Matthew Coller at the Super Bowl: The buzz from radio row
LAS VEGAS — Everything regarding the Super Bowl always seems to epic on social media, including Radio Row. It comes across like a red carpet of media and NFL celebrities prancing around with greatness abound — like the movie Zoolander for football. Beautiful people everywhere just hanging out and talking ball. Oh, how magical.
In reality, it’s more chaotic than thrilling. Several hundred radio hosts, former/current players and their handlers all walking around confused about where to go and what to do. Producers wandering around hunting for guests, organizers trying to point people in the right direction and everybody trying to find out where the coffee and donuts are located. No flashing lights or club music, though there are three slot machines at the entrance. Had to have some Vegas touch. Everyone took pictures of the slots but chose to do their gambling elsewhere.
Radio Row is inside the Mandalay Bay convention center right next to Allegiant Stadium where the Super Bowl will be played on Sunday. I arrived at 7:30 a.m. and wandered through a labyrinth of hallways to pick up my credential and headed over to where all the radio stations were set up. Local radio has tables with ethernet cords and little welcome baskets with snacks are stuffed back-to-back in the middle and the outside of the floor is lined with the heavy hitters — NBC, FanDuel, Draft Kings and XM Sirus Radio. Some of their setups must have required a tractor trailer.
Matthew Coller (right) with Sirius XM Radio at the Super Bowl in Las Vegas.
Purple Insider
I’m here for two days to promote my book Football is a Numbers Game. I’d like to thank Dan Campbell for failing on fourth down (twice) in the NFC Championship because talk hosts seem to be particularly intrigued by the analytics discussion. By intrigued I mean asking me “WHY DIDN’T HE TAKE THE POINTS?!?!?” Combine that with the fascinating Vikings quarterback situation and there is a lot to talk about with these content-desperate folks.
With a bevy of interesting topics and conversations taking place, I figured I’d take everyone through all my appearances and the types of questions I was asked and what I ended up taking away from each discussion…
8 a.m.: Solomon Wilcots and Ed McCaffrey, SiriusXM
Before going on the air, I kiddingly asked McCaffrey if Howard Griffith or Byron Chamberlain were going to be here too.
“Oh, do you know those guys?”
No, I do not but I figured it would be a better ice breaker than it was. I mentioned to him that I covered Gary Kubiak and he still seemed a little puzzled why I was throwing random Broncos names at him. So I dropped Tom Nalen’s name in the interview anyway.
Once we started, Wilcots, a former Bengal and Viking, wanted to know where we stand in the analytics boom surrounding the NFL. I think we are in a similar spot to where Major League Baseball was about 10 years ago. It is clear that the numbers are gaining trust, which we see in the massive increase in fourth downs and the focus on efficiency in play calling with more pass-first and play-action offenses, but we are nowhere near a consensus on how to interpret the information and put it into action. Some teams are far ahead of others, going for fourth downs all the time and leaning into team building theories like drafting premium positions. Others are still trying to eyeball it. Some analytics departments are growing quickly while others might have one or two people trying to keep up. In a decade the teams that are behind will look back in disbelief that they operated on a shoestring analytics budget.
It’s amazing when you think about how the origins of the recent analytics boom in football are recent enough for everyone to remember. The Philadelphia Eagles leaning into aggressive in-game decisions to win the Super Bowl in 2017 was a Big Bang type moment for teams understanding that they could no longer guess on in-game decisions. But it wasn’t just the fourth downs. In my book there is a story about the Eagles taking Peter King through their game-winning touchdown in the Super Bowl and how PFF’s data influenced the play. The analytics use in football is almost entirely taking place behind the scenes in the game planning phase where coaches use data to pinpoint tendencies and weaknesses and we are only catching traces of it in the decisions and efficiency markers. It’s oftentimes not as easy to spot as baseball’s home-run-or-strikeout mentality or basketball’s 3-pointers.
McCaffrey wanted to know if there was a way that analytics can identify momentum. What we have numbers wise is win probability models that show where games turned and the odds of winning at any given time. The funny thing about the Lions’ infamous failed fourth down in the NFC title game was that the win probability only sunk from about a 90% chance to 85% chance for the Lions to win. What really swung the game was the ball bouncing off a dude’s head and being caught for a touchdown. Was that a force of momentum or a random event that Dan Campbell never could have anticipated when he went for it on fourth down?
