NFL

Fantasy Football: Overcoming Biases to Find a Winning Edge

Utilizing statistical data to make optimal decisions is ideal. But we often get in our own way by failing to recognize our own mental biases.

I was a Sociology major in college. Like most other sociology majors, I've used that degree about as often as Hue Jackson has coached winning football games.

But through my own experience playing fantasy football, I've been able to see how well social psychology can explain our shortcomings as owners. I'm typically a numbers guy -- numbers are the great equalizer in settling fantasy debates -- but I've noticed that my social and psychological hangups are what get me in trouble.

So, allow me to put my degree to use for once (which will make my parents quite ecstatic) by telling you seven ways that your implicit biases are ruining your fantasy teams.

1. Overconfidence Bias

Quite simply, overconfidence bias is defined as "(1) overestimation of one's actual performance, (2) overplacement of one's performance relative to others, and (3) excessive precision in one's beliefs."

Said differently, we think we are better at most things than we truly are, including the evaluation of fantasy football assets.

As a general rule, we as fantasy players are horrible at really understanding a player's true range of outcomes. We use projections -- or, even worse, anecdotal evidence -- to pinpoint a singular valuation on each player. That leads us to having takes -- some hot, and some sizzling. The problem here is that your projections are wrong inherently. While projections may represent a "most likely" scenario which makes them a handy tool, they don't account for inevitable variance. We should be widening our range of foreseeable outcomes in player evaluation.

The Washington Redskins went all in on Robert Griffin III in the 2012 draft, giving up their 2012, 2013, and 2014 first-round picks (plus a 2012 second-rounder) to move up to number two overall for Griffin. While RG3 looked great early and suffered some unfortunate setbacks, the 'Skins clearly failed to give proper weight to the inevitable range of outcomes that accompany such a draft pick. They focused on what they believed was the most likely outcome -- a franchise quarterback for years to come -- and they did what it took to acquire him.

They weren't alone in their valuation of the player, but variance happened and the team was overexposed. Yet, had they recognized Griffin's full range of outcomes, perhaps they wouldn't have taken such a gamble. Peak overconfidence.

2. Primacy/Recency Bias

People have short memories. In social psychology, we call this "recency bias."

Primacy bias and recency bias point to the tendency that humans remember with greater frequency the first and last things that they experience respectively. Every year, based on an athlete's previous year's performance, fantasy owners are quick to overreact. While it's great that our opinions change as information changes, the overreaction to each individual season is textbook recency bias.

As a rookie, Todd Gurley was the 56th pick off the board on average, according to Fantasy Football Calculator. After going for 1,106 yards and 10 touchdowns in 12 games as a rookie, he was selected as the second overall pick in 2016. After a frustrating sophomore season where he rushed for 885 yards in 16 games, Gurley fell to an average draft position of 17 in 2017. He went on to crush expectations, totaling over 2,000 scrimmage yards and 19 touchdowns. Now, he's being taken as the number one overall selection for 2018. The market overreacted each year based on his most recent season.

Meanwhile, Josh Gordon truthers such as myself cling to our first real taste of Gordon from 2013 where he broke out for 1,646 receiving yards and 9 touchdowns. Gordon was a freak, and despite playing only 10 games in the last four seasons with modest efficiency, it's hard for fantasy owners to forget about Gordon's gaudy numbers from his sophomore year. Primacy bias is altering our ability to fairly assess Gordon's likely range of outcomes.

3. Groupthink Mentality

Imagine you're in a classroom. The instructor places a multiple choice question on the projector. You feel more than reasonably confident that the answer is "C." However, when asked for their answer verbally, each of your five other classmates answers "B."

Now it's your turn to communicate your answer.

All of a sudden, you aren't so sure of yourself. "Maybe B is correct," you think. "Maybe I'm an idiot for thinking C in the first place." Your initial confidence is gone and you conform to the pressure and repeat "B" when it is your turn. This is groupthink: a social phenomenon that explains cults, prison culture, and even fantasy football behavior by placing a higher value on conformity than accuracy and individual ideas.

The San Francisco 49ers were 1-10 last year heading into their 12th game of the season. They brought in a new starting quarterback named Jimmy Garoppolo who went undefeated in that role, winning all five games to end the season.

The legend of Garoppolo has grown. He's undefeated as a starter. Getting paid is his forte. He has an intriguing young head coach. He is objectively handsome.

The masses have collectively formed an opinion that Garoppolo is the next Tom Brady, and it's now inflating his fantasy value.

Despite a limited resume as a starter that left some questions unanswered about his abilities near the red zone, Garoppolo is being selected as the eighth quarterback off the board (in Round 6, no less!), which could move higher as the general public gets more involved. Even though he has an unenviable complement of receivers and a short track record of success, Jimmy GQ is being selected above more established veterans such as Carson Wentz, Matt Ryan, Ben Roethlisberger, and Matthew Stafford. Moreover, with a sixth round average draft position, Garoppolo is being drafted as an elite difference-making quarterback option.

Garoppolo and Brady are next door neighbors on Narrative Street, which has led to an overwhelmingly positive buzz and inflated valuations for the former. When opinions are shared between a huge mass of people, we are often afraid to voice a contrarian opinion, even if logical reason may take us there. After all, it's much easier to be wrong together than wrong alone. When you fail to see a lot of contrary arguments toward an individual or a team in general, that should raise red flags. Only when an evaluation is balanced and equally argued can it be trusted.

4. Gambler's Fallacy

The gambler's fallacy is exemplified in a casino. Imagine a roulette wheel that has hit black 10 times in a row. Gamblers tend to believe that the next 10 spins will heavily favor red to compensate for the prior 10 spins. Unfortunately, this is a misunderstanding of regression.

