This article is part of our East Coast Offense series.
Variance
I thought I had understood this concept, but I don't think I processed it completely until last week. I had taken variance to mean that outcomes vary due to randomness, and so you'll get some oddball results that aren't predictive. Funny bounces, blown coverages, drops, bad calls by the refs - the sort of thing that can swing a game - or a player's statline - in either direction. But while that's all true, I didn't fundamentally grasp how variance should be priced in properly on the front end.
By this I mean, when we see the preseason over/under for team win totals, for example, we'll give the best team 11 wins and the worst maybe 4.5. That means there are only 6.5 projected wins between them even though at season's end there will probably be a 13-3 team and a 3-13 one too. It's just that the 13-3 team will have had bounces go its way and the 3-13 the reverse. Or even if there is a true 13-3 team, it'll have been that team's 90th percentile outcome, tough to predict ahead of time. So the range of projected outcomes is fairly narrow. Take away a QB worth two wins, and that's one third of the way from the best team in the league to the very worst. For that reason even the best QBs in the league are worth roughly four wins at most. Four wins doesn't sound like that much - a 3-13 team going 7-9 doesn't change its fate. But it's four projected wins, so it takes the worst team (4.5) and moves it to within a couple games of the best. If a bounce or two goes its way, that means peak Aaron Rodgers could take the league's worst projected team to the playoffs - which is what we think of when we imagine a league MVP.
The implications of this dawned on me as I was handicapping last week's games. The reason getting a meager seven points is such a big deal in a league where 34-3 blowouts happen all the time is the range of projected outcomes between the best and worst teams is rarely more than 16 points and often less than that. But seeing so many blowouts fools us into thinking huge favorite X should easily cover a 14-point spread against Doormat Y because 14 points is not that remarkable of a margin. It isn't because of the variance, but it is incredibly large (essentially spans the entire range) of projected outcomes. If we eliminate variance as noise, then we see why getting four points at home or 10 on the road is so significant even though so often it doesn't affect whether the underdog covers.
This doesn't mean I'll always bet the underdog or start making my preseason player projections ultra conservative next year (though if I were entering a projecitons contest rather than trying to help people draft winning teams I would.) Variance in outcome above and beyond a "noiseless" projecton is a function of a game's volatility, and the NFL has an incredibly volatile scoring system with big plays and turnovers making such an outsized impact. While predicting how much a particular game will deviate from its projected outcome might be a fool's errand, I think we can look at situations where deviant outcomes happened in the past and develop a nose for them, so to speak. But even while doing that, grasping how narrow the range of projected outcomes is should give us a better sense of how much value is in a particular line or how many rounds a player has to fall in your draft before he's too good a bargain to pass up.
Survivor
When you pick every game against the spread and are in nine fantasy leagues, you win some, and you lose some every NFL Sunday. But there's no bigger win (or loss) than in Survivor. As a person who faded the Dolphins last week, it was about to get real when Cody Parkey lined up for a game-winning field goal. While the overall number of Dolphins backers closed around 43 percent on Officefootballpools.com, in my personal pools it was closer to 60. As some people had already gone out in Weeks 1 and 2 in my two smaller ones, I would have been down to the last handful of people in both, with $2500 and $3500, respectively, on the line. And I would have had good equity in the other (non-re-buy) bigger one. Unfortunately, he shanked it.
But I want to take a step back. Heading into Week 3 and having used the Seahawks I hated the survivor choices. The Dolphins were too highly owned for their moderate-favorite status, and I was nervous about the Cowboys, Packers and Panthers for different reasons. Like many, I was tempted to take the Dolphins as the "safe" play and live to fight another day. But the math I worked out said the Dolphins were a bad risk/reward play, and so I held my nose and put two entries on the Cowboys and two on the Packers. The point here isn't that it worked out (actually it worked out even if you took the Dolphins), but the idea the Dolphins were safe was simply false. Vegas had them as 81/19 favorites, and that's all we really had to go on. The idea that a third-string rookie QB on the road can't win, that the Dolphins played the Patriots and Seahawks tough on the road is entirely baked into 81/19. We cannot count those things as extra, e.g., let's make it 89/11 because the Dolphins are playing well, because we'd be double counting publically available information that is already accounted for in the line.
In any event, the near-loss (and yes, those people deserved to lose) was a good reminder: Trust in the odds and don't create justifications to avoid taking a sensible risk. The Packers and Cowboys were risky picks, but that risk was justified by the potential payout should the Dolphins have gone down. And if you lose with the best play, you should sleep well at night.
