Which NFL Teams Are Frauds?

A fraud isn't a bad team. They're just not who you think they are.

Five weeks in, our model has 16 potential frauds and 16 teams better than their record. A perfectly even split.

Last week we named two tests. The 3-1 Raiders went to New England and lost 45-17. The 3-1 Giants went to Washington and tied.

The biggest fraud is still the Giants. The most underrated team just lost in overtime: the 2-3 Rams. Here are the numbers behind the biggest calls.

Chart your team →

How the model makes each call

Fraud doesn't mean bad. It means a record is promising more than the team will deliver.

Every record comes with an expectation. The question is whether this team will live up to it. The model looks at three things: roster strength (how good the team is right now, with the players who are actually healthy), how they've played (not just whether they won, but how, and against who) and the schedule ahead.

If a team should finish worse than its record suggests, it's a fraud. If it should finish better, it's underrated.

Here's one signal anyone can check: every team's scoreboard against how it's played on offense and defense.

Teams below the line are getting out-gained. Since 2010, teams that outscored opponents over the first five weeks while getting out-gained won 54% of their remaining games. Teams that did both won 59%. It's a warning sign, not a verdict, which is why it's only one piece of the call. The Giants and Raiders both sit in that bottom-right corner.

Why the Giants are still the biggest fraud

Start with the quarterback. Jaxson Dart went down in Week 2. Jameis Winston has started every game since.

Look who they've played. Dallas, Tennessee, Arizona and Washington. Those four teams are a combined 4-15-1. The one good team they played, the Rams, beat them 28-6. Only two teams have had an easier schedule. Add it up and it's five games, plus-3. They were plus-3 last week too. The roster isn't there. The model rates this roster 26th. Play for play, opponents gain almost a yard more per play than the Giants do, and only three teams are worse. Vegas sees it too: on Sunday the 3-1-1 Giants are only 1.5-point favorites at home over the 1-4 Saints. The breaks went both ways this week. Down 27-13 in the fourth, Winston hit Isaiah Likely at the 1, and Likely fumbled it through the end zone. The Giants tied it anyway, then had two chances to win in overtime: a missed 46-yard field goal, then a Cam Skattebo fumble at the Washington 24 with two minutes left. Before Sunday, the breaks were all theirs. Two of their three wins ended on goal-line interceptions in the last 30 seconds.

John Harbaugh is first in the NFC East in his first year, and that's real. The model isn't grading the coach. It's grading the next 12 games, with this quarterback and this roster.

What if the breaks keep coming? Then a team getting out-gained by almost a yard a play keeps finding ways not to lose. That's the bet you're making if you believe 3-1-1.

Why the 5-0 Vikings are frauds

How can an unbeaten team be a fraud? A typical 5-0 team wins a lot from here. The model has Minnesota winning less than that. And this one is a real argument: the Vikings are plus-45, the 4th-best margin in football, and they lead the league in takeaways. But three things point to a team about to come back to earth:

Look who they beat. Green Bay, Chicago, Tampa Bay, Miami and New Orleans. No team has had an easier schedule. The last three wins came against teams that are a combined 2-13. Play for play, they're average. They're 15th in yards per play against opponents and 24th in points per drive. The takeaways are doing a lot of the work. The schedule is about to get brutal. After the bye, no team in football has a harder road. It starts with the Colts, then a trip to Detroit, then Buffalo.

If the defense keeps taking the ball away, this is the team on the board that proves the model wrong. Detroit and Buffalo in Weeks 8 and 9 will settle it.

Why the Rams are due

They're 2-3, but the model rates this roster 2nd in football. Play for play, they out-gain opponents by more than a yard a play. Only three teams are better. The record says more about who they've played than who they are.

The losses came to San Francisco, Denver and Buffalo. Those three teams are a combined 12-3. Two of the three losses were decided at the very end. At Denver, down four, Matthew Stafford was picked at the goal line with two seconds left. On Monday night, Buffalo blocked their 48-yard field goal in overtime, then kicked its own. Now the schedule gets easier:

This week the Rams are 13.5-point favorites at home over 1-4 Arizona. And look at the bottom-left corner of that chart: that's Minnesota. Last year at this point, the model called the 3-2 Rams underrated. They went 9-3 from there.

Why the Titans and Dolphins are frauds

A fraud at 0-5 means it can still get worse. For Tennessee and Miami, it probably will.

The Titans score 1.15 points a drive, the fewest in football, and the model rates their roster last. The Dolphins have trailed on 92% of their offensive snaps, the most in the league, and they've been outscored by 62, the worst margin in football.

The Chargers are 0-5 too, but the model has them finishing about where a typical 0-5 team does. That puts all three in the race for the top pick. Here's where the model has it today. Tennessee is on the clock first:

Run your own mock draft →

How often is the Fraud Meter right?

Three times out of four. Since 2010, its top three frauds and top three underrated teams after Week 5 were right 72 of 96 times. The last two seasons were rougher: 7 of 12. Each season was tested on a version of the model that never saw it. Here are all twelve, misses included:

The miss to remember is the same one as last week: the model had the Jaguars as a fraud at 4-1, and they went 13-4. So the Giants and Vikings are calls, not verdicts.

Updated weekly, proven in January

This week's tests: the 3-1-1 Giants are 1.5-point favorites at home against 1-4 New Orleans, and the 3-2 Browns are 4.5-point home underdogs to Baltimore. Lose both and the records start to come back to earth.

The standings will look different by January. When the season ends, we'll pull up this page and grade every call on it, misses included. Until then, the model updates every Wednesday. Your team might be on the stand next week.