It was a day like any other. That is, until it became anything but ordinary.
“Son, how was school? We need to talk.”
My father was waiting for me near the front door, which remained slightly ajar – something that took me by surprise considering it was September in Arizona. You see, there are effectively two seasons in the greater Phoenix area: summer, and not-summer. September is very much still summer in the Valley of the Sun.
Everything up until this point felt like any other day. School, varsity soccer practice, a ride home from a teammate. I racked my brain to pinpoint what it was my pops was referring to. I got nothing. I walked through the threshold and into the family room.
“What the f—k is this?”
My dad flashed an envelope with a logo on it that I had grown to know during the previous few months. Crap. I forgot to let my dad know that this was coming.
Three months before that awkward interaction with my dad, I was at club soccer practice when a new teammate started talking about his credit card. His family was well-to-do. The card carried a low limit. But I got this crazy idea in my head. I couldn’t get the mental image out of my mind. Chris Moneymaker standing there, surrounded by family and friends, holding up the World Series of Poker Main Event bracelet. No freaking way – a complete amateur won a seat into the Main Event and bested the entire field, taking down $2.5 million after beating 838 other players. It was the American Dream. This unknown player achieved the impossible.
Welcome to the poker boom. New sites were popping up seemingly monthly. Everyone was trying to get their piece of the ballooning pie. And I was trying to get my action in. I was 14 at the time, but I had a secret weapon. I had the same name as my dad. I had an account, and now I had the means to deposit money and hit the virtual felt.
I started off with a $50 deposit, which was torched in the first half hour on the site. Yeah, I’m sure virtual blackjack on an unregulated site was completely fair, with nothing nefarious going on. I called up my friend again.
“Yo dude, can you deposit another $50 for me? Yeah, man, I’ll get you at practice.”
I didn’t make a ton of money as a soccer referee, but it was something. 25 bucks for a linesman and 50 bucks for primary was big-time for a 14-year-old. I could only work center for teams my age or younger, which capped my potential earnings. However, it was good money for an early teen.
I was going to make this $50 last. I started playing 0.05/0.10 PLO8 (Pot-limit Omaha Eight or Better) and eventually worked my bankroll up to $500. That’s when I decided to take a shot at MTTs (Multi-Table Tournaments). It was my first time entering a large field poker tournament, so I had to keep the buy-in low.
“On snap, they’re running a $2 rebuy/add-on in an hour,” I thought to myself as I scoured the lobby.
That was it, that’s going to be my first tournament. I bought in and hopped in the shower.
Back then, rebuy tournaments allowed you to top up your stack for double the buy-in right when the tournament started, giving you double the starting stack before a hand had been dealt. You could then rebuy as many times as you wanted before a predetermined time, at which point you could buy an add-on for another buy-in to get what amounted to 2.5x the original starting stack added to your chip stack. After a couple of hours, I was sitting with an above-average stack, approaching the money bubble, after making just a $6 investment.
I played for another four hours until a sudden euphoric feeling overwhelmed my senses. I had made the final table in my first-ever MTT. Holy smokes, this is great!
I remember that feeling getting amplified with each knockout until only three players were remaining – me and two others. And I was narrowly the chip leader!
“Hmmm, pocket fours on the button, three-handed,” I mouthed after the cards were dealt. I raised 3x the big blind, as was customary practice back then — two callers as both the small blind and big blind compete.
The flop is peeled. 4QT, rainbow. “Rainbow” was a cool new term I had learned that described a flop in which none of the suits were repeating. Oh, man, I flopped a set three-handed at the final table of my first-ever poker tournament.
What’s more, there is a bet and a raise in front of me. What do I do in this spot? Do I get tricky with a call and let them hang themselves, or do I put the pedal to the metal and try to end this thing now? The amount of aggression out of position before I acted led me to the latter. I three-bet, hoping to get the money in the middle. Four-bet all-in, call for less. Wow. The money got in. Let’s run it. I call.
Small blind turns over KQ suited. Big blind turns over QT off-suit. Oh my God, oh my God. I need to dodge running hearts, a queen, and a 10, and this thing is over! Turn bricks. My heart is beating out of my chest. River bricks. The player icons disappear, replaced by a winner’s banner in the middle of the table.
“Congratulations, you have won $1,504.”