The Perils of Drafting Running Backs Early
In one of my Beat Chris Liss NFFC leagues, I took Odell Beckham with the second overall pick and then went three straight running backs who unbelievably fell to me at the end of Rounds 2 and 3 (third-round reversal) and early in 4. As a result, I took wideouts Emmanuel Sanders in Round 5 and Kevin White in Round 6. Sanders had done nothing the first two weeks, and the Trevor Siemian-led Broncos were in Cincinnati, a team ostensibly with a tough pass defense (it held the Jets wideouts in check and shut down Antonio Brown.)
Shortly before lineup lock, I put the following out on Twitter:
Kevin White or Manny Sanders, full PPR? I know everyone's gonna say Sanders, but I want to start White.
— Christopher Liss (@Chris_Liss) September 25, 2016
Most of the responses were of the "Go for it!" variety. "Trust your gut", "Play for upside", etc. I was going to do it anyway, but this confirmed for me I wasn't being crazy. So I sat Sanders and started White. (I also started Michael Thomas who I drafted a couple rounds later given the Monday night matchup and Willie Snead's absence.)
You know the results. Sitting Sanders for White cost me more than 20 points (though not the win.) But it also illustrates a key, hidden cost of taking RBs over WR in the early rounds: You're more likely to leave points on your bench. The problem with receivers is they're volatile week to week, as they're not guaranteed starting RB touches, and coverages and matchups change so drastically each game. Consequently, even top receivers like Brandin Cooks, Allen Robinson or A.J. Green will have quiet games. But because you will never bench them, you won't miss out on their huge ones.
As you get deeper into the draft, you wind up with receivers with just as much volatility but much less guarantee of finishing the year in the top-30. Maybe this Broncos team wasn't going to support two top-30 WR this year. And Sanders had only six TDs on 137 targets last year. It's hard to wait out the slumps when the matchup is bad, and you're not positive there's any payoff coming over the long haul.
Had I drafted receivers early instead, I would be counting on Charles Sims as my RB, who is now guaranteed touches with Doug Martin out. With running backs, the situation is usually clearer, though I also have Jerick McKinnon on that team, so there are some exceptions. But bottom line, you have to do more than simply compare value over replacement when choosing between a back and a WR - you need to consider how many points you'll actually get in your active lineup.
Week 3 Observations
• Ben McAdoo gave his trio of top receivers a combined total of 21 targets. On those 21 targets, they gained 264 yards (12.6 YPT). He also gave 17 targets to his tight ends and running backs, two of which resulted in interceptions, including the one that ended the game, that amassed 86 yards (5.1 YPA). McAdoo plainly adheres to his West Coast offense, dink-and-dunk philosophy, irrespective of personnel, irrespective of opponent and obviously irrespective of the success or failure of particular plays during a given game. Only game flow – and by game flow, I mean when he's absolutely forced to take chances down the field – causes McAdoo to deviate, and what happens? His team finally makes big plays to its wide receivers with ease. McAdoo simply does not play to the strength of his team because he's too wedded to his offensive philosophy.
• McAdoo also failed to use his timeouts on defense in front of the two minute warning and was only bailed out when Jay Gruden elected to run on third down rather than try to pick up the first down and likely end the game.
• One of the frustrating things about the Giants in addition to their high-school-level coach bringing down a playoff-caliber roster is their beat writers either don't know how bad McAdoo is or for careerist reasons aren't willing to call him out. Check out theseaccounts of the game and see whether there's any mention of the play selection or awful clock management. Unfortunately media incentives have evolved over the last few decades. Desire for access has turned many beat writers more or less into public relations operatives rather than adversarial representatives of fans who want a winning product.
• To his credit, SNY beat writer Ralph Vacchiano responded to my criticism and engaged it (to an extent) on Twitter. You can see some of the thread here.
•Eli Manning was sharp all game until the final interception. On the first pick, Will Tye appeared to stop his route.
•Odell Beckham mostly had his way with Josh Norman, but McAdoo doesn't set out to get Beckham involved until it's absolutely necessary.
• With Shane Vereen now out for two months at least, Orleans Darkwa ("The Dark Horse") who ran well and caught a pass for nine yards, could be the team's early-down back if Rashad Jennings misses more time. For PPR, Bobby Rainey might be the guy, though the Giants face the Vikings and Packers, two tough run defenses, next.
• The Redskins split their targets four ways between DeSean Jackson, Pierre Garcon, Jamison Crowder and Jordan Reed. Jackson is always going to win the per-target efficiency battle. He's the best big-play receiver since Randy Moss. Reed will get his red zone looks and TDs sooner or later.
