Last season, I took inspiration from Bill-In-Bangkok’s Wild Card Watch series and attempted an alternative take on playoff scenario projections. Bill is, by his own admission, an eternal optimist, and his series is intended to inspire hope, by illustrating paths that the Commanders could take to make it to the playoffs.
I am a scientist, by training, but like Bill, I’m also a diehard Redskins fan of over 50 years. So I come at this from a slightly different perspective. My first attempt was a probabilistic simulation, intended to capture the element of unpredictability which makes football equal parts excitement and disappointment. It was partially successful, in that I was able to predict outcomes of more than 50% of games in most weeks. But as a few readers pointed out, aside from some entertainment value - mainly mine - it was ultimately pointless.
This season, I am going to aim much higher by attempting accurate projections, using offensive and defensive efficiency metrics to predict the outcome of each NFC team’s remaining matches.
My approach is based on the metric Expected Points Added (EPA). The Expected Points/EPA concept was first developed in The Hidden Game of Football. It is used by betting analysts to handicap games and forms the basis of ESPN’s popular quarterback rating metric, Total QBR. The basic concept is that every play on offense increases or decreases the team’s chance of scoring. Through statistical analysis of many seasons of football, it is possible to generate models that assign an Expected Points value to the offensive team at the start of each play based on down, distance and game situation.
EPA is the difference between Expected Points before and after a play. It provides a powerful, statistically valid metric of the impact of the play on a team’s chance of scoring. EPA/play quantifies the average increase in scoring probability per play by an offense. It is a measure of offensive efficiency, and provides a useful metric for comparing teams. The best offense in the league this season, Kansas City, has an EPA/play of 0.192 points, and Houston brings up the rear with an EPA/play of -0.163.
EPA/play can also be calculated by applying the same approach when a team is on defense. The range of values for defenses is similar, but the sign is inverted. The best defenses achieve large negative EPA/play values because they tend to decrease their opponents’ chances of scoring when they are on the field. Bad defenses have positive EPA/play values because opposing offenses have their way with them. The current 2022 season leader on defense is New England at -0.108 EPA/play. Detroit and Chicago are tied for last place at 0.111 EPA/play. Washington is currently in 7th place across the whole season, with -0.053 EPA/play on defense, and has risen to 3rd place since Week 7, with an EPA/play of -0.113 in that span (New England is still on top with a -0.128 EPA/play since Week 7).
In this projection, I will use EPA/play to predict outcomes of each NFC team’s games for the remainder of the season to arrive at a playoff seeding. I will repeat the projection each week for the rest of the season until the Commanders are eliminated from contention or Ron Rivera hoists the Lombardi trophy.
Predicting Game Outcomes
My method uses EPA/play to project how each team’s offense matches up against their opponent’s defense and, conversely, how their defense matches up against the opponent’s offense. To predict the winner of each game, I calculated a “score” for each team by adding the team’s EPA/play on offense to their opponent’s EPA/play on defense. The resulting EPA/play differential indicates whether their offense is overmatched or undermatched by the opponent’s defense, and by how much. I then did the same to calculate opponent’s “score”. The team with the higher score was the winner.
That is probably enough detail for most readers. Readers who are interested in the methodological details can keep reading to end of this section. Everyone else can just skip to the Playoff Projection.
EPA/play was calculated using the rbsdm.com website, created by Ben Baldwin and Sebastian Carl. In an attempt to make the predictions as accurate as possible, I had to account for the fact that teams’ performance does not always remain constant throughout a season. As the Commanders have demonstrated, teams sometimes improve after a slow start. Other teams can implode or suffer major setbacks, such as the loss of their QB or other key players to injury. Rather than attempting to delve into the intricacies of each team’s season to date, I settled on a more simplified approach.
Probably the biggest determinant of offensive efficiency is QB play. If you don’t believe that, check out the offensive EPA/play rankings here. It is not a perfect relationship, but the most efficient offenses tend to be led by superstar QBs, while the teams near the bottom of the rankings generally need help at the position. In order to account for the impact of changes at QB throughout the season, I calculated each team’s offensive EPA/play over the duration that they have been playing with their current starting QB. For example, the Commanders’ offensive EPA/play was calculated from Week 7, when Taylor Heinicke took over from Carson Wentz.
