This article is part of our East Coast Offense series.The Biggest Edge
You can profit in fantasy football, sports betting and Survivor if you're able to predict the future better than your competitors, but that prospect is becoming increasingly difficult. For starters, your competition has better access to relevant information, and perhaps, more crucially, better methods of making decisions. Instead of drafting based on gut instinct, they're looking at ADP. Instead of looking at free-league ADP, they're using high-stakes NFFC ADP, sorted by the past week. Instead of watching SportsCenter, they're subscribing to RotoWire.
In this environment the edge you once had from doing the research and knowing the player pool has shrunk. I'm not arguing it's impossible to win by out-predicting your opponents, it's just much harder. I try to pick my spots - sometimes I'll see a consensus opinion that seems off, or a point spread that doesn't make sense to me. Sometimes, for whatever reason, I'll have a particular insight or observation that might have been missed. But on average the market is as good or better than anything I might forecast, and I'm not going to win all that often trying to out-predict it, even if my spot-picking turns out to be good and gives me a small edge over it.
The biggest edge lies in using the market's improved accuracy against itself. No matter how good our tools are, no matter how much the crowd's wisdom is better than a lone person's opinion, complex systems like NFL teams facing off against one another in a live game are inherently chaotic. Not only are there so many variables surrounding the physical and mental states of the coaches and players and the interactions between them, but teams adjust to the success or failure of different plays and schemes on the fly during the game. And of course, there are odd bounces, penalties, turnovers and injuries.
As a result, no one knows how a given game will play out, and even thinking probabilistically is problematic because each game is a separate event and unlike any other game in many relevant respects. NFL teams are not decks of cards where the probability you hit the flush on the turn or river can be known with near certainty. While an expert - if he's really good - can model relative team strength and conceivably beat Vegas 55 percent of the time, his model has no idea about any particular game, only what's likely to happen over the long haul in NFL games generally. They're simply banking on certain factors usually being relevant predictors often enough to profit against the spread including the rake. Even the best pros don't know when it comes to specific outcomes.
It seems we're in an odd place, where the available knowledge is easier to find, and it's harder to win with knowledge alone. But the ubiquity of information only makes it easier to get up to speed on what there is to know. It does not make a significant dent in the unknown, and perhaps unknowable. It's that region from which your biggest edge originates.
The market's improved accuracy and the individual's ease of accessing it can lead to overconfidence. If you know all there is to know, then you're as much an expert as anyone else. Many in your circles will also possess this expertise, and you'll reinforce each other's quite reasonable feelings of self-assurance. But your predictions will often be wrong, and your overconfidence can therefore be exploited.
One way to exploit overconfidence is in situations with asymmetric payoffs. A good example is last week's Survivor slate where the Vikings were 94 percent favorites per Vegas against the Bills, and no other team was higher than 80 percent. You didn't have to believe the Bills would win the game - no rational person would have predicted that - only that they *could* win it, and that possibility was real enough not to be dismissed entirely. Even if 94 percent was the right number - and we really have no idea because, again, NFL teams are not decks of cards - six percent is a lot different than one hundredth of one percent. But in our minds, highly unlikely and impossible often get conflated. I doubt very many people who took the Vikings were sweating the possibility of a loss before the game because they were such massive favorites, but if you were to board a plane with a six percent chance of crashing, I doubt most of you would be able to enjoy the flight. So the risk of a Vikings loss - improbable as it might have seemed - was real, and if enough people took them, the payoff was massive. (In my pool, the nine other remaining survivors took them, and I wisely faded them, but foolishly took the Jaguars, for example. Had Jacksonville beaten Tennessee at home, I would have taken down the entire pool in Week 3.)
When the sharp crowd decides that a certain running back has the best matchup and best opportunity in DFS, there might be an asymmetric payoff in using another back with nearly as good projected usage and an ostensibly bad matchup. It's a worse choice straight up, maybe even a much worse one, but the discrepancy in the reward for being right might be much larger than the added risk you're wrong. No one knows what's going to happen, and probability and modeling only get you so far. The unknown is still vast enough that you can trust in chaotic outcomes much of the time and position yourself to collect disproportionately when improbable results occur.
