five of recovery rates have declined by an average of 12 spots the following season. In both extreme cases, it is expected that these teams will fall closer to the middle of the pack the following season.

Las Vegas, Cincinnati, Jacksonville, Detroit, and San Francisco finished at the bottom of fumble recovery rate and should be expected in the aggregate to rank near the middle of the pack this season.

Carolina, Philadelphia, Chicago, New England, and Arizona ranked as the best teams in fumble recovery rate in and should be expected in the aggregate to rank near the middle of the NFL in as well.


A team’s point differential can be a better measure of future wins than its actual win total. The number of wins a team should have won purely based on their point differential is based on their Pythagorean expectation.

Many times, teams with a win total much higher than their Pythagorean expectation will decline the following year. In , the five teams who posted the largest discrepancies between their point differential and their actual records included Oakland, New Orleans, Houston, Seattle, and Green Bay. These five teams saw their win totals decline by an average of two wins in and Houston went from a win playoff team to the third-worst record in the NFL.

The opposite is true for teams who underperform their Pythagorean expectation. In , the five teams whose point differential was better than their actual records the most included Dallas, LA Chargers, New York Giants, Detroit, and Cincinnati. 

These five teams saw their win totals increase by an average of wins. Both the Cowboys and Bengals lost their star quarterbacks early in the season or otherwise, this win total increase most likely would have been much greater. The prior season the five biggest underperformers increased their win totals by an average of nearly three wins. The New Orleans Saints and the Philadelphia Eagles missed the playoffs and underperformed their Pythagorean expectation wins by close to two games each. Both won the Super Bowl the following season.

Atlanta, Jacksonville, Houston, Carolina, and San Francisco underperformed their win total the most in while Tennessee, Seattle, Buffalo, Cleveland, and Kansas City overperformed the most.

Continued Reading for the NFL Season:

NFL Futures: Win Totals, Team Futures & Player Props
Best NFL Quarterback Player Prop UNDER Bets for
Best NFL Quarterback Player Prop OVER Bets for


Nfl most stats predictive

Next Gen Stats: New advanced metrics you NEED to know for the NFL season

We understand that modern fandom can be driven just as much by one's fantasy team as it is to allegiance to one's favorite NFL team. As such, the Next Gen Stats team is making it priority to deliver actionable insights and metrics to help you win your fantasy league (or even more importantly in some leagues, avoid the dreaded last-place punishment).

In fantasy football, opportunity is king. Snap counts, touches and targets are used as proxies for a player's involvement in the offense. More volume equals more fantasy points, as the old adage goes. But not all opportunities carry the same weight in most fantasy scoring formats. A carry at midfield yields considerably less value than a rush attempt from the opponent's goal line. Expected Fantasy Points, and derivative stats like Fantasy Points Over Expected (FPOE), are two metrics you'll be able to use to make in-season adjustments to help you win it all this season.

So how does the Next Gen Stats Expected Fantasy Points metric work? Our new all-encompassing fantasy metric is calculated using a combination of the outputs of several NGS machine learning models:

  1. Completion Probability: the likelihood of a completed pass.
  2. Expected YAC: how many yards after the catch will a receiver gain?
  3. Expected Rushing Yards: how many rushing yards is a rusher is expected to gain?

Take the difference between a player's actual fantasy points scored and expected fantasy points, and you get the aforementioned FPOE. A high FPOE value indicates a player who has significantly outperformed expectations in the past, but could regress if he is unable to maintain the same level of efficiency. Likewise, a player who is underperforming in expected fantasy points is a candidate to improve so long as the volume and value of the opportunity does not change.

Traditionally, quantifying fantasy opportunity has required multi-faceted analysis of metrics across a series of splits and situations. Expected Fantasy Points distills the value of those opportunities into a single metric that is even more actionable, especially given its scale is comparable across fantasy positions.

The NGS team is not done yet when it comes to unveiling new stats and features for the campaign. We have more advanced metrics to be announced in the coming days, weeks and months.

Projected 2021 NFL stat leaders

The NFL stats that matter most

You've heard the (perhaps apocryphal) quote from Bill Parcells before. "You are what your record says you are," has some element of truth behind it, yet the phrase fails to tell the entire story. If records were the best measure of future performance, we wouldn't see upsets like the Seahawks stunning the Saints during the playoffs.

