By MICHAEL CHURCHWARD

There is a concept in football known as ‘lost yardage.’ There is an occasional reference to it and its potential impact on games, but I have not seen the impact of lost yardage amongst all the teams in the NFL. So I want to consider the specific plays in which lost yardage can be calculated and measure its impact on game outcomes.

First, an explanation of the concept of lost yardage:  This isn’t lost yardage in the sense of a punt returner losing yards on a return. This covers a situation in which a team moves the ball to a certain yard line, there is a penalty on the play and the ball is moved back to a different yard line depending on the specific call.

The difference of yardage where the ball would have been compared to where the ball was actually placed is my definition of lost yardage. 

For example, if a team receives a kickoff and returns the ball to the 50 yard line, but a penalty makes the team start the subsequent drive from their own 20, the lost yardage on that play would be -30 yards. 

Using this formula to calculate lost yardage, I wanted to find out how this concept could affect NFL teams on a per-game basis. I looked at plays where lost yardage occurs on kickoff returns and punt returns, with no turnovers on the play. I chose these plays because that team is going to be on offense the next play so there will be a quantifiable consequence for lost yardage. 

I used Sports Info Solutions’ formula for Expected Points (EP) to calculate the points lost due to lost yardage. There is an EP value assigned based on the down, distance, and whether or not the offense was the home team. Therefore, we can assign an EP value pre-penalty, and an EP value post-penalty. The difference between those two numbers is what we can define as points lost due to yardage lost. 

The data collected uses a three-season sample from 2016 to 2018, including all the playoff games and Super Bowls from those seasons. I looked at the final score in every game that had a lost-yardage situation on a kick return or punt return from the losing team, and then checked to see if the expected points from the lost yardage may have directly led to a loss. These were mostly games that were decided by three points or less and the EP lost from the penalty cost the team more than the difference in the final score. 

This research is focused on Expected Points lost on kickoff and punt returns, and while yards isn’t a perfect translation into points, I think that we can all agree that the further away an offense starts from the opposing team’s goal line, the harder it is for that team to score. While there are more plays that result in lost yardage due to penalties on offensive plays, starting offensive drives at a positional disadvantage was where I wanted to start. Offensive penalties on positive yardage plays are more of a common way in which lost yardage can accumulate, and future research will be focused on that aspect of lost yardage.

Sorting through all the scenarios that met my criteria, there were several games in the past couple of seasons in which a penalty on a punt or kickoff may have directly led to a loss. According to the expected points lost due to a penalty, there were 10 games during the past few seasons where the game could have been a win instead of a loss. I have a table below that will show the numbers for each of these games to show the starting yard lines and points that were lost

Game	Game result with winning team shown	Pre-penalty drive start	Actual drive start	Pre-penalty EP	Actual EP	EP Lost
09/30/2018, Week 4, CIN @ ATL	CIN 37 -36	-32	-9	1.56	0.39	1.17
12/03/2017, Week 13, SF @ CHI	SF 15 – 14	+ 16	-14	4.47	0.53	3.94
10/16/2016, Week 6, CLE @ TEN	TEN 26 – 28	+ 26	-11	3.70	0.13	3.57
12/11/2016, Week 14, DAL @ NYG	NYG 7 – 10	+ 29	-33	3.48	1.34	2.14
12/11/2016, Week 14, DAL @ NYG	NYG 7 – 10	-32	-21	1.28	0.56	0.72
12/11/2016, Week 14, DAL @ NYG	NYG 7 – 10	-28	-9	1.00	0.06	0.94
09/16/2018, Week 2, DET @ SF	SF 27 – 30	+ GL/ TD	-21	6.00	0.56	5.44
12/16/2018, Week 15, DET @ BUF	BUF 13 – 14	+ 49	-18	2.38	0.45	1.93
12/17/2017, Week 15, DAL @ OAK	DAL 20 – 17	-30	-10	1.45	0.40	1.04
12/17/2017, Week 15, DAL @ OAK	DAL 20 – 17	+ GL/ TD	-10	6.00	0.40	5.60
12/17/2017, Week 15, DAL @ OAK	DAL 20 – 17	-25	-11	1.12	0.43	0.69
12/11/2016, Week 14, WAS @ PHI	WAS 27 – 22	+ GL/ TD	-23	6.00	0.98	5.02
09/24/2018, Week 3, PIT @ TB	PIT 30 – 27	+ GL/ TD	-8	6.00	0.39	5.61
10/30/2016, Week 8, WAS @ CIN	T 27 – 27	-35	-21	1.43	0.56	0.87
10/30/2016, Week 8, WAS @ CIN	T 27 – 27	-22	-9	0.60	0.06	0.54

