Last week was great. We went 10-6 in what I'd call the hardest week of the season to predict, finishing a brilliant +74 units on the season. We also nailed one of the big upsets, the Giants beating the Redskins, while in all honesty, our other pick (the Eagles) could easily have gone in our favor had they actually been able to score a single TD in the entire game. Regardless, in a week where just predicting who will play for each team is difficult, we improved our yearly score.

Which is all great, but now it's business time again and all of the games are meaningful again. There are no more questions of resting players, everything is full throttle for the remaining 12 teams in contention for the Super Bowl, and we're with you right the way through.

Here's our results on the year so far:

Week 17:

Criteria

Units

Top 5

+15

Top 5 + Bottom 5

+30

All

+20

You can discard the top 5 and bottom 5 stuff from last week (although I won't obviously because they were great!) and focus on some serious successes in that 10-6 that contradicted the consensus picks. I want to just show our working out and how they went. We had 3 broad categories last week:

Upsets (1-1)

We nailed the Giants, our top pick, while the Eagles sadly pulled Nick Foles early when it became clear the game was not going to yield anything particularly useful. That game was a moral victory, but sadly it counts for nothing in the real world.

Underdog Covers (3-3)

Our logic was simple - NFL teams don't do a great job of covering in week 17 against high lines when the underdogs play their starters, but what was unusual about week 17 was the volume of games that met that criteria. Also, two of the favorite covers were the biggest spreads in week 17 for 3 years, as the Pats covered 15.5 vs the Jets and the Vikings covered 11.5 vs Chicago. Put simply, there was no statistical precursor for those games, so it was useful to see such overwhelming favorites respond as such. It was also worth factoring in that unlike a lot of teams, the Pats take great pride in not relenting, and in fact had Tom Brady and Rob Gronkowski playing late into the game.

Favorite covers (6-2)

This was the big winner for us. The logic was broken down into three types of game, but what was immediately clear when I began the breakdown was that the Lions and Colts were the best picks of the week. Part of Upset Watch's remit is not only to predict the upsets, but to point out where favorites should cover too. This week differed in that there was a specific reason to believe that some teams would cover small spreads. Indianapolis, Detroit, San Francisco, Buffalo and Tennessee all covered against opposition who had nothing to play for, while Atlanta who had more to play for covered against the Panthers at home. Our only upset was that the Chiefs failed to win against Denver, and Seattle were upset at home against Arizona.

Regular Season Overall:

Criteria

Units

Top 5

+8.5

Top 5 + Bottom 5

+26.5

All

+74

Last week we got some great results that only improved an already fantastic season so far. I'm not getting carried away here though. Even though this is big encouragement for us, we still have to adapt our strategy for the playoffs, where the trends and analytics we used in the regular season have to be adapted to take into account the 'loser goes home' nature of the contests.

Wild Card Methodology:

Methodology: I looked at the last 5 years of wild card results. I know that there is a case that can be made to go back further, but the last 5 years are the most relevant, as players are largely still active and teams in some cases even have the same coaches and QBs. That's far less common going back 10-15 years, besides which, the game has changed both on and off the field to a degree that is notable enough that we can't learn a great deal from what Ben Roethlisberger did as a rookie in 2004/5.

The first criteria I measured was the size of the spread, because all 4 wild card matchups are currently 6.5pts or above. This showed something very interesting: Teams don't tend to upset the odds and win very often when the spread is higher than the average of 4.95 points:

Not only do they not tend to win, but they tend not to cover either. Clearly, we need to base this on more than simply the size of the spread, however. My next task was to establish momentum. I started by collating records at the end of the season heading into the playoffs, however what quickly jumped out was that for the most part, these being winning teams who have made the playoffs, there wasn't a great deal of separation.

So the next factor I considered was how the emotion and momentum of earning a playoff spot in week 17 translated into the wild card round. This year 3 teams (Buffalo, Tennessee, Atlanta) sealed their berth in the final week of the season, and all will start their playoff run on the road. What I found here was far, far more interesting...

Oh boy. So here's how we read this: 4 road teams have clinched in week 17, and they have upset the odds 3/4 times, while home teams that clinched in week 17 have also won 2/3. Only once have teams faced each other having both clinched in week 17 (New Orleans at Philadelphia, 2013). That makes the 'week 17 clinchers' 4-1-1 vs the spread in that time. More importantly, there are interesting reasons behind the non-covers.