In many cases, the simplest or most traditional statistic tells either an imperfect story or a fraction of the bigger picture. In trying to break down football games and understand which elements of performance correlate best with winning, I've come to rely on a toolbox of statistics and concepts that give me a better sense of what's actually happening on the field. Let's go through them and understand why they work (and where they come up short), starting with broader team metrics.

Team statistics

DVOA (Defense-adjusted Value Over Average)

DVOA was created by Aaron Schatz of Football Outsiders and serves as his site's core metric. The stat measures a team's success on a given play (through points and yards gained or lost) versus what would have been expected after accounting for the down, distance, game situation and quality of the opponent. The result is expressed in percentages, so a team with a DVOA of 10 percent is that much better than the league average on a play-by-play basis.

1 Related

The most helpful element of DVOA is that you can split it all kinds of different ways to figure out, say, a team's performance on offense in the red zone or their defense on third down. (The scale flips for defensive DVOA since you're trying to prevent the other team from scoring, so a DVOA of minus percent is better than a DVOA of 10 percent on that side of the ball.) DVOA also does a better job of correlating with winning in the future than a team's win-loss record itself.

You can read more about DVOA here. It can also be applied to players, but it's a far less effective metric for individual team members given the difficulty in comparing players across different schemes and styles. Individual DVOA has some limited uses, such as comparing running backs on the same team who play behind the same offensive line.

impact: The Texans made the playoffs at but finished a staggeringly low 29th in DVOA, sandwiched between the abysmal 49ers and Rams. On the other hand, the Eagles finished fourth in DVOA -- between the Falcons and Steelers -- but won only seven games, thanks in part to a tough schedule. DVOA would expect Houston's record to decline and Philadelphia's to improve next season.

Point differential

A team's point differential is also a better measure of future wins than its actual win total, a reality that holds true in many other major professional leagues. As an example, consider the 99 teams who finished between and The 51 teams with a point differential greater than or equal to zero won an average of games the following year. The 48 teams who posted a negative point differential won an average of contests the next season.

We can figure out how many games a team "should" have won in a given season based off their point differential by calculating their Pythagorean expectation, a metric invented by Bill James for baseball and applied to football by Daryl Morey. The latter figured it out for Stats Inc. before going on to run the Houston Rockets. The formula spits out a winning percentage, which fans can multiply by 16 to get an expected win total. More often than not, teams whose win total outstrips their Pythagorean expectation will decline the following year, as was the case with the Panthers and Broncos. The opposite is true for teams who underperform their Pythagorean expectation, which helped push the Cowboys, Giants and Titans toward winning records last season.

impact: The Raiders won 12 games but outscored their opponents by only 31 points, producing a Pythagorean expectation of wins. That gap -- wins -- is the fourth-largest since They're likely to decline. The Jaguars, meanwhile, went with the Pythagorean expectation of a win team. Jacksonville might not be great, but that win gap suggests the Jags should be looking up in

Record in close games

Closely related to the gap between a team's point differential and their actual record is how they perform in close contests. Historically, with precious few exceptions, teams will win games that are decided by seven points or less about 50 percent of the time. (I'm using seven points as opposed to eight to make it easier to compare teams across eras when the two-point conversion was not part of the NFL game.)

Evidence suggests that teams like the Bears, a squad that went in games decided by seven points or fewer, are extremely unlikely to keep that up year after year. The following year, those same Bears went in one-score games, with their overall record falling from to

To be clear, teams aren't "due" to decline and have a subpar record the following year; that's the gambler's fallacy. Teams with particularly good or bad marks during a year of one-score games are equally likely to be great or terrible in those games the following year. Our expectation is that they'll be average, which is what we call regression toward the mean.

impact: The Dolphins went in one-score games last season, with seven of their final eight wins coming by seven points or fewer. It's unlikely they'll be as effective in close games again. Meanwhile, the Chargers were in one-score contests. They're likely to improve, but so were the Chargers after the version of the team went in those same games. Even if teams with terrible records in one-score games might improve 90 percent of the time the following year, nothing is guaranteed in the NFL.