By this time, it was after 3 AM, which meant I had played for just over eight hours. But I freaking did it! I felt like Chris Moneymaker, man. I freaking did it! I didn’t know what to do. How do I get this money? I’m playing under my dad’s name and birthday, and I used my friend’s credit card to deposit. I guess my only option is to request a check to be mailed to me (my dad). I crawled into bed, knowing I was doubtful to be going to school when 7 AM rolled around in under four short hours.
That’s where it all started for me. My childish infatuation with the glamorous life of a professional poker player drove me towards games of all shapes and sizes, game theory, human psychology, crowd psychology, figuring out how the brain works, game plan development, variance, sustainability, and pretty much anything I could get my hands on regarding the relatively new fields within the theory of games.
And yes, after a stern talking to, a massive explanation on my part, and a night to sleep it off, my dad cashed that check from Paradise Poker for me.
Turn the clocks forward nearly a decade to the launch of daily fantasy sports, or DFS. An exciting new game built around sports that pits players against each other? Oh, man, I was hooked. I wanted to dissect the game to find edges in places the field simply wasn’t looking. I crunched the data. I turned to game theory and psychology. I studied in online classes from Harvard, Northwestern, Yale, and Stanford. I wanted to take in as much as I could about this exciting new game. But the more I dug, the more I realized that players were not playing this game right, it was playing them.
The remainder of this piece will dive into the “why” and “how” of that statement. We’ll start by exploring the inner workings of the brain, how the game of DFS, and gambling altogether, preys on those tendencies, dive into the fascinating fields of human psychology and crowd psychology, discuss how dominant risk profiles work as they pertain to nonlinear payout structures, talk through variance, ownership, contest selection, and sustainability in DFS, and finish with a crash course in how to play this game most profitably. Game theory is at the core of this analysis, which aims to generate repeatable habit patterns to maximize expected value (EV) in the game of DFS.
This will not be an all-encompassing piece on how to effectively play DFS. We don’t have the time to completely go through all the inner workings of this game in a single written piece of content. Instead, my goal in this article is to generate an “aha!” moment for you. Come with me as we explore the game that is playing you.
The Pleasure Chemical
Deep in the center of your brain is a system that regulates the euphoric feeling of reward. The chemical at play has the power to influence levels of satisfaction, mood, attention, learning, and pleasure, and is also crucial for your body’s motor control and cognitive functions. If you haven’t guessed it by now, we’re referring to dopamine.
Dopamine is a neurotransmitter, or chemical messenger, controlled by a reward prediction and modulation system of the nucleus accumbens in the center of your brain. This system constantly recalibrates itself, adapting to recent levels in a predictive manner, designed to maintain your satisfaction levels, mood, attention, and pleasure. The current state of society overloads these predictive forces of dopamine regulation. Social media updates, on-demand television, video games, sugar-packed foods, immediate connectivity, and the need for constant human interaction combine to lower dopamine output in the brain over time, from activities that once felt fulfilling. The levels of dopamine released from these activities are then adjusted by the nucleus accumbens, reducing the effect on your mind (less pleasure), which, in turn, makes you seek more pleasure for the same dopamine hit.
This mind trick, if you will, is the driving force behind the success of the gambling industry. Slot machines use variable-ratio schedules, or randomized payouts, to prey on the promise of reward. Randomly shuffled cards in a shoe influence the randomness of payouts in table games like Blackjack, Baccarat, and Ultimate Texas Hold’Em, appealing to this dopamine regulation. The unpredictability of the dice in Craps drives your nucleus accumbens wild. The variance in fantasy football, from your season-long leagues to best ball to DFS, creates the optimal environment for the pursuit of dopamine in your central nervous system. But what does all of this have to do with playing the game itself? The answer to that question has only been unlocked in recent years, specifically in relation to the anticipatory phase of reward.
Recent studies in neuroscience have concluded that the promise of reward is more satisfying to your brain than the reward itself. In this research, scholars found that dopamine is released not primarily during the consumption or receipt of reward, but rather in anticipation of it. Functional MRI studies in humans reveal increased activity in the brain regions that house this system, and the greater the potential reward, the stronger the brain’s response. This anticipatory dopamine surge contributes to feelings of excitement and motivation, plundering your cognitive reason in the process while also overloading your nucleus accumbens, forcing you to seek more dopamine hits for the same feeling. What’s more, uncertainty only magnifies this anticipatory response. Studies show that the amount of dopamine released in anticipation can increase significantly when the certainty of reward is less predictable.
What this means to us is that humans are highly motivated by prospective rewards. Whether it be the hunt for the perfect item on a shopping trip, the potential of hitting a jackpot while gambling, the expected response from a post on social media, or the promise of your favorite beer or snack after a long day’s work, human beings are quite literally wired to seek out activities that carry potential recompense.