Nothing I'm saying is especially ground-breaking - most of you are familiar with these concepts already. Why then do I believe this edge is still available? Why hasn't everyone just exploited it out of existence already? It could happen, but I think there are a few reasons people don't like to play it this way.
First, you lose more often than you win. The magnitude of the win will outweigh the sub-.500 winning percentage, but people don't like to lose that often. Even a small win is validating, and it's difficult to trust your process, or keep going with a strategy where you're not getting regular positive feedback. Making it to Week 7 or 9 every year in Survivor feels better than losing in Week 1 or Week 4, even if the player taking sensible risks always has a better chance to win the entire pool than the one playing simply to survive with the chalk every week.
Second, as mentioned above, I don't think as a species we're very good at distinguishing between 1 in 1,000 and 1 in 1,000,000. Both just strike us as "not gonna happen." For that reason, unlikely but possible events to which major rewards are attached, don't appeal to most people. But if you're getting a 10,000:1 payoff, it's incredibly important to distingish between those two degrees of improbability.
Third, I think the acquisition of knowledge can create a sense of superiority - if you have the best model out there, or you understand the key predictive metrics, you might feel good about it. Every time an ignorant rube suggests a team is good because it had a winning record last year (even though its underlying, predictive metrics were below average,) it reveals a gap between you and him. You're part of the club, and he is not. You get it, and he doesn't. If you identify as a knowledgeable expert as opposed to fools who haven't yet attained the understanding you have, sticking with the most probable predictions will affirm your sense of self. But your edge in doing that will only be significant over total novices. When matched up against other experts who have also mastered the intermediate course, it shrinks to near zero. Even so, many will remain content in their relative expertise and superiority and take delight in demonstrating the difference between themselves and the fools.
In sum, you'll always have to pick your spots. Many situations do not offer enough asymmetry in the payoff to justify the added risk of selecting the less probable option. But other situations - particular weeks in Survivor, DFS, later rounds of season-long drafts - are full of such opportunities to find a significant edge, and evaluating those opportunities is probably the key skill once your competition has access to all the knowledge you do.
Week 4 Trivia
In honor of Ryan Fitzpatrick becoming the only QB in NFL history to throw for 400 yards in three straight games, can you name all the ones who did it in two straight?
Guessing the Lines
|Game||My Line||Guessed Line||Actual Line||ML-AL|
|Vikings at Rams||6.5||6||6.5||0|
|Jets at Jaguars||7.5||6.5||7.5||0|
|Dolphins at Patriots||7||6.5||7||0|
|Eagles at Titans||-2||-2||-3.5||1.5|
|Texans at Colts||3||3||1||2|
|Bills at Packers||10.5||10.5||10||0.5|
|Lions at Cowboys||1.5||3||3||-1.5|
|Buccaneers at Bears||3||3||2.5||0.5|
|Bengals at Falcons||4.5||6||5.5||-1|
|Seahawks at Cardinals||0||-1.5||-3||3|
|Browns at Raiders||0||3||2.5||-2.5|
|Saints at Giants||0||-2.5||-3.5||3.5|
|49ers at Chargers||9.5||8.5||10.5||-1|
|Ravens at Steelers||3||3||3||0|
|Chiefs at Broncos||-2.5||-3||-4.5||2|
This shapes up to be a tough week, with so many of my lines right on the actual number. At first glance, I'm on the Titans, Colts, Browns, Giants and Broncos, but of course I reserve the right to change my mind by tomorrow in Beating the Book.
Week 3 Observations
•Kerryon Johnson looks like the second best back in the rookie class behind Saquon Barkley. Johnson is smooth, shifty and breaks tackles. He went 16-for-101 and caught two of his three targets for five yards. His fantasy value might lag his skills, though, with LeGarrette Blount (16 carries, two catches) and Theo Riddick (three catches for 36 yards) clogging things up.
•Sony Michel (14 carries, 50 yards, one catch on three targets for minus-1 yard) is the only game in town. Rex Burkhead didn't receive a single carry before leaving with a neck injury, and James White's 37 rushing yards mostly came on the final drive when New England was letting the clock expire. White is still the pass catcher of the backfield, of course.
