East Coast Offense: Betting Against Gravity

This article is part of our East Coast Offense series.

Betting Against Gravity

The Dolphins not only Survived this week, but they also covered the 6.5-point spread (It was four at kickoff, due to the quarterback changes for both teams, but I locked it in Wednesday at 6.5.) That put me at 9-0 for the season on best bets, a 1 in 512 proposition by dumb luck. 

Now 1 in 512 is not that crazy -- people win the lottery, get hit by lightning, get royal flushes at the poker table, all at much steeper odds. So there's a long way to go before I can say it's definitely skill. But this framing is already a loser, seeing one's efforts and work through that lens. Every person who succeeds at every endeavor defies the odds. You want to open a restaurant? Odds are it'll be out of business in three years. And yet how many great restaurants have you been to in your life? How fortunate that restaurant owners are not all base-rate probabilists who are too savvy even to try! You want to be an actor? Terrible idea, you'll be waiting tables your entire life, hoping in vain for that big break. And yet there are great and successful actors in every generation. What about Tom Brady, slow, unathletic sixth-rounder winning seven Super Bowls? Just south of mathematically impossible. 

I started picking games against the spread for RotoWire with my brother in 1999. We went 278-202 over the first two years (58%) over a 480-game sample. This wasn't simply on best bets, or even a Super-Contest-style best-five per week, but having to pick even games we didn't want to touch. Dumb luck? You can figure out the odds -- still better than getting a royal flush -- but we had a process. That process was to go over the lines by phone on Tuesdays and revisit them several times during the week, arguing and discussing back and forth until we got clarity after which I wrote up the column. I detailed this process in a magazine article I wrote in 2001. 

Around that time, a few things changed. My brother became busy with work and eventually family, so I was left to do it solo. I didn't have the back and forth, and for that reason -- or maybe just regression some other -- I didn't do quite as well. (Notice I crossed out "regression" as the explanation as that begs the question, i.e., it implies the success was dumb luck, which is precisely what's under discussion.) 

Once I started struggling, I tried to improve by adopting various heuristics, whether they were stat-based, or market-based (fade the public) with only limited success. I started viewing handicapping against the spread as more of a probabilistic endeavor, and of course, I was doomed to perform as someone flipping coins would -- right around .500 over the second decade of doing it. I might not have been picking well, but at least I was not so naive as to believe it were possible to pick well against the spread over time!  And even though I had detailed plainly the process that had worked so well from 1999-2000, I had lost an element of it (my brother and our calls), so I went away from it. And after converting to the secular religion of my probabilist peers, my results fell in line with what they would expect. 

But probability is a more nuanced subject that it seems. While it might be hard to perfect a process for coin flips that beats the 50/50 average over time, and impossible to perfect one for virtual coin flips wherein every coin is perfect and every flipper without skill, the analogy to NFL point spreads doesn't entirely hold. While the market is good at setting lines, and therefore coin flips are a marginally useful way to model it, there is a world of difference between something we know to be 50/50 every single time, and something that's only roughly 50/50 in the aggregate. 

Put differently, There is no individual NFL game for which the probability of the favorite covering is exactly 50 percent. We don't have perfect information about the game the way we do for a virtual coin, and in retrospect, some of the lines turn out to be absurd (take the 2008 Super Bowl where the Patriots were 14-point favorites over the Giants, for example.) To see every game as a coin flip is to mistake a crude model for a complex reality. And yet, it's so tempting because it is hard to beat the market consistently, and being savvy enough to know "the game" of handicapping is a much more comfortable spot than naively making picks and being unable to overcome the vig most years. But if you're not willing to be the naive dunce who trusts his own judgement, it's a self-fulfilling prophecy -- you will be exactly as mediocre as your peers, who know better than even to try. 

Of course, understanding, but setting aside the base-rate probability belief system isn't enough. You must have some creative process, some basis for exceeding it. And that's where observation and trust come in. You watch the games, you consider them, and you trust yourself to see what's happening more clearly than the market, at least in spots here or there. (It would be awfully hard to create your own lines and let the market bet into you, but you don't have to do that -- only to find a couple weaknesses in the existing ones.)  To that end, I've tried to recreate what I was doing in 1999-2000, by going over the lines Monday night, again on Tuesday and again on Wednesday. For the SuperContest and Survivor, I go over them yet again with Dalton Del Don on the phone and even go against some of my prior picks. 