• Buffalo-Arizona is why you take the home dog no one wants, even if you can't imagine it going your way. Someone must be betting the Bills (otherwise the line would have been much higher), and without Sammy Watkins against a Cardinals team that annihilated the Bucs the prior week, it's not the squares.
• That David Johnson went off in a game with the worst possible game flow is why he was the easy No. 1 RB on the board this year.
• I didn't watch much of this game, but is Carson Palmer okay? I know he assuaged some concerns last week, but Peyton Manning looked okay on occasion last year before it was clear he was done.
•Tyrod Taylor salvaged his day with 76 rushing yards and a TD, but he's not useable without Watkins. LeSean McCoy seems healthy and useful at least.
•Trevor Siemian threw for 312 yards (8.9 YPA), four TDs and no picks in Cincinnati against what's ostensibly a good pass defense. He's already much better than 2015 Peyton Manning, and the Broncos are quite obviously a viable Super Bowl contender again.
• With Siemian looking like an NFL QB, Sanders and Demaryius Thomas are both back in play as reliable receivers – 20 of Siemian's 35 targets went to them.
•Jeremy Hill had a 50-yard TD run and scored another short one, but take away the long TD, and he was 16 carries for 47 yards. (Not that the long ones don't count, but he was not consistently finding room.)
• I had the Packers with two of my survivor entries and the Lions plus 7.5. This game could not have gone more perfectly with the outcome never in doubt even with the backdoor cover.
•Aaron Rodgers and the Packers offense got well at home with Rodgers at 8.5 YPA, four TDs and no picks. Jordy Nelson finally looked like himself too with a great TD catch. Randall Cobb not so much. Of course, Rodgers attempted only 24 passes.
•Eddie Lacy's 17 for 103 was a nice sign, though he caught only one pass. Part of that was due to game flow. The good news is James Starks had only one carry.
•Matthew Stafford put up big numbers, largely in garbage time and mostly to Marvin Jones who had 200 yards and two TDs. Eric Ebron looks like the No. 2 option (he should get more TDs, as one was called back in Week 2.) Golden Tate has been a bit player so far.
• The Lions running game was bad, particularly Theo Riddick (10 carries for nine yards.)
•Derek Carr had another modest game (7.1 YPA, one TD, one INT). Michael Crabtree saw more targets than Amari Cooper again.
•DeAndre Washington is better than Latavius Murray – that should become a more even timeshare soon.
• The Titans don't throw down the field enough to have a credible offense. DeMarco Murray has been reliably good, but Tajae Sharp is not a No. 1 WR right now.
• Apparently Jack Del Rio nearly gifted this game to the Titans after punting on 4th-and-short from plus territory near the end of the game and enabling a final drive that would have tied the game but for a penalty. Del Rio had the guts to go for two against the Saints, but as is usually the case with these troglodytes, their good decisions are mostly dumb luck, as they don't have a grasp of basic probability theory.
• I have no words for Cody Parkey keeping the undeserving cowards who took the Dolphins alive in survivor. He actually missed three field goals that game, none of which were from 50-plus.
•Terrelle Pryor completed 3-of-5 passes for 35 yards, ran four times for 21 yards and a TD and caught 8-of-14 targets for 144 yards. That's essentially 23 "attempts" for a player who qualifies at wide receiver. His biggest problem was when he played quarterback, the team's best receiver was unavailable. That will change when Josh Gordon comes back, though Josh McCown might be back shortly thereafter.
•Duke Johnson got more involved in the running game, but roles are still in flux in the offense.
•Kenyan Drake and Jay Ajayi played roughly to a draw, with Ajayi being lucky enough to get the overtime TD run. Expect a 50/50 split until further notice or Arian Foster returns.
•Jarvis Landry will get his targets no matter what, and with Adam Gase, further down the field than in the past. Ryan Tannehill's fantasy numbers were good, but he threw two picks, lost a fumble and should have lost the game to arguably the league's worst team.
• Avoid Denver's defense, but really avoid Minnesota's. Three picks and eight sacks, including a safety, of Cam Newton in his own building. For fantasy purposes, they added a punt return TD for good measure.
• Bizarre that Kelvin Benjamin after two dominant games saw only one target which he failed to bring in.
•Cameron Artis-Payne saw more work as a runner, Fozzy Whitaker as a receiver, and both were predictably useless. Jerick McKinnon got more carries, Matt Asiata and McKinnon one catch each, and like their Carolina counterparts, were both useless.