•Tom Brady did nothing, and neither did his receivers. Brady had plenty of time to throw, but no one was open. Fortunately, he should have Josh Gordon next week, and Julian Edelman the week after that. Hard to know what to make of Gordon's role, since he was a healthy scratch, but desperation is the mother of opportunity.
•Matthew Stafford played passably, spreading the ball to his three receivers and Riddick, but he wasn't asked to do a ton with a lead the entire game.
• I'd rather watch a period-piece romance than a Cowboys game, and when the opponent is a Brian Schottenheimer-led offense missing its top receiver, it features the ugliest cast in cinematic history.
• Zeke Elliott always gets his, and he should have had a much bigger game had he not stepped out of bounds before catching a would-be long TD. But that's the only thing the Cowboys have going for them on offense, and it's not nearly enough.
•Dak Prescott had 4.9 YPA, two picks and took five sacks. Why even suit up? Once again, none of his receivers are remotely worth using, and his coach is running an offense from 1973.
• So much for Rashaad Penny's smooth running last week. Chris Carson went 32-102-1 with two catches for 22 yards, and Penny only three carries for five yards and no targets. Nick Vannett caught four of five targets, ostensibly regaining the No. 1 role from Will Dissly. Not that anyone outside of my two-TE, Stopa league cares.
•Josh Rosen couldn't get it done, throwing a key pick (though it was fourth down) on the potential game-winning drive, but I have to think once that box was opened, the Cardinals can't close it, and why should they? Sam Bradford looks like Methuselah in his final decade out there. (Bradford's 8.3 YPA was fine, but he threw two picks and lost a fumble that cost the Cardinals the game.)
•David Johnson caught a 21-yard TD, but had only 12 carries for 31 yards. He remains in purgatory until the offense wakes up.
•Christian Kirk caught seven of eight targets for 80 yards, while a banged-up Larry Fitzgerald caught two passes for nine yards. I'd say this might be the "changing of the guard" but that would mean the Cardinals were Buckingham Palace.
• Mitch Trubisky is failing the eye test, and the stats back it up – 6.3 YPA, a pick and three sacks. Maybe he gets better, but this is Year 2, and the QBs that have it, like Baker Mayfield, usually prove it sooner rather than later.
• I have no idea why the Bears want to feature the 165-pound Taylor Gabriel so much (they even used him as a lead blocker on a play), but he got 10 targets for 34 yards. Will Burton, Allen Robinson and Anthony Miller (who separated his shoulder) were the other key targets.
• The Rams are easily the best team in the NFL. Picking No. 2 is the tougher call.
•Jared Goff had another 354 yards on 9.8 YPA and three TDs, despite his team leading the entire game, losing a TD pass to Brandin Cooks on a replay-overturn and his team scoring on a blocked punt. The Rams fantasy-relevant players perform irrespective of game flow.
•Todd Gurley had a ho-hum 23-105-1 and five catches for 51 yards. Business as usual. He did lose a fumble, but that's never been an issue for him.
• The big-three Rams receivers split the work again with Roberts Woods (11-10-104-2) leading the way. Cooks went 8-7-90 and Cooper Kupp 6-4-71-1.
•Philip Rivers had a passable game, but Keenan Allen (7-3-44) was shut down despite the Rams two best corners leaving with injuries. Mike Williams (7-4-81-2) looks like a serious breakout candidate in Year 2.
• Every week Melvin Gordon (15-80-1 and two catches for four yards) seems to get his, but Austin Ekeler (4-for-47, three catches for 24 yards) was more efficient on lower volume. (Ekeler fumbled, but it was while he was fighting for a first down on fourth down, so the team would have turned it over anyway.)
•Eli Manning played much better Sunday, thanks to better offensive line play (who knew replacing Erick Flowers was three years overdue?) Still, Manning took four sacks – essentially he can function competently with a clean pocket and elite weapons in Pat Shurmur's system, but once things break down, he's toast. His 10.2 YPA, two TDs and no picks pass muster, however.
