Author’s Note: The projection in this article was run earlier in the week, before half the WFT roster was placed on the COVID reserve list. This series is intended as an alternate take on standard NFL season and weekly projections, with the key point of departure being that it allows for unexpected outcomes to occur, simulating the influence of random, unpredictable events on NFL game and season results.
As if to demonstrate my point, the Sars-Cov-2a Omicron variant intervened in the NFL season this week, wreaking havoc on some teams’ rosters, while largely sparing others. The new COVID outbreak will be the key determining factor in many game outcomes this week, and no team is likely to be more impacted than Washington. As of this update, 23 WFT players are on the Reserve/COVID-19 list, including a number of key starters and contributors; while the Eagles have only lost starting WR Quez Watkins and backup RB Jason Huntley.
I came close to pulling the plug on this series when the COVID list got to 18, but then I realized that random events messing with our best predictions is really what it’s all about. There wasn’t time to include handicapping for COVID impacts in this week’s edition. I will include that next week unless something miraculous happens to turn that situation around.
Last Sunday’s loss to Dallas was like a Rorschach test for Washington fans. Everyone seems to have taken away from it what they want to see.
To the legions of disgruntled fans, haters and one actual troll who flooded the Hogs Haven comment threads, the beating inflicted by the Cowboys in the first half was further evidence that Heinicke sucks, Rivera is hopeless, Scott Turner can’t game plan to save his life and the Team is doomed to never win a playoff game this decade.
Those of us who remain hopeful for a playoff win this season saw that, with 2:35 left, the Football Team was only down by seven points and driving. If not for the dropped pass by DeAndre Carter or Antonio Gibson’s fumble, the outcome could have been different. That means the Cowboys are beatable. Since Dallas is the toughest opponent on the remaining schedule, with a few key players returning from injury, we still might have a good chance of making the playoffs.
At least that was how it looked on Sunday afternoon, before the COVID omicron outbreak swept through the WFT roster, while largely sparing Philadelphia’s. Sunday’s game at Lincoln Financial Field will be an uphill battle, with the Football Team taking to the field minus key players including Taylor Heinicke, Kyle Allen, Jonathan Allen, Kamren Curl, Kendall Fuller, Matt Ioannidis, Tyler Larsen, Cornelius Lucas, Tim Settle, Cam Sims, James Smith-Williams, Montez Sweat and Casey Toohill.
Even so, as a long suffering, old school Redskins fan, I am still thankful to be watching meaningful games in December. Like Bill-in-Bangkok’s reality-based wildcard watch, this series is for the fans who are still hanging on to hope that we’ll make the playoffs. That audience might be dwindling to a few hardcore holdouts with this week’s developments, but I will continue undeterred until the WFT is mathematically eliminated or pulls off a miracle.
Unlike traditional football season projections, my alternate take uses a simple probabilistic model to allow for the unpredictable upset wins that happen every week in the NFL, and which the Football Team will almost certainly need now if our hopes are to be fulfilled.
From some of the comments on my Week 14 projection, it appears that I may not have been clear enough about the aim of this series. Just in case there is any doubt, THIS IS NOT INTENDED TO BE A STATISTICALLY RIGOROUS SIMULATION, LIKE THE MODELS USED TO FORECAST WEATHER OR PROJECT ELECTION OUTCOMES. It is more like a game of chance, using educated guesses, instead of 50:50 dice rolls.
My model is based on a single predictive variable. It fails to take into account a host of important factors, which any sensible attempt to accurately predict game outcomes would include, such as strength of schedule, starters lost to injury, key matchups, etc., etc. Furthermore, I have intentionally built in a stochastic element to generate occasional unexpected results, which weakens the predictive power of the projected outcomes.
The aim is to generate a weekly projection in which not all of the games turn out as expected, as happens every week of the season, but is not reflected in traditional projections. If the model is successful in predicting anything, it would be the number of upsets occurring each week, rather than individual game outcomes; but even that is beside the point. What I am really attempting to achieve here is to offer an alternative vision of Week 15 game outcomes and the ensuing wildcard race which incorporates one important aspect of reality that is missing from standard projections: the favored team usually wins, but not always.
TL;DR Methodology Section
The model uses accumulated point differential over the last five games to set the probability that teams will win their next game. A random number generator is used to simulate rolling the dice to determine which team wins each game. Using this approach, the team that has achieved the larger point differential over its last five games will usually win its next game. But there remains a real possibility that the underdog will pull off an upset.
Anyone who is actually interested in the details of how it works can read about it in the Week 14 edition.
Week 14 Predictions Revisited
Before we get to the Week 15 projection, let’s see how the new model did at predicting Week 14 game outcomes. Remember, last week I brought in the point-differential model to replace the original version which performed about as well as Blewitt kicking. The table below shows last week’s predictions with correctly called winners highlighted in blue, and incorrect calls in red.
11 out of 14 correct calls isn’t too shabby. The model predicted four upsets, defined in the usual way as a team with a better record losing to a team with a worse record. In Week 14, there were, four upsets, almost exactly as predicted: MIN (5-7) def. PIT (6-5-1); LAR (8-4) def. AZ (9-3); SF (6-6) def. CIN (7-5); CLE (6-6) def. BAL (8-4). The model even correctly predicted three of the four upset winners: Minnesota, LA Rams, 49ers.
The success of the model in predicting numbers of upsets might not be a fluke. The predictive variable, point differential over the last five games, places emphasis on how teams are playing right now, as opposed to earlier in the season when they might have been playing better or worse. This week’s results suggest that recent point differential might better reflect the relative strengths and weaknesses of teams that appear to be fairly closely matched based on season to date W-L record. Or it might just have got lucky. We will see.
