This article is part of our East Coast Offense series.
Objective Models vs. Subjective Observations
I've been thinking about this for a while but have had a hard time articulating it. When we analyze something as complex as the NFL, it's useful to have a framework, grounded in historical performance, through which to understand it. At the same time, any framework - or model - has two weaknesses: (1) It can only say what's generally the case over the long haul, not specifically the case in a particular instance; and (2) it's likely to be slow to adjust when the underlying conditions change.
Let's take a simple example: Average Draft Position (ADP). There are some who believe, quite plausibly, that ADP is a better predictor of year-end performance than any one expert's personal rankings. The wisdom of the marketplace often trumps the bias-ridden observations of the individual. So it makes some sense to draft according to ADP, grabbing any player who falls significantly below his market price and never reaching for a player equal to or above it.
But while this method of drafting might work better than average generally, that doesn't mean it'll be correct for your second-round pick on August 23rd. In fact, it might be the case you thought a slimmed-down Le'Veon Bell moved especially well in the preseason, but you passed on him in the middle of the second round because his ADP was lower than that after his marijuana arrest.
While no model is perfect, and many times your hunch based on how great a player looks in preseason - Brandin Cooks, anyone? - will be wrong, and you'll wish you hadn't reached above ADP for him, the question is one of optimization: Is ADP value-taking the best we can possibly do, or are we even better when we incorporate our observations? If it's the latter, how can we distinguish helpful observations (Bell) from unhelpful ones (Cooks) in advance?
Another good example is the Massey Peabody model for picking games against the spread (ATS). This is one of the best ATS models I've seen not only because it's done well, but also because it's relatively transparent - they explain their basis for evaluating teams, and they post and publish much of their data and all of their results for free. But when the MP model identifies a particular bettable game, i.e., its line diverges significantly from the Vegas one, it's not exactly correct to say the model "likes" a certain team in that matchup. In fact, I'd argue it has no opinion whatsoever about that particular matchup, i.e., who's going against whom, but is instead making the claim that matchups of this sort based on the teams' relevant performance histories should be expected to yield a range of results that makes betting one of them a good investment at Vegas' current price.
If the pick turns out to be wrong, it could be random variance, e.g., the team it picked lost three fumbles and had a punt blocked, or it could be because the model - which works generally - does poorly in situations where one of the teams has an unusual characteristic (maybe by design it generates or risks giving up an unusual number of big plays). Or it could be wrong because in the case of these particular teams, the relevant interval of past performance diverges sharply from the average, e.g., usually the entire season is relevant with more weight put on recent games, but in a particular team's case maybe only its last two games should be given any weight whatsoever (we saw this with the Giants in 2007 and 2011 and the Ravens in 2012 during their playoff runs.) This latter error is an instance where the model might have trouble adjusting as quickly as an observer would.
Again, it's a question of optimization. Are you better off taking what the model can give you, or can you improve on the results by incorporating your own observations even if they diverge?
Finally, models typically are based on a certain set of premises or underlying conditions. Given the way the game is played today, certain team metrics are more highly correlated with winning and losing than others. Should that change due to new rules, new strategies or adaptations to either, the model will not work as well and would need to adjust. In fact, I think we're seeing that with so many coaches making fourth-down decisions using an out-of-date, old-school model from the '70s when it was harder to drive down the field in a more run-heavy league. The "always punt on fourth down" credo we mock comes from a framework that hasn't kept up with the modern-day incarnation of the sport. (I actually wonder whether that extreme risk aversion was even optimal back then.)
Even beyond football, this polarity between the general and the specific, routine and improvisation, rules and exceptions, large-sample observations vs. individual anecdotes, is often where the rubber meets the road in the decision-making process. While I'd argue you should pick a good historically useful framework as a starting point, you should also maintain constant suspicion about its long-term effectiveness and trust your own observations. When it comes to distinguishing whether that will cause you to reach for Bell or Cook, I have a simple rule of thumb: If it's something I saw with my own eyes or that struck me personally as significant, I trust it. If it's something I happen to believe because I read it somewhere, or someone else (no matter how reputable) made a good argument for it, I'll try to ignore it. The trick then is tracing the origin of your beliefs, and even that isn't always easy.