•Odell Beckham (10-9-109) catches anything thrown anywhere near him. Sterling Shepard (7-6-80-1) benefitted from Evan Engram's absence (left in the first half with a knee injury) and also has Manning's trust.
• After catching 14 passes in Week 2, Saquon Barkley went a modest 5-5-35, but also had 17 carries for 82 yards and a TD. Call me a homer, but he's already the best back in the league, breaking tackles every time he touches the ball and always a threat to get outside for big gains. His performance inspired this Twitter poll:
Will the Giants regret having taken Barkley over Darnold?— Christopher Liss (@Chris_Liss) September 23, 2018
•Rhett Ellison (3-3-39-1) could be worth a look in this offense if Engram were to miss significant time.
•Will Fuller (11-5-101-1) always gets his when Deshaun Watson is playing, though keep in mind Eli Apple, the Giants starting left corner, missed the game. DeAndre Hopkins (10-6-86) still has arguably the WR highest floor in fantasy football for receivers each week.
• Watson had a big day playing catch-up, with 385 passing yards, two TDs and 36 yards on the ground. He threw one pick and took three sacks.
•Lamar Miller contributed in the passing game (6-5-41-1), but was stuffed on the ground 10 carries for 10 yards. The Texans line is among the worst in the league.
•Joe Flacco had a passable game while playing with a lead, and John Brown – 9-5-86 – continues to be the star of the receiving corps now that he's finally healthy. Michael Crabtree 10-9-61 was his usual high-volume, modest-output self.
•Justin Tucker made two more 52-yard field goals, and his only miss was blocked.
• I can't think about the Titans-Jaguars game anymore. Suffice it to say, the way for a bad team to beat the Jaguars is to get into a slugfest with them and avoid mistakes. The Titans executed this to a tee with no turnovers, only 30 yards in penalties, three sacks taken and three of four field goals made.
•Blake Bortles neither tried nor succeeded in doing anything. Maybe the Titans defense is good, but the Jaguars were too conservative and shouldn't have punted the ball away on 4th-and-4 on the game's penultimate drive. I get being aggressive against the Patriots last week and being conservative against the Titans here, but they took it too far.
•Andy Dalton racked up a ton of yardage despite losing A.J. Green (groin) for much of the game, but he threw four picks. With Green out, Tyler Boyd, seemingly a second-round bust, went 7-6-132-1. If Green misses any more time, Boyd's presumably the team's No. 1. Tyler Eifert also had a mini break-out, going 8-6-74. If he can stay healthy, he'll be a red-zone threat too.
• Gio Bernard ran well (12-for-51) and caught five of nine targets for 25 yards. He's the only game in town until Joe Mixon comes back.
•Cam Newton did what he had to – two rushing and two passing TDs on modest yardage.
•Christian McCaffrey finally got the workload Ron Rivera seemed to promise and delivered big – 28 carries for 184 yards – though he had only two catches for 10 yards. It's a fierce competition between McCaffrey and Julio Jones for who can avoid the end zone the longest, however.
•Devin Funchess led the Panthers with a 7-4-61-1 line, but there wasn't a ton of passing from the Panthers.
•Aaron Rodgers took four sacks and struggled to generate offense against the Redskins. Washington's defense might be good, but Rodgers is gimpy and has below-average weapons. He has to be ranked below Drew Brees and Pat Mahomes at the very least right now.
• The Packers split carries between three backs, but unsurprisingly Aaron Jones (6-for-42) was the most effective. Who knows what Mike McCarthy will do, though?
•Davante Adams went 9-7-52-1 and Geronimo Allison 4-2-76-1, while Randall Cobb did little. Jimmy Graham went 7-5-45, and Ty Montgomery 7-6-48 in addition to carrying the ball four times for 16 yards. Jamaal Williams had two catches for 16 yards (and five carries for 29), while Jones caught one pass for five yards.
•Alex Smith attempted only 20 passes, but managed 220 yards (11.0 YPA), two TDs and one pick, while not taking a sack.