I ignored cases in which a starting QB missed a single game due to injury. When the original starter missed two or more games, I started the calculation from the when the current starter took over or returned to the lineup for good. For example, Dallas’ EPA/game was calculated from Week 5, when Dak Prescott returned from injury without missing any more games.
This approach created a bit of a quandary for two teams. Arizona is expected to start Colt McCoy in Week 15, due to Kyler Murray’s injury on MNF. My simple solution was to calculate Arizona’s EPA/play from Weeks 10 to 11 when McCoy last started. Baltimore was trickier, because they are likely to activate one of the third stringers, from the practice squad to replace injured backup Tyler Huntley. Since neither Anthony Brown nor Brett Hundley has started for Baltimore, the best I could do was to calculate EPA/play in Weeks 13 and 14 when Lamar Jackson was out of the lineup, making the assumption that Brown or Hundley is likely to play more like Huntley than Jackson.
My approach for calculating defensive EPA/play was simpler. Washington’s defense came into its current form around Week 6 to 7, so I just calculated every team’s defensive EPA/play over the last eight weeks. This might impart a bit of home team bias, but I figure the scrappy underdogs could probably use a little help.
Last of all, in this simulation, teams’ EPA/play on offense and defense remains fixed for the remainder of the season from Week 15 on. That should mean that the projection gets less accurate as the season progresses to Week 18, since there is no way to predict which teams will get better or worse. I will update the EPA/play values each week until the end of the Commanders’ season.
Week 15 Results
Eagles W (13-1) – Bears L
Cowboys W (11-3) – Jaguars L
Commanders W (8-5-1) – Giants L (7-6-1)
Commanders – Giants Matchup: Coming off the bye, follow the Week 13 tie with the Giants, the Commanders’ defense (-0.113 EPA/play) provides the decisive edge in the rematch, resulting in a dominating win by Washington. This time, the Giants’ vulnerability on defense (0.104 EPA/play) allows the Commanders’ middle of the pack rushing offense (-0.013 EPA/play, ranked 11th Weeks 7 to 14) to dominate time of possession and control the pace of the game.
The win moves the Commanders ahead of the Giants in the Wild Card race and provides them with a tiebreak advantage over their division rival and a potential tiebreak advantage over competitors outside the division by adding a conference win.
Vikings W (11-3) – Colts L
Lions L (6-8) – Jets W
Packers W (6-8) – Rams L
Bears L (3-11) – Eagles W
Bucs L (6-8) – Bengals W
Panthers W (6-8) – Steelers L
Saints W (5-9) – Falcons L (5-9)
49ers W (10-4) – Seahawks L (7-7)
Cardinals L (4-10) – Broncos W
Rams L (4-10) – Packers W
Week 16 Results
Eagles L (13-2) – Cowboys W (12-3)
Commanders L (8-6-1) – 49ers W
Giants L (7-7-1) – Vikings W
Commanders – 49ers Matchup: Unfortunately for the Commanders, the expected drop off in offensive play with Mr. Irrelevant Brock Purdy replacing Jimmy Garoppolo as the 49ers’ starter has not materialized. Purdy was the highest rated QB in the league in Week 14, with an amazing total QBR of 92.8 and a stat line of 16/21 for 185 yards, 2 TDs, 0 INT and 1 rushing TD. The 49ers had the third most efficient offense in the league in Week 14, with an EPA/play of 0.188, trailing only Philadelphia and Detroit. While the Commanders had the advantage on defense (-0.113 EPA/play vs. 49ers -0.077 EPA/play) it was not enough to offset the 49ers’ superiority on offense (-0.01 EPA/play vs 49ers 0.188 EPA/play).
Vikings W (12-3) – Giants L
Lions W (7-8) – Panthers L
Packers L (6-9) – Dolphins W
Bears L (3-12) – Bills W
Bucs W (7-8) – Cardinals L
Panthers L (6-9) – Lions W
Saints W (6-9) – Browns L
Falcons L (5-10) – Ravens W
49ers W (11-4) – Commanders L
Seahawks L (7-8) – Chiefs W
Cardinals L (4-11) – Bucs W
Rams L (4-11) – Broncos W
Week 17 Results
Eagles W (14-2) – Saints L
Cowboys W (13-3) – Titans L
Commanders W (9-6-1) – Browns L
Giants W (8-7-1) – Colts L
Commanders – Browns Matchup: Fortunately for the Commanders, the much awaited Week 13 debut of Deshaun Watson with the Browns has been a massive disappointment so far. In two starts to date, he has failed to perform as well as journeyman place holder Jacoby Brissett, throwing 1 TD to 2 INTs, with a completion percentage of 59.4% and achieving a mediocre QBR of 48.3, compared to Brissett at 59.6. With Watson under center, the Browns have had the fourth-least efficient offense in the NFL (-0.197 EPA/play).