The 2012 Vikings lost starter Christian Ponder after clinching in week 17, so Joe Webb started their playoff game against the Packers. Of course, the Packers easily beat the Vikes and covered the 11 point spread. The 2013 Eagles lost at home despite covering in week 17 with their 'starter' Nick Foles falling into that void of technically being the starter, but having been a backup and having been replaced during the season (Mike Vick started 2013 for the Eagles).

Finally, the Niners overcame the Packers who had clinched in week 17, albeit without covering the spread. However, they were actually road favorites, who have an exceptional record. In fact, only one road favorite has lost a wild card matchup in the last 5 years.

So where does this leave us? Well, the biggest takeaway is that road underdogs who clinched in week 17 are 3-1, and in fact they are 3-0 when they have a starting QB playing rather than a backup. I'm not sure that 3 upsets are on the cards, but the analytics do support that the Titans, Falcons and Bills are the picks. I don't have the courage to back all three, but I like the following:

Wild Card Picks:

1 - Atlanta (+6.5) @ LA Rams

This is the best chance of an upset primarily because the Falcons have a better team than most underdogs. Meanwhile, it's always worth considering that Jared Goff will start his first playoff game and the Rams - as good as they are - may be set up to fail against a team that have gone deep into the playoffs a number of times collectively. Matt Ryan knows how to play this game, and the Falcons are largely the same team that made it to the Super Bowl last season.

I do like the Rams, and there's a chance they have one of those enormous blowout wins that makes the NFL stand up and collectively take note (as if they weren't already...), but we are looking for upsets, and frankly, the experienced, playoff veterans of Atlanta have a much better chance of at least covering than most teams who find themselves underdogs by 6.5 points.

3 - Buffalo (+8) @ Jacksonville

There are two factors at work here, first of all the Bills have broken their playoff hoodoo in emotional, slightly fairytale fashion (it wouldn't be the Bills if it didn't involve scraping in on the back of someone else's achievements in week 17). I think the emotion favors the Bills slightly despite the Jags also having their first playoff game in 10 years upcoming, and they are seriously lacking at receiver as I pointed out last week. People have cited that the Jaguars rebound well, often blowing out teams after a loss, but I can't emphasize enough that their starting wide receivers are likely to be rookies Keelan Cole and Dede Westbrooke, with only Marqise Lee (ordinarily the Jags' third receiver) likely to even be close to playing. last week I liked the Niners to beat them, and this week I think the Bills will at least cover - even if the Jags find a way to get it done.

3 - Tennessee @ Kansas City (-8)

The Titans aren't great, and they lost starting RB DeMarco Murray today to injury. A lot of people might consider this addition by subraction, allowing Derrick Henry the chance to finally run free as the starter, but Henry has already done so and largely underwhelmed. I'm a big believer that a power back always needs a complimentary guy to keep defenses honest, and Henry has some odd stats. He has two 100yd rushing games this season, yet in both he had a run over 70 yards. While not discounting those games, the rest of his season was not just average, but bad. He averaged 2.9yds per carry in the rest of the games he played in, managing <2.5yds per carry on no less than 5 occasions.

The Chiefs are of course, an odd one, but they have turned a corner with Matt Nagy calling the plays and covered their final 4 games in wildly varying scenarios. They covered spreads of -11, -4.5, PK and +3 in their meaningless season ender. I like that they'll be fresh and I particularly think that their offense will enjoy the matchup against Tennessee at home.

4 - Carolina @ New Orleans (-6.5)

If you use statistics and algorithms, you have to accept that sometimes, you have to pick against your gut. My gut was that this is a big spread for New Orleans to cover, but these two teams have met twice already as they're divisional opponents, with the Saints winning both matchups by 10 points at home and 21 points on the road. Neither team has been convincing vs the spread in the run-in, but the Saints have a ridiculous offense when they need it, and I think their regular season changed when they became almost assured of a playoff spot relatively early.

But this team can rebound like no other, and I fancy the Saints are the better of the two teams here, with the more explosive weapons and an underrated defense. Yes, Cam Newton is an X-factor, but his biggest weapon is Greg Olsen who simply never looks comfortable out there at the moment. He almost pulled up lame in week 17 and if that happens on Sunday, the Panthers lose a big part of their equation for winning. More importantly, they are strong against the run, but the Saints found ways to both run and score against the Panthers in the previous two matchups, notching almost 150yds and scoring 4 TD's in both games. Meanwhile, Cam Newton didn't manage to throw for 200 yards in either game and had a disastrous 0TD/3INT showing in week 3.