Quarterback statistics

Yards per attempt and adjusted net yards per attempt

The simplest individual metric with which to judge quarterbacks is yards per attempt (YPA), which shouldn't require much explanation. YPA correlates well with winning, but the complicated passer rating statistic is better.

Passer rating is built on an antiquated framework and doesn't fit the modern game, so if we're going to use raw data to create a complex quarterback stat, we might as well use one built more recently that boasts a stronger quantitative underpinning. Adjusted net yards per attempt (or ANY/A) uses more modern research by Chase Stuart to estimate the value of touchdowns and interceptions while also incorporating sacks, which evidence suggests has plenty to do with quarterbacks despite being commonly blamed on the offensive line. You can find out more about ANY/A here.

impact: Despite receiving praise for his hot start, Carson Wentz had a dismal rookie season by ANY/A, ranking between Blake Bortles and Case Keenum at 27th among qualifying passers. NFC East rival Kirk Cousins, meanwhile, finished fourth overall, ahead of Drew Brees and Aaron Rodgers.

The index statistics

One of the problems with comparing quarterbacks is accounting for the era in which they play. Right now, for example, we're in an era when both passing stats and scoring are at all-time highs. What passes for average in the modern game would've been deemed superstar numbers as recently as 25 years ago.

The indispensable (PFR) adjusts for era in several key metrics with their index statistics, such as Sack Rate+ (Sack Rate Index) or ANY/A+. PFR measures the number of standard deviations above or below the mean that a player accounts for in a particular category, and multiplies it by 15 to create the index stat. It's not a perfect methodology, but this does an excellent job of putting things in context in terms of key quarterback rate stats.

impact:Jared Goff was staggeringly bad as a rookie, posting the worst ANY/A+ since the AFL-NFL merger in among quarterbacks with passes or more. He narrowly beat out a group of terrifyingly awful rookie passers, including Ryan Leaf on the negative side and, more promisingly, Terry Bradshaw on the positive path. Nobody wants to start with a terrible campaign, but with a much-improved offensive line, Goff could still get better.

Total QBR

QBR, a metric developed by ESPN Stats & Information, incorporates several elements of quarterback play that aren't often accounted for in other quarterback metrics, including penalties and fumbles. It adjusts for context, giving far more credit for a 7-yard gain on third-and-6 than it does for the same yardage on third-and, because it's built on an expected points framework. It also uses evidence to divide credit for a play between a quarterback and his receiver, which makes sense on a fundamental level. When Dak Prescott tosses a screen pass 1 yard downfield to Ezekiel Elliott and the latter jukes four defenders out before taking it to the house, it's debatable whether Prescott deserves 10 percent or 15 percent of the credit for the play. It's far less plausible to suggest he deserves percent of the yardage.

I wouldn't suggest QBR is perfect, although its biggest problem previously -- the fact that it wasn't adjusted for the quality of the opposing defense -- has been fixed. If one passer has a QBR of 60 and another is at 55, I wouldn't use QBR to suggest one is definitively better than the other.

At the extremes, though, QBR is useful. If a quarterback is sixth in the league in QBR when he's not pressured but 29th in QBR when the defense is on him, I'm confident the game tape will back up the idea that he struggles more under pressure than most passers. If a quarterback is 10th in passer rating and 28th in QBR, I'm going to see whether there are mitigating factors that could be inflating his traditional stats. No measure is perfect, but QBR is the most effective one-number metric for quarterbacks dating back through

impact:Tyrod Taylor was far more effective as a quarterback by QBR than he was by popular perception last year, finishing ninth in the league with an opponent-adjusted Total QBR of The Bills can move on from Taylor after this season, so the QB may very well hit the market next year underrated by traditional metrics.

Running back statistics

Success rate

A Football Outsiders statistic that serves as a check on the efficiency implied by yards per carry, success rate measures the rate at which a rusher keeps his offense "on schedule." In most situations, a successful run picks up 40 percent of the needed yardage for a conversion on first down, 60 percent on second down, or percent on third/fourth down, with adjustments for game situation in the fourth quarter.