The facts are that future outcomes of a random event are not tied in any way to previous outcomes of the same random event. While fantasy football is not completely random and sometimes there are legitimate reasons that are creating spikes or dips, over a large enough sample, we'll always assume that performance will return to normal levels based on that player's abilities and opportunity.

"Regression" may be the most misused word in fantasy football. We sort of get it, right? When something unsustainable is occurring, it will eventually level back out.

First, regression can be either positive or negative. Amari Cooper figures to be a positive regression candidate this year because his poor 2017 (where he saw career lows in targets, completions, and yardage) was an obvious outlier compared to his strong 2015 and 2016 (where he went for over 1,000 receiving yards in each). Deshaun Watson, on the other hand, would figure to be a negative regression candidate because his 9.3% touchdown rate in 2017 was the second-best rate among all 200-attempt quarterbacks since 2000.

Second, we misunderstand what regression looks like. We tend to assume that if a player has significantly underperformed expectations, then he is due to soon significantly overperform expectations. Regression is an indication that outlier results will return to a mean, not that a counterswing will occur.

5. Loss Aversion

Psychologists coined the term "loss aversion" to explain that human behavior typically gives preference to avoiding losses over acquiring an equivalent gain. In an experiment, people were asked if they would prefer a guaranteed $3,000 or an 80 percent chance of $4,000 (and 20 percent chance of $0). 80 percent of participants elected the guaranteed money although the other option yielded a higher net expected value of $3,200 ($4,000 * 80%).

Once the participants were offered the $3,000 guaranteed, they viewed it as their possession. Taking the gamble meant taking a 20% risk of losing what they had just gained even though the $1,000 additional payout was more than fair for such odds.

Have you ever offered a trade to a leaguemate -- a good, legitimate trade that helped both sides improve -- but they still couldn't pull the trigger? Perhaps the roles were reversed and you were the one who couldn't click "accept."

My leaguemate's team was loaded with running backs while mine was flush with wide receivers. He proposed to send his fourth-best running back, Alvin Kamara, for my fourth-best wide receiver, Chris Hogan. The numbers at the time showed this was a mostly fair trade. But my brain played mental gymnastics to delegitimize Kamara while elevating Hogan. I was more afraid of what I might lose in my player than what I might gain in my opponent's. Why?!

We want to hang on to what we have. We fall in love with "our guys." We have inflated views of players on our roster because we were the ones smart enough to take them in the first place, right? This makes us suboptimal when engaging in negotiations with other teams. Remember that an opportunity not gained is as bad as an opportunity lost.

6. Hindsight Bias

"I knew it all along" is perhaps the most face-punch-worthy phrase in fantasy football.

Don't get me wrong, analysts and fans who call their shot by using sound logic and historical evidence to forecast a result should be proud at how they arrived at that conclusion. But folks who look through the rear view mirror to claim credit for that which was unpredictable are wrong for doing so. The hindsight bias is our human tendency to try to explain unpredictable events as predictable. As such, we often misinterpret how to make future decisions because we are bad at evaluating previous decisions.

Take a look at Alex Collins' 2017 season. He opened the year competing with the less-than-stellar group of running backs employed by the Seattle Seahawks and was waived on September 2nd, just prior to the season. On September 5th, the Baltimore Ravens picked him up to add to their practice squad. A week and a half later, an injury to Danny Woodhead opened the door to Collins being called up to the active roster. In a surprising turn of events, Collins shined in limited action and soon took over feature back duties for the Ravens.

If you think that Alex Collins' rise to the top was in any way predictable as you drafted your team in late August, you're simply using hindsight to back your way in to a decision-making process. You simply can't invest your time in identifying these unicorns until more data is there to suggest it has legitimate possibilities.

7. Confirmation Bias

Confirmation bias is "the tendency to process information by looking for, or interpreting, information that is consistent with one’s existing beliefs." It explains our news preferences, our selection of similar friends, and how we interpret fantasy-related data.

What will Dion Lewis' role be in the Tennessee Titans offense alongside Derrick Henry in 2018? The prevailing opinion is that Henry will dominate early down work and Lewis will be used as the third down back.

But why?

Lewis is 5'8" and 195 pounds while Henry is 6'3" and 247 pounds. They don't even look like they play the same position. But our experience and the prevailing narrative tells us that little guys catch passes and run outside the tackles on third downs. Meanwhile, big backs are used on first downs to run between the tackles. That leads us to this conclusion about the Titans' running backs.

Yet, throughout his career and despite his size, Dion Lewis has been used as an early-down specialist. Both his percentage of rushes and targets have been higher than league average on first and second downs. Meanwhile, Derrick Henry has rushed at a rate higher than the league average on second and third downs. The data suggests that Lewis might continue to see work on first down ahead of Derrick Henry if both players' trends continue in the future.

But this data flies in the face of our generally held opinions. Often, when we are presented with contrary data such as this, we will ignore it. Similarly, when presented with information that is favorable to our preconceived opinion, we will absorb and embolden our takes with that information. The opposite, however, should be true. We should seek information which challenges our opinions, not dismiss it.

Conclusion

It takes awareness to combat any of these implicit biases. We do them all subconsciously, without any mental effort, all while harming our chances to succeed. By knowing these basic tendencies of human reaction and interaction, we can easily see how they make us suboptimal in this game. By bringing the bias from the back of your brain to the front, you'll begin noticing it live as it happens. And you can adjust.

This is the winning edge.