Exploiting Recency Bias And Overpaying For Last Year: Fade Dragons, Don’t Chase Them
When players like Cam Newton, Devonta Freeman, and DeAndre Hopkins win championships, they’re usually overvalued the next year from chasing the dragon.
No article I wrote last season was more polarizing (that’s a nice way of saying it made people angry, and they insulted me on Twitter about it) than this one.
It said to fade Rob Gronkowski as a first-round pick (and top-12 overall flex player), to avoid Andrew Luck as the QB1, and to pass on Keenan Allen as an overall WR2.
Here’s how that turned out:
|Player||Call||Overall Finish||PPG Finish|
|Rob Gronkowski||Not Top 12 FLEX||FLEX15||FLEX20|
|Andrew Luck||Not QB1||QB27||QB6|
|Keenan Allen||Not a WR2||WR41||WR6|
While definitely bailed out by Allen’s injury, both the Gronkowski and Luck predictions came true on a season-long and points-per-game basis.
THE CONCEPT OF CHASING THE DRAGON
The introduction to that article still rings true:
If we dwell on our mistakes, we’ll make illogical choices in the future. This is a key issue for fantasy players: accepting that you cannot predict the future, and being comfortable playing for the most likely outcome every time, regardless of whether or not that outcome actually happens.
If you let hindsight tell you that you made the wrong decision, even though you played for the most likely outcome, you will start to do illogical, overly exuberant things.
FOMO (Fear of Missing Out) is very real, and leads people to badly overpay, abandoning any rational approach to value in hopes that they’re catching lighting in a bottle that either has already struck, or never will.
This article isn’t so much about these specific players, but more of a general way to think about recency bias, identify it when you see it, and weigh whether or not it’s playing for the most likely outcome.
You may disagree on the individual names, but what is more important is weighing the concepts, and applying them broadly to team construction.
This season, people are chasing that dragon again, for fear of missing out on the players they didn’t own last season, when they were underpriced, fantasy juggernauts.
Three examples of this unbridled enthusiasm are Cam Newton, Devonta Freeman, and DeAndre Hopkins.
It’s extremely hard to project the number one player at any position accurately. Last season, Gronkowski and Antonio Brown were both correctly drafted at the top of their positions, but Le’Veon Bell and Andrew Luck were injury-laiden disasters.
In 2014, fantasy prognosticators got none of the four positions correct. In 2013, they got only Jimmy Graham correct. In 2012, they got both Calvin Johnson and Graham correct; however, Johnson was drafted after two quarterbacks, and Graham was drafted after five. In 2011, drafters once again went zero for four.
|ADP QB1 (Real finish)||ADP RB1 (Real finish)||ADP WR1 (Real Finish)||ADP TE1 (Real Finish)|
|2011||A. Rodgers (QB2)||A. Peterson (RB15)||A. Johnson (WR76)||A.Gates (TE13)|
|2012||A. Rodgers (QB2)||A. Foster (RB3)||C. Johnson (WR1)||J. Graham (TE1)|
|2013||A. Rodgers (QB26)||A. Peterson (RB10)||C. Johnson (WR5)||J. Graham (TE1)|
|2014||P. Manning (QB4)||L. McCoy (RB12)||C. Johnson (WR16)||J. Graham (TE2)|
|2015||A. Luck (QB27)||L. Bell (RB47)||A. Brown (WR1)||Gronkowski (TE1)|
In other words, out of twenty tries over the last five years, fantasy players have gotten it right fives times total: twice at wide receiver and three times at tight end. You can see on the table, if you can remember some of the things at play, that people really do just bet on what happened the year before happening again.
Peyton Manning was the QB1 in 2013, a year in which Aaron Rodgers broke his collarbone. So of course in 2014, people bet on Manning to do it again, instead of Rodgers returning to form, even though Rodgers had been QB2 or better five years in a row prior (he was the QB2 in 2008, QB1 in 2009, and QB2 in 2010).
Well, shocker, Rodgers stayed healthy and was the QB2, Manning was the QB4, and Luck was the QB1. And, of course, despite Luck’s résumé being finishes as the QB8 and QB7 in the two years prior, people drafted him over Rodgers, Manning, and most damaging, Drew Brees.
Brees was coming off nine straight seasons of QB6 or better finishes, and people still drafted Luck several rounds higher, and got absolutely destroyed for it.
Newton, unlike Luck headed into last season, does actually possess that kind of résumé. In 2011 he was the QB4, in 2012 he was the QB4, and in 2013 he was the QB5. In 2014, he struggled with broken ribs, which affected his rushing ability, and he finished as QB17. Fantasy drafters, in their infinite wisdom, with their exceptional memories, then drafted him as the QB11 last year.