The strength of this stat is also its weakness: It penalizes players who rack up most of their yardage with a few big runs if they aren't also efficient. That sounds like it would hate a boom-or-bust back like Barry Sanders, but the stats suggest Sanders was more efficient than you remember, especially earlier in his career. Big plays are always nice, but unless you're Barry Sanders, it's far tougher to sustain those bigger plays from year to year.

impact: Jay Ajayi turned into a franchise back once the Dolphins gave him the starting job, as the second-year man averaged yards per carry, which was good for eighth in the league. Those numbers are buoyed by several big plays: Ajayi was the only back in football with four carries of 40 yards or more. Ajayi's success rate on runs was just 43 percent, which was 32nd among 42 qualifying backs.

Wide receiver/tight end statistics

Catch rate

One of the more basic statistics on this list, catch rate is simply the number of passes a receiver catches divided by the number of targets in his direction. Targets can be murky -- there are some passes that get arbitrarily assigned to a receiver even though they're not remotely catchable or get batted away before the receiver ever has a shot at catching the ball -- but overall, a receiver's catch rate is a worthwhile measure of efficiency. If two players each catch nine passes for 80 yards, the receiver who caught those nine passes on 10 targets is far more effective than the one who needed 17 targets.

Impact: Brandon Marshall saw his catch rate fall from percent in to percent last year, the worst figure in football for receivers with targets or more. Playing with Ryan Fitzpatrick didn't help matters, but then again, Fitz was playing quarterback in , too. He'll have to hope the presence of Eli Manning under center -- arguably the best quarterback Marshall has caught passes from during his decade-long career -- helps him turn that catch rate around.

Air yards per target

The other element of receiving -- one that influences catch rate greatly -- is the degree of difficulty on a player's reception attempts. A deep threat like DeSean Jackson can be wildly effective if he posts a catch rate of 55 percent, while an underneath wideout like Danny Amendola needs to be closer to 70 percent to justify his spot in the receiving rotation. The range of air yards per target for wide receivers varies from more than 16 yards per target (Jackson, J.J. Nelson) down to fewer than 6 yards per target (Anquan Boldin, Adam Humphries).

The classic example is Colts tight end Jack Doyle. Over the past three seasons, Doyle has caught percent of the passes thrown to him, the best figure in the league for a wide receiver or tight end. Not coincidentally, the average pass to Doyle has traveled fewer than 5 yards in the air, which was also the lowest figure in the league for any wide receiver or tight end by more than half a yard.

impact: The Raiders signed Cordarrelle Patterson, presumably to pitch in as a returner and help stretch the field on offense. By the time he finished his tenure with the Vikings, though, Patterson was almost exclusively a target on screen passes. The average pass to the speedy Patterson traveled just yards in the air last season, the lowest among wideouts by a comfortable margin. The second-lowest average among wideouts was the yard mark recorded by Humphries.

Receptions per route run

A measure of how integral a player is to a passing game, receptions per route run analyzes the frequency with which a receiver demands the football on the field. Receptions aren't created equal -- some players come onto the field for only designed passes in their direction, while others are catching checkdowns when the offense breaks down.

The leading reception rate among wideouts last year was the percent mark posted by Kansas City's Tyreek Hill. Theo Riddick trailed him, but led the way at running back with percent. The leading star wideout in this category is A.J. Green, who caught the ball on percent of his routes. Perennial rival Julio Jones was below him at percent. The top tight end? C.J. Fiedorowicz at percent. I didn't see that one coming, either.

impact: Hill has gone from being a third wideout and part-time offensive weapon for the Chiefs to the team's presumed top wide receiver this season. Can he continue to rack up receptions at a league-best rate as an every-down wide receiver this year?

Pass rusher statistics

Quarterback knockdowns

Sacks are the most meaningful statistic used to judge pass-rushers, but they're too few and far between to be our only gauge. The difference between a great season (12 sacks) and a solid, unremarkable campaign (eight sacks) is one sack per month. Judging players that way tends to be dangerous, which is why we generally discount stats like rushing and receiving touchdowns because of their year-to-year volatility.

Another way to judge a pass-rusher's effectiveness is the number of quarterback knockdowns (also called quarterback hits) he racks up in a given season. This number includes sacks (where the quarterback hits the turf), but doesn't include strip sacks (where the edge rusher bats the ball out of a quarterback's hands).