• It made little sense why the Vikings would give up a 1st and a 4th for Sam Bradford, but maybe they knew something we didn't. If you're a Super Bowl contender, having a competent rather than terrible QB is worth nearly any price. Incidentally, since the start of last year, the Vikings have been a spread-covering machine (16-3 ATS).
•Justin Tucker kicking a 54-yarder is like Steph Curry draining an open three from four feet beyond the arc. It's much easier than it should be.
• Every week a different player leads the Ravens in targets and receiving yards. This time it was grandpa's turn (was going to put in a link to the Texas Chainsaw Massacre, but couldn't find a good one.)
•Terrance West ran competently, but rookie Kenneth Dixon should still get a shot when he returns. Justin Forsett caught six passes, albeit for only 12 yards.
•Allen Robinson finally got his touchdowns, but 11 targets for 57 yards shows how badly the offense is still struggling. Blake Bortles was abysmal, and with a minute left and no timeouts, he was scrambling and buying time, but instead of getting rid of the ball, he took what amounted to a game-sealing sack. And Chris Ivory was of no help to the running game which was again non-existent.
•Marqise Lee saw another seven targets – he's a regular part of the offense now.
•Christine Michael is a top-10 back now that Thomas Rawls' is out for 5-6 weeks. Michael could have had a bigger day but for being pulled in garbage time.
•Russell Wilson looked sharp before exiting with an MCL injury. He says he could have returned to the game had it not been a blowout, but we'll see what the MRI says.
•Doug Baldwin had another big game and is clearly the team's go-to receiver this year. Should he and Wilson stay healthy, he'll shatter his career-high in targets from last year.
•Jimmy Graham had a big game, proving to be an unlikely survivor of a patellar tendon tear (Victor Cruz is also doing well, but another year removed.) If Graham is anything close to his pre-injury self, and Wilson's injury proves to be minor, the Seahawks might be the best-rounded team in the NFL.
•Carlos Hyde did some damage in garbage time. There's nothing else to say about the 49ers offense.
•Jameis Winston had 405 yards, but it took him 58 throws to get there, i.e., only 7.1 YPA. The Bucs are short of weapons. After Mike Evans, slot man Adam Humphries saw a lot of work, but with Vincent Jackson looking old and irrelevant, and Austin Seferian Jenkins gone, there's no legitimate No. 2 weapon.
•Charles Sims played decently in his debut as the starting running back with six catches, 124 yards from scrimmage and a score.
•Cameron Brate caught two TDs but is not a threat to make plays down the field, i.e., he's mostly red-zone only.
•Todd Gurley slugged his way to two scores, but it was tough going. Case Keenum made a couple long TD throws, but the offense has no rhythm.
• I bitch about bad beats quite a bit, so it's nice to see a lucky one bounce my way – the Colts had no business covering against the Chargers, but a long TD to T.Y. Hilton when Indy only needed a FG sealed it.
• It was hilarious how bad the announcers were in this game. After Indy went up four on the TD and went for two (and failed) to make it six, they kept harping on how big San Diego's earlier missed PAT was, thinking the Colts would only be up three for the final drive. But had the Colts been up three, they would not have gone for two and failed! They would have kicked a PAT to go up four anyway.
•T.Y. Hilton had his way with all-world cover corner Jason Verrett. Matched up one-on-one, Hilton spun him around a few times and caught a nice back shoulder throw from Andrew Luck. He finished with eight catches, 174 yards and a TD.
• Incidentally, I watched only part of this game, and while Luck's numbers were padded by the late Hilton score, he looked sharp to me. Luck is a polarizing player – the scouting types love him as does the mainstream media, but the advanced stats people think he's overrated. I'm with the squares here – put him with a better coach, in a better system, and he'd be one of the league's undisputed superstars.
•Melvin Gordon got another score, but 2.2 YPC against the Colts? And the Chargers were unable to put the game away behind Gordon. He did have four catches for 43 yards, but 25 of those were on a play where he ran up the sideline untouched. Again, I'm selling high even though he's the only game in town.
• For whatever reason, the Chargers didn't target Tyrell Williams at all in the first half, but he finished with six catches for 69 yards and led the team with nine targets. He's the team's best receiver.
•Hunter Henry had five catches for 75 yards, but lost the fumble that sealed the game.
• The Chiefs didn't have to do much with the opposing quarterback throwing six interceptions. Ryan Fitzpatrick got 4.3 YPA, to boot, and Eric Decker and Brandon Marshall combined for four catches.
•Spencer Ware averaged only 3.9 YPC, but that was heavily weighed down by what were essentially kneel-downs as the Chiefs looked to kill the clock with a huge lead.
