Since the AFL-NFL merger in 1970, only four playoff games have had spreads of at least 15 points. None of the underdogs team wound up winning, and only one managed to cover the spread.

You could say the odds are stacked against the Houston Texans this weekend.

They enter Saturday's game against the New England Patriots as 15-point underdogs with the victor advancing to the conference championship. The Texans got a win last week to advance, but facing a Derek Carr-less Oakland Raiders team is a whole different ballgame than going toe-to-toe with Tom Brady in the playoffs.

Right now, numberFire's algorithms give the Texans 18.3% odds of victory in Foxboro. Because randomness abounds in the NFL, there's always a chance things could break their way on a given day. But how, exactly, do they tap into that 18.3%?

Let's try to figure this out using a couple of different tools, including numberFire's Net Expected Points (NEP). This is the metric we use to track the efficiency of teams and players with the team totals being adjusted for strength of schedule. There's a big difference between a three-yard completion on 3rd and 2 and one on 3rd and 4, and NEP helps quantify those discrepancies.

Is there a path to victory for the Texans against the Patriots? Let's check it out.

A Tough Test for Brady

You don't need me to tell you that, if the Texans are going to win, they're going to have to force turnovers. That would likely be obvious. The problem is that the Patriots never turn it over, doing so just 11 times the entire regular season. That's tied for third-best among all teams since 1970. This one task isn't as easy as it seems.

Thankfully for the Texans, they may have one of the units best suited to generate said turnovers that the Patriots have faced all year. They finished the year ranked sixth in Adjusted Defensive NEP per play. In the time after Brady's suspension ended, the Patriots faced only one team ranked higher than that, and that was the Denver Broncos. They help contribute to this formulation of a blueprint.

In their matchup with the Patriots, the Broncos kept the game within one score all the way up until late in the third quarter, eventually falling, 16-3. Brady didn't throw any interceptions, but he did fumble twice (both of which were recovered by the Patriots) and pass for just 188 yards on 32 attempts. That's something the Texans could overcome.

Unfortunately, the Texans aren't the Broncos. Denver easily had the league's best pass defense this year, and that game also wasn't in Foxboro. It does help show what the Texans need to do, but it doesn't get us much closer to finding them a victory.

That doesn't mean the two teams are without similarities. A big part of the Broncos' stout pass defense is their ability to get to the quarterback, an area where the Texans can excel when both Jadeveon Clowney and Whitney Mercilus are healthy. They got to Connor Cook three times last week, and the Raiders lost 4.69 NEP on those sacks. They may not force Brady into interceptions, but sacks can result in high-upside plays, as well, and that may be the best avenue for success the Texans have.

We can see a similar thread here in turning to numberFire's game profile, which is available to all premium subscribers. This shows us similar games to this one since 2000, which can provide us with a glimpse at what the Texans need to do to snag a win.

When we scroll to the strongest predictors section, the top match has 94.20% similarity. It was between the Seattle Seahawks and Los Angeles (then St. Louis) Rams back in 2015 when the Seahawks were at home and 12-point favorites. This was during their run of destruction when Russell Wilson and Doug Baldwin scorched the league and bathed in efficiency, similar to what Brady has done this season.

The Rams won that game, 23-17. It's the only game of the 17 most-similar games that has the Texans' equivalent winning, but at 94.20% similarity, we'll take it.

A big part of that victory for the Rams was getting Wilson and the Seahawks in negative game script. The Rams took a 3-0 lead on their second drive, and then they got their negative game script on the Seahawks' next possession.

The NFL's embedding laws are semi-archaic, so you'll just have to take our word for it when we say that Akeem Ayers took a fumble from Will Tukuafu 45 yards to the promised land to put the Rams on top, 10-0.

While it seems unlikely that Bill Belichick will be handing out rush attempts to their equivalent of Tukuafu, it just shows that fluky things can happen early in a game, putting Brady and the Patriots in a hole. That's not something they have faced much this year.

Of Brady's 432 pass attempts, only 77 have come while trailing. For context, Landry Jones -- who started just two games for the Pittsburgh Steelers -- had 76 attempts while trailing. Quarterbacks have lower efficiency when they're behind (they average 0.09 Passing NEP per drop back compared to 0.13 when tied or leading), meaning that if you want to neutralize a quarterback as good as Brady, you need to get that puppy in a hole.

The problem with this is that, even if the Patriots trail, the Texans' offense will likely need to score at some point. Can they do that with Brock Osweiler at quarterback? It may not be easy, but it's certainly a possibility.

Exploiting the Patriots' Weakness

Prior to the postseason, we went through and assessed the weakness of each playoff team to decipher what their Kryptonite would be. For the Texans, it was obviously Osweiler. For the Patriots, though, it was their pass rash, and that could provide an opening for the Texans.

The Patriots finished the year ranked 28th in Sack NEP per drop back, which tracks the expected points opposing teams lost on sacks on a per-drop back basis. The Texans were 12th best on the offensive side, meaning they were largely able to give Osweiler time. He obviously didn't do much with that, but it is at least an encouraging sign.

It becomes even more critical when you see just how bad Osweiler was under pressure.

The Texans were at their best here last week. They didn't allow any sacks against the Raiders despite facing the likes of Khalil Mack and Bruce Irvin. This allowed Osweiler to finish with 3.61 Passing NEP, his highest output in a game since Week 11 and his fourth-highest mark of the season. It's not much, but it is at least a glimmer of hope.

We know the Texans would prefer to grind it out on the ground, but when you're a 15-point underdog, that's not usually an option. If they want to win this game, they'll likely need Osweiler to rise to the occasion. The Patriots' lack of a pass rush provides a window there, and it means the Texans do, in fact, have at least a chance -- however slim -- to pull off the upset this weekend.

Conclusions

We've seen some bonkers stuff happen in the playoffs in the past. Mark Sanchez went to back-to-back conference championship games, Trent Dilfer won a Super Bowl, and Peyton Manning finally got a second ring while playing the worst football of his career. Why can't the Texans pull off the upset on Saturday?

Obviously, the odds are very much stacked against them, and most scenarios result in a victory for the Patriots. But with how much variance there is in just 60 minutes of football, it would be foolish to say they don't at least have a chance.

The key -- on both sides of the football -- will be the pass rush. Clowney and Mercilus will need to force Brady into mistakes that he doesn't usually make, and left tackle Duane Brown and company will need to keep Osweiler upright. If both of those things happen, then this becomes anybody's ballgame.

There are plenty of "ifs" involved here. Most of them have higher likelihood of breaking in the Patriots' favor than the Texans'. But there's always a possibility that everything skews toward the Texans, in which case they could find themselves in the AFC Championship Game while the Patriots wonder what could have been.