While the best pass-rusher in the league might make it to only 15 sacks, the league leader in quarterback knockdowns will often approach 35 hits. The knockdowns put J.J. Watt's dominance in perspective. When healthy, he puts even other great edge rushers to shame:

YearJ.J. Watt HitsRankSecond PlaceHits
431Cameron Wake33
461Robert Quinn34
511Carlos Dunlap28
501Aaron Donald37

Vic Beasley Jr. led the league in sacks last season with , but Aaron Donald topped the hit charts with More on Beasley and his prospects in a second.

Sacks per knockdown

While any pass-rusher getting to the quarterback is doing the right thing, the difference between a sack and a knockdown can come down to a fraction of a second or a lone step. Over the past five years, regular pass-rushers (guys with 10 or more hits in a given season) have turned about 43 percent of their knockdowns into sacks.

Players who have a disproportionately high or low percentage of sacks per knockdown are likely to see their sack total rise or fall accordingly the following year. On the low side, the obvious candidate to improve after was Jets defensive end Leonard Williams, who turned his 21 hits into just three sacks ( percent). Last year, he jumped from three sacks to seven and made his first Pro Bowl. The opposite example was Washington linebacker Preston Smith, who had eight sacks on 10 hits during his rookie year. Despite moving into a starting role last season, his sack total fell to five.

impact: Beasley is a prime candidate for regression this year. He racked up sacks on just 16 knockdowns, and while he had several strip sacks that wouldn't count as knockdowns, it's extremely likely that his sack total will fall back to earth in his third season. His percent sack-per-knockdown rate is the second-highest over the past five seasons. For comparison, Nick Perry had the second-highest rate in all the way down at percent.

It's impossible to produce a worse rate than Jihad Ward, who did not record a sack during his rookie season despite producing 10 knockdowns. He'll get his first sack in Datone Jones ( percent) and the wildly underrated Tom Johnson ( percent) also qualify. One more notable candidate is Lions star Ezekiel Ansah, who recorded just two sacks on 15 hits during an injury-plagued campaign.

Kicker statistics

Adjusted kicker stats

Football Outsiders tracks the efficiency of kickers, expressing them versus league-average in a given range after adjusting for the weather and altitude of the kick. The latter variable is critical, given how much easier it is to hit from distance in Colorado. The result is expressed in points above or below league-average. FO also tracks the same stats for punters, kickers and return men, though those are also far more subject to the abilities of the blocking units than the field goal kickers themselves.

impact: The worst kicker in football last year was Tampa Bay's Roberto Aguayo, who the team traded up for in the draft and was worth a league-low minus points last season, missing nine field goals and two extra points. The Bucs signed Nick Folk this offseason and gave him a $, signing bonus, suggesting Aguayo's time in Tampa might not last.

Hidden special teams statistics

Hidden football stats sounds like the secret menu at a restaurant, but it's an amalgamation of numbers tracked by FO. Their "hidden" special teams statistic consists of elements of special teams that matter but are out of the opposing team's control. The stat takes the opposing team's kickoff placement and punt distance into account, but most crucially, it accounts for the opposition's performance on field goal attempts.

FO expresses this metric in terms of points of field position, and the range is quite enormous. The luckiest team in the league last year was the Dolphins, who received points of "hidden" help. Meanwhile, the unluckiest team was Chicago, who lost points of field position from the opposition. That's a point swing. Indeed, despite Chicago's volatile weather conditions, opposing kickers connected on a league-best percent of their field goals against the Bears, while teams hit only percent of their field goals and percent of their extra points against Miami.

impact: You would expect the Dolphins to regress toward the mean, as teams haven't displayed much ability to hold on to these hidden benefits, but Miami doesn't appear to be budging. They've ranked in the top three of special-teams luck since and haven't ranked outside of the top seven since Indeed, since that season, opposing kickers have successfully converted a league-low percent of their field goal tries against the Dolphins. The Patriots ( percent) are the only other team in the league below 80 percent.