• I was skeptical of Carson Wentz last week, but no more. His scramble to buy time and perfect throw on the move to a running Darren Sproles turned into a 73-yard score that broke the game open. He also attempted a deep throw to Dorial Green-Beckham up 34-3, a great sign for fantasy purposes. Wentz finished with 301 yards, 9.7 YPA and still hasn't thrown a pick this year.
• Perhaps more impressive was the Eagles defense that held the Steelers to three points and Ben Roethlisberger to 5.8 YPA. DeAngelo Williams was also bottled up for the first time this year.
• Of course, Antonio Brown still got his – 18 targets, 12 catches, 140 yards.
• The Eagles vowed to get Green-Beckham more involved, and while he got only four targets, no receiver saw much work given the game flow. Pick him up now as he has the chance to be a league-winner for you if he pops.
•Ryan Mathews saw only two carries for negative yardage, and it's unclear whether he's injured. In any event, Wendell Smallwood, and Kenjon Barner ran well, each scoring a TD.
• It's a small sample, but I wonder whether Doug Pederson (and possibly Hue Jackson) aren't two of the better under-the-radar NFL coaches. As a Giants fan, I'd give anything for either one rather than the sub-moron we have.
• It was surprising to see Massey-Peabody have the Bears as their "Big Play." I realize they're going strictly by the numbers and not anything personnel related, but that line actually seemed low for a home game against a bad team down to its backup QB. Full disclosure, I took the Bears as a "the sharps must be on them because no square in his right mind would take this awful team," but I had it as a coin flip, and I didn't feel good about it. The game went exactly as you'd expect with Dallas running roughshod over the defense and throwing from hitter's counts on third down. Anyway, I think M-P is a valuable resource and respect their work quite a bit, but of all the games to have as a Big Play, that was a strange one. But I suppose it's just what their indicators (which are strong over time) told them. It's also easy to point things like this out after the fact, but had the Bears covered, and the game flow been surprisingly Chicago-favorable, I would have written it up t
•Ezekiel Elliott had 140 yards on the ground, but Jason Garrett trolled Elliott's owners with rushing TDs from Dak Prescott, Alfred Morris and Lance Dunbar.
•Dez Bryant salvaged a potentially terrible day with a late TD. Still, as Antonio Brown is to Odell Beckham, Beckham is to Bryant, i.e., as comparatively neglected as Beckham is, Bryant's lack of use (six targets) is even worse. And now Bryant is awaiting MRI results on his right knee.
•Cole Beasley is Prescott's favorite target, it seems.
•Zach Miller is Brian Hoyer's favorite red-zone target, and that resulted in two TDs. The Bears receiving tree is thin.
•Kevin White made a ridiculous one-handed catch, but he and the Bears offense still look very raw. Still he saw 14 targets, by far the most on the team.
•Jordan Howard ran well. He has good burst, but not much wiggle. That should still be enough to supplant Jeremy Langford, who left the game with an ankle injury.
• Streaming the games from Berlin via my laptop and Apple TV make following them on Twitter an annoying experience. Like watching the Giants on their final drive near midfield, and 100 people saying "Oh Eli!" before I've seen what happened.
•Drew Brees always gets his at home. He'll have tough tests against Denver and Seattle, but no QB has more pronounced home/road splits.
•Coby Fleener had a big night. It's worth noting it took Ben Watson a few games to get going last year too. Fleener's arguably a top-five TE at this point.
•Devonta Freeman looked great, but unfortunately he gave up goal-line carries to Tevin Coleman who wasn't shabby, either. I advised a random caller on a radio show who had a tough call to use both backs, and I feel good about myself. (You might see this as self back-patting, but believe me I'm so often aghast at the advice I've given when it goes wrong I feel entitled to it.) Going forward, this is a tough race to handicap as Freeman probably solidified his role as the starter with 207 YFS, but losing goal-line work would hurt quite a bit.
• How is it possible Julio Jones got seven targets but only one catch against the league's worst secondary? He must be playing hurt. If you drafted him in the top three, hoping for another 200 targets, you should be worried.
• Cooks had eight targets and only two catches, but it's less alarming given Desmond Truffant's coverage and Drew Brees' tendency to spread the ball around. But Coby Fleener's and Michael Thomas' emergence could be a problem once Willie Snead gets back. Cooks will get his 130 targets minimum, but he needs to crack 150 to be top-seven fantasy receiver.
•Mark Ingram finally scored, albeit on a pass play. He should be fine, but that defense is killing his game flow, and Travaris Cadet caught six balls to Ingram's four.
For the podcast version of this article, click here