This process does not guarantee success, obviously. In fact, while the Super Contest picks and best bets have been good, my overall record is still just under .500. Moreover, I am liable to fall into the same traps, errors and biases as anyone else. But to me the biggest trap is the probability one -- seeing the world (and the NFL games) as an extension of lifeless coin flips, settling for mediocrity and believing the only edge is capitalizing on the mistakes of others. I'd rather invert this logic and risk being the person on whose naive mistakes others can capitalize for the chance to do something worthwhile -- even if improbable.  

Perhaps this is folly, and regression is like gravity -- you don't want to bet against gravity, do you? But how else are we going to the moon?

Week 10 Sporcle Quiz

Can you name all the players with 200 rushing yards over the last four weeks?

Guessing The Lines

At first glance, I'm on the Dolphins (again), Jets, Colts, Broncos and Raiders. Of course, I reserve the right to change my mind in Beating The Book. 

Week 9 Observations

• I've made fun of Justin Fields this year because the Giants could have drafted him and instead wisely (in my opinion) traded down, but he passed the eye test for me. There's no doubt about his scrambling ability, but he seemed unusually poised for a rookie on the road against a nasty defense and got 10.0 YPA against them. Chicago is a tough place to play quarterback, but Fields has a chance to make it.
• David Montgomery (13-63-0, 2-2-17) was back, and he was essentially the only back as Khalil Herbert (4-13-0) didn't see the field much.
• Cole Kmet (8-6-87-0) and Allen Robinson (6-4-68) were involved at least. Darnell Mooney (1-15-1, 6-3-41-1) scored twice, but had only 56 total yards.
• Najeh Harris (22-62-1, 3-3-16-0) always gets his. He didn't have much room to run, but showed good patience and didn't go down easily. Oddly, the player he most reminds me of is Le'Veon Bell.
• Diontae Johnson (6-5-56) got me just enough to win my NFFC primetime matchup. Well, that and Khalil Herbert's 1.3 points.
• The late taunting penalty that gifted the Steelers a huge first down on their penultimate drive mercifully did not decide the game, but that was as egregious as it gets. Borderline Tim Donaghy s***.
• The announcing is this game was intolerable, and keep in mind I watched only the 40-minute version.
• The Titans held the Chiefs to three points two weeks ago and kept the Rams out of the end zone until the final minute of garbage time. Their defense is somehow good now, and after losing to the Jets, they've reeled off wins against the Bills, Chiefs, Colts and Rams. Mike Vrabel has to be on the short list for Coach of the Year.
• Cooper Kupp (13-11-95-0) gets his no matter how the game goes.
• The Cardinals destroyed the 49ers defense without Kyler Murray and DeAndre Hopkins.
• James Conner (21-96-2, 5-5-77-1) had a monster game after Chase Edmonds left early. Conner now has 11 touchdowns over his last seven games.
• Jimmy Garoppolo is so easy to sack – he has little mobility or strength. It boggles the mind the 49ers traded up for the No. 3 pick, yet still keep Trey Lance on the bench behind this guy.
• George Kittle (8-6-101-1), Brandon Aiyuk (8-6-89-1) and Deebo Samuel (9-5-63-0) split the work as we might have expected before the year. Aiyuk lost a fumble, but otherwise looked great, snatching everything near him with his giant wingspan.
• The 49ers defense could not make a tackle or get a stop to save its life.
• Jordan Love got the cover, and that was about it. I doubt he'll be starting for the Packers in 2022.
• The Packers defense is good this year, but the Chiefs offense is a shell of its former self. (Yesterday when I was shucking oysters at a friend's house, and we gave a shell to my dog, I said, "Here you go, Oscar, but this oyster is only a shell of its former self."
• Justin Herbert reminds me of Dan Marino, only more mobile. The Chargers for some reason have reverted to dink and dunking to Keenan Allen (13-12-104-0) rather than downfield shots to Mike Williams (5-2-58-0.)
• I started Jalen Reagor in a 14-team half-point PPR because I had DK Metcalf on bye and DeAndre Hopkins out. He got me negative one tenth. Real receiver.
• Jordan Howard (17-71-1) was the guy, but keep in mind that was against the Chargers who are one of the easiest teams against which to run. Jalen Hurts completed only 11 forward passes.