A lot of people missed out, and a lot people are chasing that dragon.
Unfortunately, there’s a problem. Exploiting people drafting Newton, simply by avoiding him, is complicated by an unpleasant fact this year: people woke up. Years of drafting quarterback high may have finally ended, we won’t know for sure until a few weeks from now, but Joshua Lake says that Late Round QB Is The New Black.
It’s even trickier because getting Newton at the turn of the third and fourth rounds also means you can get QB2 through QB5 much later than normal. There’s not a great case for why Newton is likelier than Rodgers, Luck, Russell Wilson, or Brees to be the top scoring pass-thrower, and if he has become a tremendous value, then they have as well, at much cheaper prices.
Luck threw 45 touchdown passes the season before last, while Newton averages 19.4 per season. Rodgers and Brees’ résumés speak for themselves, as detailed above. Over the second half of last season, Wilson scored at a pace that absolutely dwarfed Newton’s season-long total.
Devonta Freeman is also a little tricky because he’s not being drafted like someone who just had the kind of season he did.
Every single person on the ADP chart above, beginning in 2012, first appears following a year where they were in the top three at their position. Freeman wasn’t just top three at his position — he murdered the field. The gap from Freeman to RB2 Peterson was larger than the gap from Peterson to RB8, David Johnson. Yet, somehow, Freeman is being drafted as the RB7.
The reason is that the majority of Freeman’s scoring, or at least the games that made him seem like an otherwordly fantasy league-winner, are antithetical to the psychology we are trying to exploit.
As University of Rochester Professor of Brain and Cognitive Sciences, and contributor to ESPN and RotoViz, Dr. Renee Miller explains, Freeman’s best games occurred at the time that has the least effect on people’s thinking and memory:
“The Primacy Effect, the finding that things that happen first in a sequence of events are preferentially remembered and hold a more prominent place in our memory than subsequent events, with the exception of the most recent event (Recency Bias).”
He’s been such a popular fade, since before the season even ended, that The Contrarian had to write an article called “I Don’t Even Like Devonta Freeman And I Think He’s Wildly Undervalued” in February. In that piece, he explains that running backs with 200-carry, 50-reception seasons in their first two years are usually long-term fantasy superstars. He also argues that Freeman’s Sim Score ranges are multiples of some running backs being drafted above him.
This is a tough thing to exploit by fading.
I tried to make the case already, in my defense of Tevin Coleman and the other sophomore backs, and it basically comes down to three things:
- Coleman was the Week 1 starter last season
- Freeman scored most of his points in a very easy four game slate as shown above
- His workload was so massive to get there
His workload could be reduced, he could get injured, or Coleman could just simply outplay him at certain things. Or none of that happens, and he absolutely murders the field again.
Here we go.
|2011-2015 ADP WR4||255|
|RotoViz Writers' Composite Projection||290.5|
|Sim Scores||273.6 - 326.4|
I absolutely destroyed predicting DeAndre Hopkins last season, and it felt great. I absolutely love him, and let me tell you right now: he’s probably being drafted above his ceiling.
For starters, he’s in all likelihood not seeing 192 targets again. Hell, Dave Caban only projected him for 180 in an article about how he’s going to lead the damn league in targets.
Arian Foster played in two full games last season, after being eased back in his first game following a groin tear in the preseason, then leaving early in his fourth game with a torn Achilles’. Cecil Shorts and Nate Washington missed seven entire games, and left several others injured, between the two of them. Jaelen Strong was a raw 21-year-old rookie that barely played at first, following wild fluctuations in his weight through the offseason and into the season.
Hopkins’ competition for targets couldn’t have been sparser. This year, Shorts is back, Strong is a year older with a year of experience in the offense and the league, they signed Lamar Miller in free agency, and drafted two wide receivers in the first three rounds of this year’s (real, NFL) draft.
On top of that, when you split out the second half of last season, Hopkins’ targets absolutely plummeted.
Now let’s consider that the team signed Brock Osweiler to be their new quarterback, and people are assuming he’s competent enough to provide Hopkins with the same opportunity that the combination of Brian Hoyer and Ryan Mallet were.
Let me ask you this: what if Hopkins isn’t actually QB-proof? What if Hoyer is actually kind of not-that-shitty?
What if Osweiler and Mallet are both absolute dumpster fires? Looking at the adjusted yards per attempt, and the location graphs, of all involved raises a giant red flag about if Osweiler can get Hopkins the ball on the type of plays where he thrives:
Hopkins has the lowest Volatility Score of any wide receiver, which means his range of outcomes is extremely narrow, albeit relatively, safely high. He’s also very unlikely to finish in the top four fantasy wide receivers this season.