•Jordan Reed (7-4-65) was Smith's primary target, while Vernon Davis led the team with 70 receiving yards and Jamison Crowder and Paul Richardson caught TDs. Josh Doctson and Chris Thompson did nothing. (This is not an easy offense to project each week.)
•Adrian Peterson had a monster day – 19-120-2. He didn't receive any targets, but there's no doubt who the bell cow is here. There aren't more than half a dozen backs who should be projected for more rushing TDs right now.
•Carson Wentz was pedestrian in his return, but made some characteristic amazing escapes and throws on the run. He passed the eye test despite modest stats.
• Eagles wideouts did almost nothing with tight ends Zach Ertz (10-5-73) and Dallas Goedert (7-7-73-1) seeing most of the looks. Wendell Smallwood went 10 for 56 on the ground with a score and caught three passes for 35 yards. Corey Clement went 16-for-56 on the ground and caught three passes for 19 yards. So long as Jay Ajayi is out, the two will split the work fairly evenly, it seems.
•Andrew Luck managed only 4.1 YPA (40 passes, 164 yards) and one TD. He had a 33-yard run and didn't throw any picks, but took two sacks.
• The 16.5-point underdog Bills shocked the Vikings, blowing them out at home in one of the NFL's all-time upsets. Josh Allen starred with 8.9 YPA and a TD in 22 attempts. But it was his 10-for-39 on the ground, including an epic hurdle of a Vikings defender to pick up a first down, along with two rushing scores that dropped jaws. While it was an amazing play (and game) from Allen, he still has arguably the worst weapons in the league, and the Vikings – who struggled in Green Bay and were lucky to win at home against the 49ers – might not be the defense we thought heading into the year.
•Kirk Cousins had a poor game, but Adam Thielen always gets his – 19-14-105. Stefon Diggs was held to 10-4-17. Latavius Murray had seven targets, five catches and 30 yards receiving, but saw only two carries for one yard in the blowout.
•Derek Carr had another efficient game minus the soul-crushing picks. He went 27-for-39 for 345 (8.8 YPA), but threw two interceptions vs. only one TD. Jordy Nelson went 8-6-173-1, Jalen Richard went 7-6-59 out of the backfield and tight end Jared Cook went 6-5-31. That left nothing for any of the other receivers, including Amari Cooper, whose every disappearance brings back memories of last year.
• The Dolphins are 3-0, and Ryan Tannehill is now 10-1 in his last 11 games. He went 17-for-23 for 289 yards (12.6 YPA), three TDs and no picks. He took only one sack and led the team with 26 rushing yards.
•Kenyan Drake did nothing, and the receiving yards were mostly spread between Albert Wilson and Jakeem Grant. Wilson and Grant each scored TDs from Tannehill, and Wilson also threw a 52-yard TD to Grant on a trick play. Kenny Stills caught the other TD, though no receiver had more than five targets. (Personal note: I benched Wilson in my 14-team non-PPR league less than an hour before the start of the games for D.J. Moore.)
• Pat Mahomes continued his early-season record-book onslaught, racking up three more TDs in the first half. But the Chiefs, nursing a big lead, scored only three points in the second, and Mahomes' final stats – 24-for-38, 314 yards (8.8 YPA), three TDs, no picks – were merely very good, not jaw-dropping for the first time this year. It's unclear whether the 49ers found something other teams can copy, or whether the Chiefs just took their foot off the gas with a big lead.
•Kareem Hunt struggled on the ground, but had 18 carries and scored twice. He failed to catch his only target, however, and it's concerning in PPR formats that Damian Williams and Spencer Ware are more involved in the passing game.
•Matt Breida appeared to suffer a serious injury himself in the first half, but returned to lead the team with 90 yards on 10 carries. Alfred Morris had 14 carries for 67 yards and a score. Breida is iffy for Week 4, while Morris, who sprained his ankle, seems slated for a heavy workload.
•George Kittle led the team with five catches on seven targets for 79 yards, while Marquise Goodwin (quad) and Kyle Juszczyk scored TDs. But all bets are off with Garoppolo out. Pierre Garcon caught only one of four targets for 11 yards.
