Elo ratings predict double digit wins for 2020 Dallas Cowboys

The Cowboys managed to come away with a win on Sunday by the narrowest of margins, and a lot of things had to go just right in the final minutes of the game for the team to end up with a 1-1 W/L record. Had they not, we’d be discussing an 0-2 record and a must-win game against the Seahawks next week, as only 3% of teams with an 0-3 start have historically made the playoffs.

But as Bill Parcells was fond of saying, “you are what your record says you are,” and that’s certainly true of the 2020 Cowboys. And it’s also true for all the other teams in the league, some of which look much less formidable after an 0-2 start than they did heading into the season.

The Vikings (0-2) and Eagles (0-2), for example, are both on the Cowboys’ schedule and don’t look anywhere near as daunting as some some had seen them entering the season. And other parts of the 2020 schedule may also look different after two weeks of play than they did when the schedule was initially released.

One site that provides an up-to-date look at the remaining schedule is fivethirtyeight.com, where they use an Elo points system to calculate an expected point spread and win probability for each coming opponent while factoring in the first two weeks of play.

Elo ratings are a relatively simple rating system originally developed by the physicist Arpad Elo to rate chess players. The system can be easily adopted to other head-to-head competitions, and that’s just what Nate Silver of fivethirtyeight.com did in 2014 when he extended the ratings system to football.

Elo ratings estimate each team’s skill level using only the final scores of previous games and the locations of each game. That skill level is translated into a points system where an average team has 1,500 points, and teams usually fluctuate somewhere between 1,300 and 1,700 points. The Cowboys for example are currently ranked 12th overall with a middling 1,567 Elo points.

The Cowboys’ next opponent, the Seahawks, are ranked third overall with 1,610 points, which gives Seattle an Elo point spread of -3.5 and a win probability of 62% next Sunday. And what can be done for the Seahawks can be done for all of the remaining 2020 opponents. Here’s what that looks like, according to fivethirtyeight.com:

The Elo ratings system currently has the Cowboys favored in 11 of their remaining 14 games, which would translate to a 12-4 season record. And here’s how that could go:

• 5 wins over the next six games: The Elo system has the Cowboys as underdogs in Seattle, but then has them favored in five consecutive games. 6-2 at the halfway mark of the season? Most Cowboys fans will gladly take that, even if some will still find things to gripe about.
• A 2-2 stretch midseason: The Cowboys are currently underdogs against the Steelers (WK 9) and Ravens (WK 13), but are favored against the Vikings and Washington. That leaves them at 8-4 after after 12 games, with a playoff berth firmly in sight.
• 4 wins to close out the season: The Cowboys are currently favored in their last four games to finish the season 12-4 and ride the hot hand into the playoffs.

But that projected 12-4 record is not written in stone, not by a long shot.

The Elo rating system always improves a team’s Elo rating after a win, just as it always reduces a team’s rating after a loss. And the system also factors in margin of victory. A blowout loss against the Seahawks on Sunday, for example, could lower the team’s rating to such a degree that the win probability for a game like the Week 15 match against the 49ers, or the road games in Minnesota (WK 11) or Philly (WK 8) might change so much as to make the Cowboys the underdogs in those contests. And before you know it, we’d be at a projected 9-7 W/L record.

Of course, that works both ways: If the Cowboys win in Seattle, the win probability for the remaining games will increase, and the Cowboys might even end up being favored against the Steelers, depending of course on how the Steelers play on Sunday.

Ultimately though, because we’re dealing with win probabilities, mathematically we can’t just add up individual probabilities to arrive at a W/L record. Which is where the 538.com algorithm comes in: They simulated the remainder of the season 100,000 times to arrive at a projected W/L total for the Cowboys, and that currently sits at 9-7, with the closest divisional rivals at 6-10 (Philly and Washington).

In any case, the season is far from over, and it can develop in wildly different ways. Where do you think this team is headed in terms of W/L? Let us know in the attached poll or in the comments below.