Peyton Manning vs. Tom Brady - Who's the Better Interview?
After Manning broke Brett Favre's TD-pass record Sunday night, he was interviewed after the game and said all the right things about the quarterbacks who came before him, his teammates, etc., etc. And it wasn't the dead-eyed "I'm reading a script" kind of interview, but one with the perfect tenor and tone. I've written before about how Manning's not only mastered the on-the-field game, but the off-field one as well. While there's nothing necessarily wrong with being an exceptionally polished spokesman for oneself, I have to disagree with Bob Costas who said something to the effect of "Manning isn't only a great player, but a great interview as well."
He's a great interviewer if by that you were describing someone likely to land a sought-after job, but unless I needed an instructional video for that purpose, I'm not interested in it. I'm not implying he's lying or doesn't really think all those things - only in the same way Upton Sinclair noted: "It is difficult to get a man to understand something, when his salary depends upon his not understanding it," it's easy to see how someone could believe what he's supposed to when his salary (and popularity) depend upon it. Manning is the perfect corporate advertisement for the NFL and is unsurprisingly featured in so many of them. Good for him - he's his own best spokesperson, in stark contrast to this poor bastard.
On the other hand, you have Tom Brady, who after the Patriots win Thursday night was boring as hell, repeating cliches and saying off the field it's all about family. But then he added:
I'm always sort of frustrated. I've been playing for 15 years, and I'm sort of pissed off about everything.
Brady has none of Manning's polish, sounds dumber than he is during the interview, offers none of the obligatory self-deprecation Manning does and yet admits he's pissed off about everything despite being a multi-millionaire, two-time MVP, three-time champion with an all-time supermodel for a wife. He admits he has it pretty good when put in perspective, but for a moment we got a glimpse of who Brady actually is - a driven perfectionist not especially concerned with his "brand" who's frustrated the Patriots haven't won a Super Bowl since 2004.
He's also the same guy who in 2009 when the Jets were on Hard Knocks, told WEEI:
You know, I haven't turned it on. I hate the Jets. I refuse to support that show... I'm sure it's great TV, and I'm glad people are liking it. But that's something I have no interest in watching. I'd love to say a lot of mean things, too, but I'd rather not do that, either.
Again, unless I'm looking for a course in social and corporate self-advancement, give me Tom Brady over Peyton Manning the interviewer any day.
Tipping the King
With coaches making so many cowardly punting decisions even in blowouts, I think the NFL should add a rule that would preserve player health and spare viewers and fantasy owners wasted time: the chance to resign mid game.
Here's how it would work. Instead of punting on 4th-and-7 from your own 30, down 35-7 in the third quarter (which is tantamount to resigning anyway), the coach could actually resign the game to the opponent, get his players off the field and get ready for the next game.
You could have the coach toss in a white flag instead of a red challenge one, but I'd actually opt for a 10-foot chess-piece king on the sideline he could tip over if the game got too out of hand. Why go through the motions of snaps, handoffs and penalties while risking injuries? Only beginner chess players make each other go through the formality of the checkmate when the outcome is clear.
Week 7 Observations
• So much for my Orioles World Series prophecy. When they won the AL East and swept the Tigers, I started taking it a little too seriously. I can now out the man whom I let persuade me not to bet them, too. It was @Jonahkeri. (This was before I bet them anyway at much lower odds in September.)
•Antonio Brown is a hellish player to be rooting against with a lead on Monday night. He didn't have a huge game, but did just enough - thanks to his TD pass - to beat me. For a receiver he's just so damn involved in the offense.
•Arian Foster and Le'Veon Bell are both firmly in the top tier of running backs. Only DeMarco Murray, Matt Forte, Jamaal Charles and LeSean McCoy are close.