Of course, real people play the game. Things like pressure and anxiety matter. The failed fourth down may have caused an “uh oh” feeling on Detroit’s sideline that impacted what happened afterward. They fumbled and dropped another third-down pass on the following drives. Those things may have been a force of feeling the weight of the moment. It’s not anti analytics to say that there can be human explanations for results that go beyond what we can track with numbers. Does that mean Campbell should have kicked the field goal? Well, would the feeling have been different if he missed the field goal?
I talked about how it’s odd that the go-for-it decision struck a nerve with people who want to lash out against analytics because the numbers did not have a clear answer on either fourth down that Campbell called for. He was aggressive all year and it worked, plus he probably felt that they would need more points to hold off the awesome 49ers offense. He also didn’t trust his kicker. Not trusting your kicker isn’t exactly analytics — ask Mike Zimmer.
It’s always great to talk with former players about numbers because we can talk about how to combine what we see in the data versus what they see on the field.
9:00 a.m. : Bill Michaels, 96.7FM/1670AM The Zone
Bill is old school, yet a believer in analytics. He understands the extra layers that it has added for media people like ourselves. Imagine a world where the only numbers we had to tell the story of the Vikings and Packers were touchdowns, interceptions and yards. It wasn’t too long ago that we couldn’t give you anything about how often a team ran play-action passes or their success with screens or how many pressures a left tackle allowed. While none of it is perfect, it has opened the door for so much better understanding of what’s actually happening on the field and many more questions for coaches and gives us a cross check on their answers. In fact, one person suggested in my book that explains why coaches want to downplay Pro Football Focus in public — they don’t want the questions about the struggling left guard’s pressure rate.
What I found surprising is that Bill is a little skeptical about Jordan Love. He mentioned that he needs to see two full years of Love before deciding that he’s the real deal. The standard is different in Green Bay, isn’t it? If Love was in Minnesota they would have given him the keys to the cities by now. But he has a point. Love’s poise and arm talent impressed me but if you had watched Carson Wentz in 2017 there is no way you would have dreamed he would one day be backing somebody up. Small sample sizes are a thing for quarterbacks. Still, the Vikings’ brass should treat their plan like Love is going to be good until 2034.
We discussed how Love handled his first season. We have seen many talented QBs get frustrated early in their career and struggle with confidence. Love found his way midway through the year and handled the problematic first half of the year without collapsing and then managed success with equal grace. The Vikings should be most concerned with how Love and Packers head coach Matt LaFleur paired together. “He operates the offense,” Bill noted. He also had a running game to work off. Maybe the Vikings should try that with their next young QB, whenever that person arrives. Snark aside, it’s a good example of a coach adapting to the circumstances.
9:30 a.m.: Bob “The Moj” Marjanovich
Bob used to work for a Vancouver sports station and one day their company decided to randomly fire everyone so he started his own thing and he’s loving it. Sounds familiar. He is a hardcore football guy who was not thrilled with Campbell’s fourth down shenanigans and mentioned that his CFL coaching friend did not subscribe to PFF’s new CFL grading.
Nonetheless, we had a good discussion about how we have decided to bucket all aggressive decisions into a monolith that is THE ANALYTICS. By the actual numbers it was much more of a 50-50 type call and Campbell went the way he had been going all year: All gas, no brakes. Still, there is a particular frustrating that you can sense with some folks about analytics that seems to have been unleashed by a single fourth down that went wrong. I’m not sure why. You would think that the most FOOTBALL people wouldn’t want to side with a bad kicker in this instance.
It reminded me of what Michael Lopez of the NFL said about analytics and football: The numbers have leaned toward a more exciting game. More passing, more fourth downs. It feels like people should like that. Yet the numbers have also given defenses more ammo to battle back and the average point totals per game have not changed over the years. The fourth downs give and take away. That’s compelling sports, man.
Anyway, I explained the layers of the grading system. Some coaches may not like it but PFF’s grades are checked by folks who have proven they can track games at a high level of accuracy and then the grades are cross checked by former players and coaches. Do they know the assignment? No. Neither do opponents. One GM told me that using the grades is all about understanding them, what they can tell you and what they cannot. Some teams like to use PFF as a comparison tool to their coaches to see if there’s a gap and then look for why that might be.
Bob and I had a really good dialogue about it all. Plus, the CFL is awesome.