•Drew Brees has aged even less than Tom Brady. His spin move to beat two defenders for the game tying TD was characteristic of a much younger player. All told he had an absolutely monster fantasy day – 396 passing yards, three TDs, no picks and two rushing TDs.
•Michael Thomas is on an all-time pace after a 10-10-129 day. His 38 catches in three games prorate to 203 catches over 16. The single-season all-time catch record is 143 (Marvin Harrison), and only three times has a receiver ever cracked 130. I'd be more optimistic about Thomas getting there were he seeing uncanny volume, but instead he's doing it with an insane 95 percent catch rate – he's failed to haul in only two of his 40 targets. Last year he caught 104 of 149 looks (70 percent), still an excellent rate for an outside receiver, but probably the upper limit even for him and Brees.
•Alvin Kamara went 20!-15-154 and had 16 carries for 66 more yards, while Mike Gillislee and Jonathan Williams combined for four carries. Mark Ingram should have his full role waiting for him when he returns.
• Not to be outdone, Matt Ryan threw for 374 yards (10.7 YPA) and five TDs, three of which went to first-round rookie Calvin Ridley, who seems to be the badly-needed complement to Julio Jones. Ridley went 8-7-146-3 while Jones a more modest but still efficient 6-5-96. Of course, Jones caught none of the five TDs (Mohamed Sanu, Austin Hooper.)
• Fitzpatrick was 30 of 50 for 411 yards (8.2 YPA) and three more scores. He did throw three picks, but one was on a tipped ball. Two of the picks were bad, especially the pick-six which was a poor throw and a bad decision. He also missed a couple throws, but made several great ones under pressure. I doubt he gives up the job to Winston in Week 4, and it would probably take a couple bad games/losses before Fitzpatrick is dislodged as the starter.
• Fitzpatrick also ran for 27 yards to go along with his 411 passing. He's the first QB in league history to throw for 400 yards in three straight games, has 1230 yards on the year to go along with 11 TDs, plus 62 rushing yards and a rushing TD. Those numbers prorate to 6,560 passing yards, 59 TD passes, 331 rushing yards and five rushing TDs.
•Mike Evans had another 11-6-137-1. He's targeted often, in the red zone and down the field. Hard to ask for a much better situation for him. Chris Godwin went 10-5–74-1 and could have had another TD but for a drop. He also fumbled, but that didn't deter the Bucs from going back to him. Desean Jackson went 5-3-37 and should have had a punt return TD, but it was called back for a hold.
•O.J. Howard could finish as a top-five TE, going 8-6-72 and showing speed, burst and tackle-breaking ability. Fitzpatrick is almost certainly better for him as Jameis Winston loves Cameron Brate who went 4-3-34-1.
•Peyton Barber is ostensibly the team's lead back, but they so rarely run the ball it doesn't much matter.
•Ben Roethlisberger had a rare big road game – 30 of 38 for 353 yards (9.3 YPA), three TDs and one pick. It seems like Juju Smith-Schuster is taking over as the team's No. 1 receiver though – 11 targets to Antonio Brown's nine, 116 yards to Brown's 50. Brown did score the TD though.
•Vance McDonald had the play of the night, catching a pass in the flat and straight-arming a defender to the ground before rumbling for a 75-yard TD. It was Gronkowski-esque. McDonald finished with a 5-4-112-1 line while the other tight end, The Outlaw (Jesse James) went 1-1-7. McDonald is the better pass catcher of the two, and as long as he's healthy, the one to own.
• James Connor went 15-for-61 on the ground and caught five of six targets for 34 yards. He had some big runs on the team's final drive, but the team misses Le'Veon Bell's pass-catching skills and power-running ability.
• The Steelers defense seemingly blitzed every play. They were gashed for some big plays, but overall the aggressiveness worked decently – three picks, three sacks. I wonder whether the book on Fitzpatrick is to rush him into bad decisions because the Steelers so rarely sat back and played coverage. Or maybe it's just because the Bucs have no running game of which to speak.
• There were a lot of penalties in the Monday night game. At least three roughing the passer calls on the Steelers and one on the Bucs. It seems like the league is willing to pay a hefty price to get the defenses to change the way they play. I'm not sure it'll work.