The Commanders enjoy advantages on both sides of the ball (-0.113 EPA/play on defense vs Browns 0.001 EPA/play; -0.01 EPA/play on offense vs Browns -0.197 EPA/play), making this one a decisive win.
Vikings W (13-3) – Packers L (6-10)
Bears L (3-13) – Lions W (8-8)
Bucs L (7-9) – Panthers W (7-9)
Saints L (6-10) – Eagles W
Falcons W (6-10) – Cardinals L
49ers W (12-4) – Raiders L
Seahawks L (7-9) – Jets W
Cardinals L (4-12) – Falcons W
Rams L (4-12) – Chargers W
Week 18 Results
Eagles W (15-2) – Giants L (8-8-1)
Cowboys W (14-3) – Commanders L (9-7-1)
Commanders – Cowboys Matchup: Much as it kills me to admit this, the Cowboys are a much stronger team on offense (0.134 EPA/play vs. Washington -0.01 EPA/play), while the two teams are fairly evenly matched on defense since Week 7 (Washington -0.113 EPA/play vs. Cowboys -0.115 EPA/play). Unfortunately, Washington’s 20th ranked ball-control offense is a major mismatch against Dallas’ 2nd ranked defense. This one could be as bad or worse than the first Dallas game this season, unless the Cowboys rest their starters for the playoffs, which is a very realistic possibility.
Vikings W (14-3) – Bears L (3-14)
Lions W (9-8) – Packers L (6-11)
Bucs W (8-9) – Falcons L (6-11)
Panthers W (8-9) – Saints L (6-11)
49ers W (13-4) – Cardinals L (4-13)
Seahawks W (8-9) – Rams L (4-13)
Final NFC Playoff Seeding
NFC East Champion – Eagles (15-2), 1st seed
NFC West Champion – Vikings (14-3), 2nd seed
NFC North Champion – 49ers (13-4), 3rd seed
NFC South Champion – Panthers (8-9), 4th seed –tiebreak over Bucs 2-0 head to head record
Wild Card 5th seed – Cowboys (14-3)
Wild Card 6th seed – Commanders (9-7-1)
Wild Card 7th seed – Lions (9-8)
In this scenario, the Eagles clinch the first seed, earning a bye in the Wild Card round and home field advantage throughout the playoffs. The Vikings are the second seed. The 49ers take the 3rd seed. The Panthers win the NFC South, despite the losing record, thanks to a 3-1 late-season run. They beat the 8-9 Buccaneers for the division title on the first tiebreak due to their 2-0 record in head-to-head competition.
The Wild Card seeds are determined entirely by Win-Loss record alone, with no need to rely on tiebreaks. Washington finishes with the 6th seed, which means they play the 49ers in the Wild Card round. Had they ended up with the 7th seed, they would be facing a weaker opponent in the Vikings.
Much as I hated Rivera playing for the tie against the Giants, rather than going for the win, the close race with the Lions in this scenario vindicates the wisdom of that decision. In the scenario projected here, an additional win would not have elevated the Commander’s playoff position, as they trail the next highest seeded team by five games. On the other hand, losing the Giants game instead of tying would have dropped them out of the playoffs entirely.
What I like best about this scenario is that the Commanders end up with a very realistic chance of beating their Wild Card round opponent. The last time the Commanders played the Vikings, the game was tied until Minnesota’s game winning field goal with 12 seconds remaining. Had Taylor Heinicke not taken the sack on 3rd and 6 to end the Commanders’ last offensive drive with 6:59 remaining, the outcome could have been different. Since that game, the Commanders have been playing consistently well on defense (2nd ranked by EPA/play) and better than league average on offense (13th ranked by EPA/play). If they can contain the mistakes on offense and maintain their form on defense, they have a realistic opportunity to make their deepest playoff run in 17 years, or perhaps longer.
Acknowledgement: Edited by James Dorsett
Which Wild Card contender is the biggest threat to Washington’s playoff aspirations?
This poll is closed