There's no evidence that teams can pull this off deliberately from year to year, so it's interesting to see what's happening with Miami. They've turned over most of their special teams personnel during that four-year stretch, but one exception has been special-teams coordinator Darren Rizzi, who has been on the books since Miami hasn't been especially impressive on special teams over that time frame, with an average rank in FO's metrics of 19th.

It's bizarre that the Dolphins would be middling at special teams on the whole, but great at this single, seemingly uncontrollable element of the game. Bruce Arians criticized Rizzi and the Dolphins for barking out snap counts before an extra point last season, though Arians has a history of complaining about special-teams plays. The snap count maneuver would be illegal, but it's hard to imagine the Dolphins executing such a tactic for the better part of a decade without being scolded by the league at some point.

It's tempting to credit Miami's fans for inducing misses, but opposing kickers have hit percent of their kicks against the Dolphins at home and a nearly identical percent of their tries at Miami over that time frame. The same fans were also around in the previous decade, however, and opposing kickers hit on a far more standard percent of their tries back then.

That's one of the fun things about pairing advanced statistics with football: Sometimes you stumble onto something important and seemingly meaningful -- and have absolutely no explanation for it.

© ESPN Enterprises, Inc. All rights reserved.

You will also be interested:


Throughout my season previews and my picks during the season, I refer to certain statistics as being either predictive, predictable, or not. I wanted to discuss what I mean by that and put all my supporting statistical evidence in one place. First, let’s define some terms.

Predictability is probably the easiest term to understand. A statistic is predictable if it can be easily predicted from one year to the next. This is measured by calculating the correlation between a team’s performance in a specific statistic in a specific season with its performance in that same statistic the following season.

Essentially, the higher the correlation the easier it is to predict how a team will perform in a certain statistic based solely on how they performed the season before. This is important because if a statistic can’t be reasonably predicted on a year-to-year basis, it doesn’t provide us much predictive value.

That leads into the second term, predictiveness. A statistic is predictive if it can be used to predict a team’s likelihood for winning. I am going to measure two types of winning percentage predictiveness, one measuring same-season win predictiveness and one measuring next season’s win predictiveness. 

Same-season win predictiveness is measuring the correlation between a certain statistic and a team’s winning percentage in the same season. For example, as a team averages more yards per play, their likelihood of winning goes up, but not at a perfect rate and likely not at the exact same rate as other statistics, which have their own statistical relationship with winning percentage. Measuring correlation allows us to see which statistics most closely vary with winning percentage. 

That being said, while same-season win predictiveness is definitely worth taking into account, it’s not a particularly useful stat for handicapping purposes because it only works with data from games that have already happened. Once I know how many yards a team gained in a game, I can give you a pretty good guess as to whether or not they won the game, but that isn’t all that useful.

Next season’s win predictiveness is really what we want because we want to be able to take last year’s statistics and use them to most effectively predict future winning. Rather than just measuring the correlation between a statistic and the same season’s winning percentage, we also want to measure the correlation between a statistic and the next season’s winning percentage to see how closely those variables relate.

If this isn’t making sense yet, hopefully it will when I get into some examples. Let’s start with a common one, turnover margin. All statistics included in this post are over a sample size of the past 10 seasons ().

StatisticYear to YearWinning %Next Year Winning %
TO Margin%%%

We all know intuitively that winning the turnover margin has a significant impact on winning, but this puts it into context. A team’s turnover margin correlates with same-season winning at close to a 70% rate. However, while it is predictive of same-season winning, it is highly unpredictable year-to-year, with a correlation of just about 11% year-to-year, meaning from a statistical standpoint, a team’s turnover margin almost might as well be random year-to-year. 

As a result, while turnover margin is predictive of same-season winning, it really isn’t predictive of next year’s winning percentage. I will break this down further later, but I wanted to use this as an example right off the bat.

Another good example is winning percentage itself.

StatisticYear to YearWinning %Next Year Winning %
Win %%%%

Winning percentage correlation is obviously going to be % because we are correlating a statistic with itself within the same season, but on a year-to-year basis, winning percentage only correlates with itself at about a 25% rate, meaning winning percentage can’t be used to accurately predict itself on a year-to-year basis. 

It’s well-known the NFL is a parity league that is highly unpredictable every season, but this just puts into context how unpredictable and how tough it is to handicap a team’s future success. Fortunately, there are statistics that are significantly more predictive of future winning percentage than winning percentage itself.