At this point, I would like to reflect back on the comments on last week’s article by one poster who claimed to be a trained statistician and pointed out a variety of methodological flaws and faulty assumptions. First, if a trained statistician read through the methodology of these articles and left with the impression that this was a serious attempt at a statistically rigorous predictive model, they might want to consider calling up the institution where they did their training to inquire about a refund. Second, show me a model that does a better job of predicting NFL game outcomes with just one predictive variable and two lines of code.
Week 15 Predictions
After Sunday’s loss to Dallas, the WFT’s possible routes to the playoffs have narrowed considerably. They still control their own destiny, but they have almost no margin for error. The COVID Omicron variant outbreak adds a new element of unpredictability to the Week 15 outcomes, as teams continue to add starters and key contributors to their COVID reserve lists. I didn’t factor that into the model because it was just gaining steam when I ran the simulation and too hard to figure out how it would affect different teams at that point in the week.
The WFT travel to Lincoln Financial Field as 0.431 to 0.569 underdogs, according to the model’s handicapping. The random number generator’s dice roll was extremely close, coming up 0.435. That’s probably about a one point loss. Here are the complete projections for Week 15:
The LA Rams maintain their grip on the 5th seed wildcard despite an upset loss to a resurgent Seattle. San Francisco’s win over Atlanta moves them up into the 6th seed position, thanks to the WFT’s loss to Philadelphia.
Fortunately for the Team, the season doesn’t end at Week 15, because the loss to Philadelphia and Minnesota’s win over Chicago pushes them out of the playoff bracket altogether. There is still time to get back into the playoffs, but they will now have to overtake both of those teams. To do that, they will probably need to get a lot of players back from the COVID list over the next week or so, without losing many more.
In addition, they now have new competition from Seattle, pulling even at 6-8 due to their upset of the Rams, and the once moribund Panthers springing back to life with a huge upset win over Buffalo. The latter result might ordinarily seem completely improbable, but maybe not this week with COVID Omicron randomly disrupting starting rosters.
This week the model predicts three upsets. In addition to Seattle and Carolina, Pittsburgh pulls off a win against favorite Tennessee.
If the season ended at Week 15 in this scenario, the NFC playoff seeding would be as follows:
1st Seed - NFC North Champion Green Bay (11-3)
2nd Seed - NFC South Champion Tampa (11-3)
3rd Seed - NFC West Champion Arizona (11-3)
4th Seed - NFC East Champion Dallas (10-4)
5th Seed Wildcard - LA Rams (9-5)
6th Seed Wildcard – San Francisco (8-6)
7th Seed Wildcard - Minnesota (7-7)
Philadelphia (7-7) is now the first team out of the playoffs, trailed by five teams tied at 6-8: Washington, Seattle, Carolina, Atlanta, New Orleans. There are still a lot of ways the wildcard race could unfold, but if Washington loses another game, their chance of making the playoffs is more or less gone.
Week 15 Season Projection
Can Taylor Heinicke’s ragtag band of misfits bounce back from successive losses to Dallas and Philadelphia and overcome the coronavirus to work their way back into the playoff race? Let’s see what the model decides. This week’s projection for the rest of the NFC East Division race is as follows:
This is my least favorite season projection yet. Not only does the WFT drop three of its remaining four games to finish with a losing record and out of the playoffs, but Dallas wins out. Thanks to a Week 17 win over Arizona, the Cowboys move up the NFC playoff seeding.
As atrocious as this outcome might be to WFT fans, the NFC East Division results are plausible. Plugging the rest of the projected game outcomes into the ESPN Playoff Machine gives the following final outcome to the NFC playoff race:
1st Seed - NFC East Champion Dallas (13-4)
2nd Seed - NFC North Champion Green Bay (13-4)
3rd Seed - NFC West Champion Arizona (12-5)
4th Seed - NFC South Champion Tampa Bay (12-5)
5th Seed Wildcard – LA Rams (11-6)
6th Seed Wildcard – San Francisco (9-8)
7th Seed Wildcard – Minnesota (9-8)
Just missing the playoffs at 8-9 are Philadelphia, Seattle, Carolina, and New Orleans. The 7-10 Washington Redwolves go home to lick their wounds (assuming the new name is announced after Week 18 to distract from the losing record and the continuing Washington Post investigation of Dan Snyder’s indiscretions). This outcome gives the Wolves the 9th selection in the 2022 draft.
I hate this outcome, but I have to concede it is looking increasingly likely with so many starters and key contributors on the Reserve/COVID-19 list or out with injury. Meanwhile, of our remaining opponents, Philadelphia has only lost two players to COVID, Dallas just one, and the Giants a still modest seven.
On the positive side, it is less likely that Dallas wins all their remaining games, particularly with a Week 17 matchup with Arizona remaining on their schedule. Hopefully some team like the Rams knocks them out of the playoffs in the wildcard round.
I realize that I said at the start that this series is for WFT fans who are holding onto hope of a playoff finish, and then proceeded to spin a scenario where the Team crashed out of the playoffs. Well, that is how things turn out sometimes, when game outcomes are determined by unpredictable, random real-world events or a random number generator, filtered by estimated win probabilities. Most fans who have stuck with the team this century should be used to being disappointed.
Better luck next week.
Acknowledgement: Thanks as usual to James Dorsett for excellent editorial assistance
What would you do with the 9th pick in the 2022 draft?
This poll is closed
Best QB available - Matt Corral, Kenny Pickett, Malik Willis, Carson Strong
OT Evan Neal or Charles Cross
Pick an elite playmaker - S Kyle Hamilton, CB Derek Stingley Jr, WR Treylon Burks, CB Andrew Booth Jr, LB Nakobe Dean, WR Drake London, WR Jameson Williams, WR Garrett Wilson
Trade down for a ton of picks
Ask a Philadelphia fan. We’re picking in the 20’s.