• The Vikings can lose a close game against anyone.
• Lamar Jackson had 21 rushing attempts and 41 passing attempts.
• Le'Veon Bell is getting more involved – if Adrian Peterson at 36 is still scoring TDs, why can't 29-year old Bell?
• I love the Josh Allen on Josh Allen violence as much as anyone, but Scott Hanson beat it to death so badly, I can't muster the enthusiasm for more than this sentence.
• I'm starting to feel badly for Sam Darnold and even the stats guys who insist the Giants should have taken him over Saquon Barkley. Stop the fight already!
• As I said last week, I'm starting to get a 2001 vibe from the Patriots.
• I made this point in last week's podcast, but maybe the all-time greats aren't necessarily better at game-planning or strategizing, but instead error-correct faster. A great quarterback who is struggling against a defense more quickly diagnoses the best way to attack it. Bill Belichick's teams almost always improve as the season goes on, and this unusually efficient error-correction might explain why his teams went 58 percent against the spread from 2001-2019. The market looks only at past-performance and perhaps cannot keep up with Belichick's unusual rate of improvement. In any event, by Wednesday night, I realized I should have made the Pats minus four my best bet.
• Rhamondre Stevenson (10-62-0, 2-2-44-0) looks like a player.
• Josh Jacobs (13-76-0, 4-4-19-0) looked awfully spry against the Giants, but he's so rarely at full health.
• Devontae Booker (21-99-0, 3-3-23-0) is a quality backup RB. He breaks tackles and makes good decisions, but lacks an extra gear.
• The Saints' QB play devolved this week. I'd expect Taysom Hill to start in Week 10.
• The Broncos were up 30-0 in Dallas before the Cowboys' two garbage-time TDs in the last 4.5 minutes.
• Baker Mayfield really prefers life without Odell Beckham.
• Nick Chubb (14-137-2, 2-2-26-0) is still the top-five real-life back we thought he was, but Kareem Hunt could be back before long to ruin it.
• Remember when the Bengals were the top seed in the AFC a couple weeks ago, and Ja'Marr Chase (13-6-49-0) was averaging 14.8 YPT? He's down to 11 YPT, though still leading the league by a full yard. The Bengals are now one of five 5-4 teams tied for the 7th-11th spots in the conference.
• I'm pretty sure no one on the planet was watching Dolphins-Texans as closely as Dalton Del Don and me.
• The Dolphins turned it over three times in the first half, lost their starting center, and of course, their quarterback. But Tyrod Taylor gifted them a TD late in the second quarter by throwing an accidental pick near the sideline when he meant to throw the ball away.
• The Dolphins, up eight and in range, could have sealed the game with a field goal midway through the fourth quarter, but instead of running the ball, chewing clock and locking it up, they called pass plays, and Jacoby Brissett was strip-sacked. Think Kyle Shahanan in the Falcons-Patriots Super Bowl.
• It just goes to show any game, no matter how bad its teams, can be epic if you have enough riding on it.
• Carson Wentz played well, made good throws, didn't turn it over and scrambled when necessary. More importantly he looks completely healthy.
• Jonathan Taylor (19-172-2, 2-2-28-0) is the 1.1 now. It's amazing how often he's gotten stuffed at the goal line, though, but he'll keep getting his chances, especially against all the bad teams on the schedule.
• Nyheim Hines (6-74-1, 6-4-34-0) also had a good game. Marlon Mack was a healthy scratch, so this is now a two-man backfield with Hines likely seeing more targets.
• Mike White was off to a good start before getting hurt, but Josh Johnson generated a lot of successful drives. Both looked much more ready than Zach Wilson.
• Michael Carter (13-49-0, 2-1-37-0) is the lead back with Ty Johnson (4-21-0, 3-2-40-1) as the sole backup.
• Elijah Moore (8-7-84-2) is living up the preseason hype now that Wilson is out. I'm still annoyed I didn't have the guts to use him.
• The Bobby Okereke interception at the end of the game not only saved the cover for me, but gave me five points in my lone IDP league.
• The Jets really have a problem on their hands with Wilson. Management needs to play him when he's healthy to see what it has, but the team has shown a lot of fight the last two weeks, and I can't imagine benching White (whose forearm injury is minor) or Johnson for Wilson will be popular with the players on the team unless Wilson improves in a hurry.