Let’s start with one I’ve already mentioned, yards per play. The below chart breaks out yards per play, yards per play allowed, and yards per play differential. Note: any “allowed” statistics will have a negative correlation with winning percentage because the less a team allows, the more they win.

StatisticsYear to YearWinning %Next Year Winning %

All three statistics are reasonably predictable on a year-to-year basis, with yards per play allowed actually being the most predictable of the three by a slight amount, although offense correlates with winning at a much higher rate and is much more predictive of future winning than defense. This is a theme we’ll see throughout this analysis, offense being more predictive than defense.

In terms of overall differential, this statistic correlates with same season winning slightly less than turnover margin does, but because it is significantly more predictable, it’s significantly more predictive of future winning, correlating with future winning at about a 32% rate, already a significant increase from the 25% predictiveness we get just from looking at winning percentage.

We can do better than that though. Let’s look at another obvious one that would correlate heavily with winning, points, more specifically points per play, points per play allowed, and points per play differential.

StatisticYear to YearWinning %Next Year Winning %

Right off the bat, we see this correlates with same season winning at a very high rate, which is to be expected, considering points are what decides games. It’s not a perfect correlation as teams can win a high percentage of close games in a single season sample size, and, as a result, would have a better record than their point differential would suggest, but point differential is the most predictive statistic of same season winning that we’re going to find. 

It’s also a good predictor of next season’s winning percentage, as this is the best predictor of future winning that we’ve seen yet by far.  However, there are a couple big problems with points per play differential as a statistic. For one, while it is relatively predictive, it’s not all that predictable, predicting itself at just a 35% rate, which gets even worse when you look at points per play and points per play against, which only correlates with itself on a year-to-year basis at about %. 

That leads into my second big problem with this statistic, that it does a relatively poor job of breaking out offense versus defense, which is likely why points per play and points per play allowed are relatively unpredictable statistics. Return touchdowns by special teams or defense count towards points per play and against opponents’ points per play allowed and field position skews this statistic even more, as good defenses can easily look bad in this statistic if their offense constantly gives them terrible field position to start, and vice versa. 

At first glance, it might seem like a good thing that the gap in predictiveness between points per play and points per play allowed is less than other offensive/defensive statistics, but I think that is a result of neither stat accurately representing the side of the field it is supposed to represent. As we’ll see more going forward, if offense and defense are broken out from each other properly, offense always is significantly more predictive.

The next statistic is a personal favorite of mine, first down rate differential, which includes first down rate and first down rate allowed.

Year to YearWinning %Next Year Winning %

Right away what stands out is that, across the board, first down rate and its associated statistics are significantly more predictable than anything we’ve seen thus far and, in fact, it is the most predictable statistic year-to-year. It also does a great job separating offensive and defensive performance and, unsurprisingly, there is a significant gap between the predictiveness of offense and defense performance, more so than any statistic we’ve seen thus far. As a result, first down rate correlates with future winning more than yards per play, but yards per play allowed correlates with future winning more than first down rate allowed.

The disappointing thing about first down rate differential is that, while it is significantly more predictable year-to-year and higher correlated with same-season winning than yards per play differential, it isn’t actually more predictive of future winning year-to-year than yards per play differential, at least over the year sample of this study. On top of that, in comparison to points per play differential, it is less predictive of future winning, despite the problems with points per play differential. 

However, there is still a lot to like with first down rate differential and there is a key thing that points per play differential takes into account that first down rate doesn’t that likely explains why it is more predictive. That key thing is special teams, which both yards per play differential and first down rate differential both lack, likely the reason they are not as predictive. Reliable special teams statistics are hard to come by, but one that does a great job is DVOA, Football Outsiders’ signature statistic.

To illustrate this point, I’ve broken out overall, offensive, defensive, and special teams DVOA.

Year to YearWinning %Next Year Winning %

Across the board, DVOA does very well, correlating with next year’s winning at about the same rate as points per play differential, while effectively separating out performance in all three phases of the game. The standout here is special teams though, having year-to-year predictability in line with other phases in DVOA and surprisingly correlating with winning and future winning relatively well, given how small a part of the game special teams is. 