12:15 p.m.: SportsGrid with Scott Ferrall
Scott is a legend. He hated Dan Campbell’s fourth downs more than anyone. He asked me if Chuck Knoll would have gone for it. I reminded him that in the Super Bowl against the Cowboys, Chuck went for 4th-and-8 with 1:20 left to go and the lead despite the fact Terry Bradshaw got knocked out of the game throwing an unbelievable touchdown on the drive before. The fourth down failed and the Cowboys almost hit a Hail Mary to win. It pays to watch NFL Films Super Bowl recaps.
That was in good nature though. He was interested in why I wrote the book in the first place. He asked if I had always been a math wizard. Funny, I’m not a math wizard at all. I couldn’t do any of the data science that PFF has produced or develop win probability models, I’m just interested in them. They have shined so much light on football and given us so much more to analyze and discuss about the game that I wanted to write a book about it.
12:45 p.m.: Phil Perry, NBC Boston
Phil is a beat reporter covering the Patriots. He’s very knowledgeable when it comes to analytics so it was like talking to myself. His theory was that Bill Belichick had probably fallen behind when it came to analytics usage. I pointed out that Cris Collinsworth told me that Belichick was once the farthest ahead of the game, despite public comments otherwise. Yet some of his recent in-game stuff seemed to slip behind, as did decisions with players. Maybe we wouldn’t have noticed if Tom Brady was still around. Could that be why teams didn’t hire Bill and wanted someone more progressive? I’m not sold on that. I’d guess it’s because Bill wants to run the entire organization.
Phil asked if Kirk Cousins was a fit for the Patriots. If they have a plan to replace four offensive linemen and three receivers, maybe. They are in a similar spot as the Vikings with an owner who won’t want to rebuild but a team that really needs it. The only difference is that they can draft Jayden Daniels and the Vikings probably cannot without a huge trade package (and even then it might not be enough).
2 p.m.: Jared Stillman, 102.5 The Game in Nashville
Jared is annoyed at the Tennessee Titans because they are using buzz words surrounding analytics but he isn’t sure if there is substance behind it. A fair question. We are still kinda asking this about the Vikings’ GM too, right? That got into a discussion about how hard it can be to evaluate decisions in a vacuum and separate where the analytics were involved. Normally with a front office the numbers have a seat at the table but they aren’t making the call alone. There are scouts, personnel people and coaches.
Sometimes I’ve wondered if a team was run by one person and a laptop how they would perform over a decade versus front offices of 100 people. Certain folks are painted as being “analytical” like Kwesi Adofo-Mensah but he’s really supposed to be the leader of the organization and taking into consideration everyone’s voice. He might do something that has positive analytical feedback that had nothing to do with the numbers or he might go against the numbers because other front office people and coaches believe in it.
We went down a road about outliers. The Christian McCaffrey trade. Analytically terrible. But McCaffrey is different and the 49ers are different. They can do something crazy for their final piece because they are that good as a complete team. Analytical statements about fourth downs or running backs are broad. 99% of the time you wouldn’t trade for a RB because they can only make so much impact (I used Derrick Henry as an example — he was good this year and it didn’t matter). But 1% of the time it’s a good choice.
2:30 p.m.: Aniello Piro, Mile High Sports
Another host who gets analytics. He wanted to understand the different levels of analytics impacting the game, so I talked about the in-game elements, striving for efficiency in play calling and front office decisions. It has really permeated every play and every choice being made by NFL teams, even down to using a data-based approach to study their own scouting reports. It’s really about pairing it all together to be 1% better than your opponents.
He asked about George Paton, who is the GM of the Broncos and used to work for the Vikings. I didn’t know George well outside of his reputation for being a very highly respected member of the front office and key in building the 2017 team. I feel bad for George. He’d be known as a brilliant GM if the Broncos had gotten the Russ Wilson from 2020. Instead I’m not sure how long he will be there. Sometimes you can do everything right and it just goes wrong. That’s football.
3 p.m.: Grant Wisco, 96.7FM/1670AM The Zone
My final spot of the day is in Green Bay again. Grant asked why he sees Vikings fans on Twitter always arguing that if Kirk Cousins just had X, Y or Z that they would have been in the NFC Championship game.
“That’s all they have,” I said.
We talked about how the Vikings got the best version of Kirk that they could have imagined and it wasn’t enough, which speaks to how difficult it is to build around an expensive QB that isn’t among the best playmakers. If Kirk’s 2019 self with that team wasn’t enough, why would 2024 or 2025 be different?
3:30 p.m.: Lunch and podcast time
I ate lunch and went back to my hotel to livestream on YouTube and write this. Wake up time tomorrow: 5:15 a.m. Yet somehow I think I’ll be the most well rested person in Vegas.