There are many special teams plays throughout the course of an NFL game and season that are overlooked by most people, including coaches. With such a razor-thin margin for victory in the NFL, special teams play needs to be noticed as a major factor. Hopefully, this statistic of lost yardage and loss of expected points will highlight the major role special teams plays in every single game and how teams could potentially lose due to the impact of one penalty in a game.

BY ALEX VIGDERMAN

Many (hopefully all) of you know that we recently concluded the initial judging of our first Football Analytics Challenge. We released some previously-locked-down defensive alignment data to the public and asked people to come up with an answer as to which defensive line position is the most valuable. To go with the competition, we also asked registrants to donate whatever they could to the United Negro College Fund.

To bring in 133 donations totaling $3,300 so far was beyond our expectations. And we ended up with a solid crop of 34 submissions for the competition, with the finals being presented on YouTube tomorrow night (Wednesday July 29)!

While we are obviously excited to show you the research that was done by the finalists, we didn’t want to turn away from the work of the other 31 teams off to the side. So here are a few highlights of the efforts of the rest of the participants.

As a company that dabbles in multiple sports, we appreciate it when analysts draw from multiple sports in their work. Both Nate Rowan and Sam Chinitz cited baseball’s Weighted On-Base Average (wOBA) as the inspiration for their approach to valuing the events on a play. Rowan called his key metric “Points Gained,” which essentially measures the value of a charting data point by taking the difference in EPA/play between plays with and without that event occurring.

Matthew Reyers, Meyappan Subbaiah, Dani Chu, and Lucas Wu leveraged two key resources outside of the provided data set to aid with their research. The first was the nflWAR paper by Yurko et al (whose work multiple submissions referenced), and the second was the predicted yards at the time of the handoff from the 2019-20 NFL Big Data Bowl winners.

Sam Struthers and Adrian Cadena used ideas about division of credit from the Yurko paper to distribute EPA among the players who had a chance to be involved on a play. They also estimated the extent to which edge pressure affects the performance of the interior line and vice versa, which was a unique approach.

Alex Stern invoked multilevel modeling (which does a good job in measuring player-to-player variation when sample sizes can differ wildly) to evaluate the same concept of Individual Points Added. In the passing game, the model focused mostly on generating pressure, which was a decision that many teams made thanks in part to recent research from Timo Riske of Pro Football Focus.

Calvin Smith used a linear model to predict the EPA of a play based on the existence and direction of pressure. Unsurprisingly, avoiding pressure altogether is the most valuable, with outside pressure being the most effective at reducing the offense’s EPA.

Matt Colón, Silas Morsink, Robbie Thompson, and Peter Gofen were one of a couple teams (including one of the finalists) who used Madden ratings to help quantify player talent. The group’s approach to evaluating play outcomes was what stood out the most, however. They figured that defensive linemen don’t have much impact on the specific final result of the play, but they do affect what kind of play it was, roughly. So, when evaluating the contributions of each defensive line position, instead of using actual play results, they replaced each play’s EPA with the average EPA value for many different kinds of play results (e.g. “Rush big loss”, “Screen under pressure”, “Medium pass”).

Dan Rees used some notions of how to break down a play using charting data that we use ourselves within our Total Points statistic. He also focused on the range of possible EPA values on a play when judging a player’s opportunities instead of just the EPA itself, which he called a play’s EPA Range. David Schmerfeld also took a Total-Points-esque angle at valuing plays, and added in explicit measures of “Indirect Impact” that allowed interior linemen to receive identifiable credit for their more subtle play-to-play value.

