Source: Jamie Squire/Getty Images
So here we are, the Final Four of the National Football League. At this point, there isn’t much in the way of surprise with the teams that have made it to this point. You have the two best teams in both conferences playing each other for the right to go to the Super Bowl in Atlanta in two weeks. Then the football fan inside of all us can mourn the idea of no more football until the first week in September several months hence!
Divisional Recap
I mentioned it on LinkedIn, but, in the Wild Card Round my predictions were 0-4. Then in the Divisional Round, I had all four home teams winning, which they did with very little drama. Maybe Philadelphia going up 14-0 in the 1st quarter of their game with New Orleans made things a bit competitive, but New Orleans came back winning 20-14 not allowing the Eagles to do much of anything after that initial quarter.
However, that’s not really how I score my models’ performance. I look at the Brier Score represented in the following table for the Divisional Round:
| Models | Brier Score |
|---|---|
| Original | 0.0503 |
| ELO (HFA=52.5) | 0.097 |
| FO+ELO | 0.1537 |
| Bayesian | 0.034 |
| Average | 0.0592 |
As it turned out, the Bayesian model running 10,000 simulations of the games, was the most accurate. That was helped primarily with the predictions for Kansas City winning to be at 80.97% and the Los Angeles Rams winning at 91.54%.
Conference Championship Win Projections
This weekend’s game feature the top four teams in the league as well as the teams that have the four highest scoring points per game for the 2018 season. I don’t think there’s really any reason to think that there will actually be much, if any, defense being played in these games this weekend. As it turns out, each conference championship game has already played out in the regular season.
Los Angeles Rams vs. New Orleans Saints (Sunday 2:05 PM CST on Fox) – O/U 56.5
In Week 9, the Rams went into New Orleans, and the game ended up an offensive show with 80 total points scored between the two teams. QB Drew Brees walked out of the game with 346 yards passing and four touchdowns. His primary target was Michael Thomas who finished with 211 yards receiving and one touchdown.
For the Rams, QB Jared Goff had 391 yards passing with three touchdowns and one interception. He was a little more generous with the ball, but WR Brandin Cooks had the most yards with 114 and one touchdown.
New Orleans won at home 45-35.
Can we expect the same out of this game? Absolutely! My projections have New Orleans winning 32-30 in another shoot-out.
| Models | Proj. Winner | Win % |
|---|---|---|
| Model 1 (Original) | New Orleans | 72.78% |
| ELO Model 2 (HFA=56) | New Orleans | 62.73% |
| Model 3 (FO+ELO) | New Orleans | 81.52% |
| Model 4 (Bayesian) | Los Angeles | 50.30% |
| Average | New Orleans | 66.68% |
| Vegas | New Orleans | 62.27% |
| Cynthia Frelund Picks | New Orleans | 56.00% |
In the above table, you’ll see the different win projections for my four models as well as the Vegas pick from Pinnacle and the NFL’s Cynthia Frelund’s picks. The one interesting thing to note is the Bayesian model. In the past, I would run 10,000 simulated games and see which team won the majority of those games. For this weekend, I ran 25,000 simulated games, and in 50.3% of those Los Angeles won. As it turns out, that’s the only model projection of an away team winning this weekend. Unfortunately, that’s basically just a coin flip, right? All the other models are confident (defined as 60% or above) that New Orleans will win.
At the start of the early evening hours on Sunday, I’m expecting that New Orleans will be the team lifting the George Halas Trophy as the best team in the National Football Conference!
New England Patriots vs. Kansas City Chiefs (Sunday 5:40 PM CST on CBS) – O/U 56
These two teams met in Week 6, but at the time, New England was the home team. New England led the game at Halftime with a convincing lead of 24-9. The Chiefs came roaring back right after that making it 27-26, New England at the end of the 3rd Quarter.
The 4th Quarter of that game was as exciting as one would hope for these two teams. New England scored first with a FG going up 30-26. Kansas City scored a TD making it 33-30 in their favor. New England answered back making it 37-33. Kansas City couldn’t do anything with their next possession, so NE kicked another FG making it 40-33 with 3:15 left on the clock. Kansas City then tied the game with a one play 75 yard pass from QB Patrick Mahomes to WR Tyreek Hill!! Unfortunately that left way too much time on the clock as QB Tom Brady and the Patriots do what they do and scored a FG with time running out.
What an exciting game that was!
So what can we expect Sunday night? This time around Kansas City has the home field advantage. I’m still expecting another high scoring game with Kansas City winning 31-30 in another shoot-out.
| Models | Proj. Winner | Win % |
|---|---|---|
| Model 1 (Poissson) | Kansas City | 75.03% |
| ELO Model 2 (HFA=56) | Kansas City | 60.02% |
| Model 3 (FO+ELO) | Kansas City | 54.40% |
| Model 4 (Bayesian) | Kansas City | 95.05% |
| Average | Kansas City | 71.13% |
| Vegas | Kansas City | 62.27% |
| Cynthia Frelund Picks | Kansas City | 60.00% |
On Twitter last week, I mentioned that, in my Bayesian 25,000 simulation model, I had Kansas City with a 95.05% chance of beating New England. That is pretty darn confident…for a machine. Intuitively most people would probably go with the FO+ELO model which gives KC a 54.4% chance of winning. You really just can’t count New England out. They have a fantastic coach and a really good and seasoned Quarterback. Being that confident in KC just doesn’t seem right. But that’s what the models say. Overall, I’m giving KC a 71.13% chance of hoisting the Lamar Hunt Trophy at the end of the night.
Enjoy the weekend!
*Statistics used in the FO model come from Football Outsiders.