• Marcus Wheaton started off strong this year before tapering off, but the 6-4, 211 Martavis Bryant makes more sense as Brown's complement.
• On the game's final onside kick, Darrius Heyward-Bey was on the Steelers "hands" team. Apparently, they couldn't find a defensive tackle to do the job instead.
• I get why breaking the all-time TD record is a big deal, but I'm against acknowledging it during the game and having a late-second quarter celebration. Records are something derivative you deal with later, not during the action. Plus, that Manning would get it either this game or next was never in doubt. Celebrating (and calling plays specifically to get) records in game is similar to those ridiculous Derek Jeter, Cal Ripken and Mariano Rivera farewell tours. Have some respect for the particular game you're playing, tally achievements later.
•Ronnie Hillman is a better alternative to the passing game than Montee Ball because unlike Ball he doesn't give the defense a break.
•Demaryius Thomas is the No. 1 overall pick if I'm drafting today.
• The Niners have a lot of quality receiver depth with Michael Crabtree, Anquan Boldin, Stevie Johnson and Brandon Lloyd, but you never know which one will be involved in a given game. And that's not counting Vernon Davis who should eventually secure a catch.
• The Giants had their chances before Larry Donnell's first fumble, but the Cowboys were the better team Sunday because their scoring drives were easier, thanks largely to the blocking of their all-world offensive line. Tony Romo had so much time on a couple TD throws, they were inevitable.
• If Thomas is the No. 1 overall player Dez Bryant isn't far behind. DeMarco Murray's in that conversation, obviously, but he tweaked his ankle on the turf and had to get it taped before returning. While it didn't affect him much Sunday, I'd be surprised if he's not on the injury report this week.
•Odell Beckham scored two touchdowns, but Rueben Randle and Donnell saw more targets. In fact Donnell caught all seven of his, though he fumbled after catching two of them. While the quick-hitting, short-passing offense has been largely effective, the Giants need to take an occasional shot down the field. Otherwise, it's too easy to stall, something you can't afford when Romo's getting 12.1 YPA against you.
•Jason Pierre-Paul, who had two sacks and six solo stops, looks close to his 2011 level. Apparently Pro Football Focus agrees.
• I know he's had some disaster games, but Ryan Tannehill has a few good ones too. I still like his upside in the second half this year.
• The Dolphins defense held Jay Cutler to 5.6 YPA at home.
•Brandin Cooks is looking more like Tavon Austin by the week.
• What a terrible blow to those who waited out C.J. Spiller's slow start only to see him knocked out for the year after a 56-yard gain with starter Fred Jackson out. (Likely an unwelcome development for Spiller too.)
• So much for the Kirk Cousins era. Three games ago, people were saying they'd take him over RGIII. Now he's benched for Colt McCoy. Or maybe I'm overestimating the distance between those two poles.
• The Titans are the most pointless team in the NFL. Unless their purpose is to troll my ATS picks (whether or not I back them.)
• The Bengals offense without A.J. Green, Marvin Jones and Tyler Eifert is not a top-30 unit.
• Many predicted the Seahawks would stomp the Rams after the embarrassing home demolition by Dallas. But the Seahawks have a bad offensive line, and the defense can't generate a consistent pass rush. It took some trick plays by the Rams, though, including a heroic call by Jeff Fisher to fake a punt from his own end on 4th-and-3 to seal the game.
•Russell Wilson became the first player in NFL history to throw for 300 yards and rush for 100 in the same game.
• The Panthers +7 seemed like an easy call, but when they decide not to show, they don't even fake it. Jordy Nelson is a victim of the Packer blowouts where he catches one long bomb and is no longer needed.
• People have been emailing and tweeting me about my two-seconds of screen time in that Samsung ad. It's very exciting, like I achieved some important goal or won a major award. The truth is it's a clip from last year that barely anybody watched, i.e., I did absolutely nothing except give them permission to use it. Still, if people think I'm the man because I was an extra in a television commercial, who am I to dispute that?
