Through the first six weeks of the year, the New England Patriots' pass defense was hemorrhaging yardage. Each of the first six teams to face them threw for at least 292 passing yards, including three games of at least 330 yards. The last game in that stretch was against Josh McCown and the New York Jets. That's not a championship-caliber unit.

Since then, though, it has been a completely different story on the back end. Over what is now a 12-game sample, the Patriots have allowed an average of 212.2 passing yards per game, maxing out at just 273. By this one metric, they have seemingly gone from the worst pass defense in football to one of the best.

Now, in order to take home yet another Super Bowl trophy, all they have to do is stop Nick Foles, a passer who has made just six starts in the past two years and wouldn't even be starting if it weren't for an injury to Carson Wentz. That would seem to indicate the Pats' secondary is in line to continue their marked improvements from the start of the year.

It just ain't that easy.

It turns out that surface-level statistics can be fluky in smaller samples, and that's what we get when we omit what was easily a team's worst stretch of the entire year. And if we do this, we're going to overrate this Patriots defense. Those same small samples could also lead us to underrating Foles. This matchup is a lot closer than it appears at first glance.

To explain why, let's take a deeper look at this pairing with the help of numberFire's Net Expected Points (NEP). NEP is the metric we use to track the efficiency of both teams and players. The team metrics are adjusted for the strength of opponent, which will be important later on. Overall, though, NEP tracks the expected points added on each play, illustrating the cavernous difference between a three-yard completion on 3rd and 2 and that same three-yard completion on 3rd and 4. Using NEP allows us to see which players and units truly operate in a desirable manner over the course of the year.

Let's start things off here by addressing the Patriots' defense and why those baseline numbers are misleading.

A Blessed Schedule

Let's go back to our 12-game sample from above. You would think that in that stretch, the Patriots would have faced a good sampling of quarterbacks of differing ability levels. While they did see some decent competition, a lot of it was just pure garbage.

In those 12 games, the Patriots faced four quarterbacks who finished outside the top 30 in Passing NEP per drop back (out of 45 quarterbacks with at least 100 drop backs during the regular season). That means one third of their games were against the league's worst signal callers, including Bryce Petty, who was dead last on that list. That'll certainly help pad the stats a bit.

On the flip side, only three of those quarterbacks were in the top 15 in Passing NEP per drop back. Two of those -- Matt Ryan and Philip Rivers -- were back in Weeks 7 and 8, respectively. That means they've faced just one top-15 passer in the past 10 games, including the playoffs. This was not a tough road.

The one top-15 passer they did face in that stretch was Ben Roethlisberger in Week 15. Roethlisberger finished that game with 0.42 Passing NEP per drop back. That was almost double his season-long mark of 0.23 and well above the league-wide average of 0.07. He did that with his best asset in Antonio Brown in the locker room for most of the game. You can absolutely pass on this team if you're good enough.

That has continued into the playoffs, as well. Even though neither Marcus Mariota nor Blake Bortles tore the Patriots apart, they both finished their respective games with a higher Passing NEP per drop back against the Patriots than they had in the regular season.

They certainly weren't great, but they were both above-average passers against the Pats and outperformed their season-long metrics. That matters.

As mentioned, we can take this schedule into account when looking at team NEP numbers. If we narrow the scope of our search to just be from Week 7 on -- when the Patriots seemingly turned things around -- they rank 14th against the pass, according to numberFire's schedule-adjusted metrics. That's certainly not bad, but it's worse than the Philadelphia Eagles' sixth-place ranking. It's also worse than the Atlanta Falcons (12th) and Minnesota Vikings (1st), the two teams Foles conquered to make it to this point. This doesn't really look like a cakewalk anymore.

But is Foles truly good enough to take advantage of the matchup? After all, the Patriots did a decent job of suppressing inferior foes in the regular season, and if Foles were to qualify as another sub-par passer, then this would all be for naught. Let's look at Foles' metrics to try to answer that question.

The Conflicting Data on Foles

If we were to look just at the regular season, then Foles would probably be just next in line to bolster the Patriots' defensive resume. Once you add in what he has done in the playoffs, though, things look a bit different.

In the regular season, Foles was a below-average passer, finishing with -0.01 Passing NEP per drop back, ranking 26th out of 45 qualified passers. That's below both Mariota and Bortles, and neither of them truly gashed the Patriots.

But we're also limiting ourselves if we look just at the regular season. Foles had 106 drop backs then, and he's at 65 drop backs in the playoffs. Why would we willingly ignore 38% of the data?

So, instead, let's combine all 171 drop backs for Foles to see how he grades out. Below are the blind metrics of two separate quarterbacks from this year. Success Rate is the percentage of drop backs that increase the team's expected points for the drive.

The differences between these two guys aren't overly massive. Although Quarterback A has the edge in per-drop back efficiency, the additional nudge in Success Rate helps Quarterback B.

Quarterback B is Foles. Quarterback A is pre-injury Wentz. Foles doesn't look like a scrub any more.

Foles' 0.20 Passing NEP per drop back would have ranked ninth in the league if it were for the full season, right between Alex Smith and Deshaun Watson. What do those two have in common? They both torched this New England defense, combining for 669 yards and 6 touchdowns on 68 attempts. If they can do that, why can't Foles succeed on Sunday?

But even this may not be giving Foles all the credit he deserves. Two of his appearances this year were in relief of Wentz, and coming in cold off the bench is a difficult ask for a passer. numberFire's Brandon Gdula noted earlier this week that if we look just at the four complete games Foles has played, his metrics are right in line with Tom Brady's full-season marks. This isn't to say that Foles is as good as Brady, but it would be foolish to say he's incapable of picking this secondary apart.

As such, we've got someone in Foles who is performing better than the league-average quarterback facing a pass defense that has benefited from an easy schedule of late. Brady, on the other hand, is facing an Eagles defense that has a ferocious pass rush and performed well from an advanced metrics perspective the second half of the year. That would seemingly set things up for a tight game in Minneapolis.

Foles isn't Wentz, and his success this year is still based on extrapolating numbers from a small sample size. And we do need to give the Patriots credit for performing well against lesser competition. But with the Patriots entering as 4.5-point favorites, Vegas doesn't seem to believe that Foles and company will be able to keep pace for the entirety the game. Based on both Foles' effectiveness of late and the underlying metrics on the Patriots' defense, it appears as if he could continue to shatter expectations once the game kicks off on Sunday night.