Keegan Abdoo and Mehmet Erden used a similar approach, using a linear model that controlled for situational factors to estimate the EPA contribution of a streamlined set of charting data points on each of run plays (forcing the rusher to bounce or cut back) and pass plays (pressuring the quarterback or breaking up the pass).

A few teams used clustering to robustly characterize player positions using some combination of roster position and play-to-play alignment. One of the better implementations of that belonged to James Hyman, Colin Krantz, Brendan McKeown, and Kushal Shah, who used a random forest to model the most likely roster position for a player (including a Hybrid DE/DT position) and then combined those with defensive line techniques to feed the clustering algorithm.

We’re so glad to have received so many great submissions to our competition. Feel free to check out work by the finalists or by anyone else in the competition on the competition’s GitHub repository.

This weekend, Sports Info Solutions (SIS) announced the three finalist groups for its Football Analytics Challenge.

Team 1 – Andrew Rogan and Robert Bernhardt

Team 2 – Zachary Feldman, Michael Egle, and Anthony Reinhard

Team 3 – Emmett Kiernan

The finalists, who were chosen from 34 total submissions, will compete in a live final event on Wednesday, July 29 at 8pm ET.

The live event will feature presentations by each of the finalist groups, given to a panel of judges consisting of Matt Manocherian (SIS), Aaron Schatz (Football Outsiders), and John Park (Indianapolis Colts).

A live feed of the event will be streamed on YouTube for public viewing, with the link distributed via SIS social media shortly before the scheduled 8:00pm ET start time (Twitter: @SportsInfo_SIS).

The Football Analytics Challenge is a public competition that tasked participants with analyzing SIS play-by-play charting data to answer questions about Defensive Line position value in the NFL. To read more about the event, the rules, and judging criteria, visit this link.

All proceeds generated by the event go to the United Negro College Fund (UNCF). To date, over 130 donors have helped raise over $3,300 dollars. Donations are still being accepted through July 31^st, via GoFundMe.

By John Shirley

A few weeks ago on this site, we published research showing the effects weather has on a quarterback’s expected completion percentage. Now we will be applying that same question to the running game to see if any weather effects exist.

For this analysis, weather effects on running backs will be analyzed with a few different statistics, including Yards Per Carry, Positive%, and Broken/Missed Tackles. Data used will be from the past two seasons and include only rushing attempts by the road team to help with sampling bias. This results in a data set that includes a little over 11,500 carries.

Dome vs Outdoor

As a first look into weather effects on rushing, lets see how performance is impacted by whether or not the game was played in a covered stadium. 

Running Back Rushing Performance of Road Teams by Roof Type (2018-2019)

Roof Type	Yds / Carry	Avg Yds After Contact	Positive%	BT+MT / 100 Carries
Open	4.3	2.6	42%	15.7
Dome	4.1	2.4	42%	14.6

By looking at the numbers simply grouped by outdoor games versus games in a dome, it seems there could be a very slight positive benefit to playing outdoors. Running backs had an increase in three of the statistical categories when they were outside, though, Positive% was noticeably unchanged. This is the opposite effect that was seen previously in quarterback performance.

Weather Effects

Similar to the previous research into quarterback performance, we will also look into the role weather plays in rushing performance, if any. The weather variables used will once again be Temperature and Significant Precipitation (defined as any time the precipitation intensity was greater than or equal to 0.25 mm/hr). 

As with passing performance being adjusted for other variables such as throw depth and throws outside the numbers, running back rushing performance needs to be adjusted for down, distance, and defenders in the box.

After attempting to model a relationship between the two weather variables and the first two rushing metrics of Yards and Yards After Contact, we found both variables not statistically significant. 

However, when modeling a relationship between the two weather variables and Positive%, Temperature becomes statistically significant, albeit with only a marginal impact, even on the season level. Temperature was determined to have a negative relationship with Positive%, meaning that as the temperature increases, the Expected Positive Play Rate decreases.

One way to quantify which performances were most affected is comparing a running back’s Expected Positive% without accounting for weather conditions to a model that includes Temperature as a variable.. The 2019 game which was most positively impacted by the weather was the Bears vs Packers matchup in Week 15. After adjusting for weather, both starting running backs Aaron Jones and David Montgomery had close to a 2% boost in Expected Positive%. 