In fact, special teams DVOA is actually slightly more predictive of winning than defensive DVOA, at least over the course of this year sample. I would take that with a bit of a grain of salt, but it’s clear that special teams performance has a much bigger impact on winning than most, including myself, would expect. Because of this, I am going to go back and factor special teams more significantly into my season previews.

Given that special teams is likely what makes points per play differential more predictive than first down rate differential, I decided to add special teams DVOA to first down rate differential and see what that does to predictiveness. I played around with different allocations of offensive, defensive, and special teams performance, but I found that 45% offense, 30% defense, and 25% special teams was most predictive, which once again reinforces the importance of special teams.

Year to YearWinning %Next Year Winning %

Just by adding special teams to first down rate differential, we get a statistic that is more predictive than anything we’ve seen so far. We can do better than this though. Since we know that yards per play allowed is more predictive than first down rate allowed, let’s see what happens when we swap yards per play allowed into this hybrid statistic. Once again, I found the 45/30/25 split was most predictive.

Year to YearWinning %Next Year Winning %

This gets us to an impressive number when you consider that winning percentage itself predicts future winning percentage at just a 25% rate. NFL records are very tough to predict year-to-year, but having a statistic that correlates with future winning percentage at a % rate is a very useful tool for handicapping. 

For the record, I tried swapping in points per play allowed and defensive DVOA and both lowered the predictiveness significantly. Points per play allowed didn’t surprise me because, even though it was predictive, it includes things that the offense is already being given credit for. Defensive DVOA surprised me a little, but it’s not a very predictive statistic year-to-year, so it’s not a huge surprise that including it did not have a positive effect on predictiveness.

Let’s see how each team performed in this metric in


Obviously, this can’t be blindly followed, as % correlation is still not that high and a lot changes for teams from season to season to affect their performance from year-to-year, but this is a much better base point to start with than win/loss record.

I also wanted to show a few other breakdowns. This one shows yards per play differential broken out into pass offense, pass defense, rush offense, and rush defense.

Year to YearWinning %Next Year Winning %

Unsurprisingly, offensive statistics are more predictable and predictive than defensive statistics and, also perhaps unsurprisingly, pass statistics are more predictable than rush statistics and by a significant amount.

Let’s take a look further at passing statistics.

Year to YearWinning %Next Year Winning %
Completion %%%%
INT %%%%

We see that completion percentage is much more predictable year-to-year than any other metric, but yards per play correlates better with winning and next year’s winning. Touchdown rate also correlates with winning and next year’s winning, but is tough to predict on a year-to-year basis. Interception rate is as well, but it’s notable that it’s significantly more predictive than turnover margin, which brings me to my next chart.

Year to YearWinning %Next Year Winning %
INT %%%%
Def INT %%%%
Fumbles Lost%
Fumbles Recovered%

While turnover margin itself is very unpredictive, interception rate seems to at least have some predictive value, which makes sense, given that passing offense is what tends to be most consistent year-to-year. Teams who fare well in turnover margin as a result of having a quarterback who had a low interception rate are more likely to see their turnover success continue than teams reliant on defensive takeaways or avoiding fumbles. For fumbles, I didn’t even bother calculating its relationship to winning because of how unpredictable it is year to year. There is no predictive value to a statistic you can’t reasonably predict and fumbles are a perfect example of that.

Year to Year
1st/2nd vs. 3rd/4th differential%
1st/2nd allowed%
3rd/4th allowed%
1st/2nd vs. 3rd/4th allowed differential%

This is the last one I want to show for now. I may add more to this later, but this breaks out the year-to-year predictability of first down rate and first down rate allowed between early downs (1st and 2nd) and later downs (3rd and 4th). I didn’t correlate these statistics with winning because it’s obvious that better success on 3rd and 4th down leads to better results on the scoreboard, but it’s worth noting that those downs don’t tend to be any more predictive than early downs and there is minimal, if any, evidence that teams can consistently outperform their 1st and 2nd down performance on 3rd and 4th down year-to-year, as there is very little year-to-year correlation in the differential between early down and later down performance.

Like this:




1468 1469 1470 1471 1472