Player	Week	Expected Pos%	Weather Adjusted Expected Pos%	Difference
Aaron Jones	15	37.9%	40.0%	2.1%
David Montgomery	15	42.2%	44.2%	2.0%

This shows that even though Temperature is a significant variable in the model, even in the most extreme case its impact is relatively small.  

Grass vs Turf

Although the field surface type is not a weather variable, we did find it to be an interesting piece when attempting to model a predicted rate for running backs forcing a broken or missed tackle on each individual tackle attempt. The two weather variables were both insignificant in this analysis, but field surface type was found to be a significant variable. This differs from each of the previous models mentioned within this article, in which field surface type did not play a role. 

Running backs attempting to force a broken or missed tackle have a higher success rate of doing so on grass fields rather than turf fields. This is somewhat shown in the first table, where outdoor stadiums had a higher Broken+Missed Tackles / 100 Attempts. But it doesn’t tell the whole story, as a stadium’s roof type is not the deciding factor. A stadium’s field surface type seems to be the actual factor providing the difference. Though, this research is currently limited by a binary classification of field surface type, when in reality there are multiple different types of turf and grass being used in the NFL. 

Overall Findings

• Temperature and Precipitation are NOT significant variables when trying to predict Rushing Yards,  Rushing Yards After Contact, or Broken+Missed Tackle Rate.
• Temperature IS statistically significant when predicting Rushing Positive%. This results in running the ball being slightly more efficient in cold weather. Though, the overall impact is relatively small, even in the most extreme cases.
• The field’s surface type IS a factor in forcing broken and missed tackles, where it is easier to do so on grass than turf. Though, this too has a relatively small impact overall.
• Overall, weather and the field’s surface type have limited impacts on rushing offense.

By BRYCE ROSSLER

EA Sports has begun to slowly unveil their player ratings for Madden 2021. Getting angry at video games is silly but this is typically a dead period for football news and the ratings are the only thing the developer ever changes, so it’s something for people to do. And as long as EA continues to undertake the impossible task of accurately quantifying player skill, people will continue to critique them.

We’d like to approach this with more nuance than anger, so we’ll intake ratings as they’re released and contrast anything noteworthy to our suite of advanced stats.

Just two QBs over a 95. Rodgers outside the 90s club.

Here are the top 10 Madden QB ratings. Thoughts? 🤔 pic.twitter.com/EkAeCYawiz

— ESPN (@espn) July 13, 2020

Tom Brady a 90?

Tom Brady showed signs of decline last year, so Madden rating him a 90 overall and making him the fifth-best quarterback in the game seems nostalgic. He had his worst season of the Total Points era (2016-present) in 2019, ranking 19th in Passing Total Points/60 Snaps (8.2) after ranking 1st (16.2), 6th (12.9), and 4th (11.3) in 2016, 2017, and 2018, respectively. While I’m sympathetic to the talent sink in New England last season, that’s a pretty steady decline.

Good luck tackling these guys 💪 #Madden21 @EAMaddenNFL pic.twitter.com/eTpV4DOXoa

— NFL on ESPN (@ESPNNFL) July 14, 2020

Running Backs are … not badly rated

EA Sports also released the top ten running backs in broken tackle rating, and did a pretty good job with it. Saquon Barkley (12th), Joe Mixon (13th), and Dalvin Cook (17th) all fell outside of our top 10 in broken tackle rate among RBs with at least 150 carries in 2019. 

Kareem Hunt didn’t qualify for that leaderboard this year, but ranked eighth in that metric in 2018, which is about where he’s slotted in Madden 2021. Mark Ingram, who ranked first among that group with 18.8 Broken Tackles Per 100 Carries, arguably deserved the highest rating, but there aren’t a lot of gripes to be had with these ratings otherwise.

Because the ratings reveal is ultimately a marketing campaign, EA Sports still has a lot of cards up its sleeve. Top 10 lists can be hard to argue with just because they inevitably include good players, so the real dissection begins once they release full rosters. We will continue to keep an eye on the information they release and update this piece with anything that sticks out as interesting.