In a DFS setting, this typically drives players towards the most top-heavy contests, where the potential for reward is much greater – casuals playing a few entries per week in the Milly Maker on DraftKings instead of a contest that is more tailored to their play style, seasoned veterans taking shots on contests beyond the means of their bankroll, players max-entering contests when their bankroll doesn’t support it, and immediately jumping in stakes after a large score are all examples of the effects of dopamine on DFS play.
Human Psychology
While it’s true that humans are driven by the promise of reward, an arguably greater impact on our decision-making processes is the fear of what we stand to lose. Like the field of neuroscience, this psychological aspect of human existence is rooted in our evolution. Fear of loss can most commonly be associated with the fundamental human need for security and protection, triggering great loss aversion in our core. Moreover, recent research in psychology points to the pain of losing something being felt more intensely than the pleasure of gaining something of equal value, leaving humans biased towards a perpetual state of fear-based anxiety.
Parlaying that research into the field of game theory yields a basis for risk-dominant tendencies (more on this below) in games involving competing agents, or players, effectively leading those players to adopt strategies that minimize losses or the potential for loss. Studies also found that uncertainty only exacerbates these predispositions, leading to a stronger yearning for safety and control in situations where variance exists. Sound familiar? That’s because this behavior is present in the game of DFS, which can be described as a Bayesian game of incomplete information.
One of the key assumptions in Bayesian game theory is that players hold private information and beliefs about the unknowns, which triggers a fear-based anxiety when combined with uncertainty. However, playing any game optimally means adhering to equilibrium theory (Nash equilibrium), which considers both potential payouts and potential losses as we develop an optimal approach to the game. This process is defined as game plan development, which we will finalize at the end of this piece. Suffice to say, inherent psychological aspects of DFS place us in an uncomfortable situation, driven by potential reward while fearful of what we stand to lose. What this does in a DFS setting is create a scenario where an imbalance exists between the contest we’re playing and our strategy in that contest.
Crowd Psychology
Crowd psychology is a subset of human psychology that aims to describe how humans interact with each other in certain group situations. A notable conclusion that has been widely accepted by scholars in this field in recent years is that individuals often deviate from typical behaviors when placed in a group setting. Gustave Le Bon, one of the founders of the field, found that crowds promote irrationality and heightened suggestibility. These tendencies are found everywhere – political parties, stock market bubbles, emergent situations, social media influence, pedestrian flow, and even traffic patterns. How do people react when faced with opinions of a different political party? How do those parties interact with each other, and how do members of the same political party interact with each other? How does a market bubble occur? How do people behave when there’s a fire? Do we see orderly exits or primal senses of survival take over? How about the effects of social media influence? Ever thought about why “influencer” is even a thing?
Here’s a fun one – check out the lines at the next theme park you go to. And I’m sorry in advance; once you notice this, you will never be able to look at such a normal situation the same again. Why do cars keep going to the line with the most cars in it already, when they could skip 10 cars in the line right next to it? The same phenomena are present when going through security at the airport. People flock to the longest line because there must be something about that line that makes it safe. You see the same things exiting a highway when multiple turn lanes are going the same direction. Why are there seven cars in one turn lane and zero in the other? All these examples involve irrationality and suggestibility, things that are present in the game of DFS.
“My favorite analyst says this is a good play, I better play him this week.” Sound familiar? Or what about, “wow, that player is approaching 30% expected ownership, he must be a good play!” Or even better, “this game is clearly the game of the week, it has the highest Vegas total.” I can almost guarantee that you have thought all these things to yourself at some point in your recent DFS career. That is crowd psychology, or groupthink, at work – suggestibility and irrationality resulting from outside influence.
What Do You Win When You’re Right?
There’s a simple game in game theory that highlights the interaction of what you stand to win versus what you stand to lose: the prisoner’s dilemma.
In this game, two people are apprehended for a crime, but the authorities have no hard evidence. They give both people two options: either hold out or confess to the crime. But the stipulation is that if both hold out, both serve one-year jail sentences. If both confess, both serve two-year jail sentences. If one holds out and one confesses, the holdout does no time and the confessor does four years. In this game, the minimum amount of combined time served would be two years if both individuals hold out; however, the minimum time served by one individual is zero if the other person confesses. Each player would then be incentivized to hold out and not confess, because they would either serve two years or four years by confessing. There is no incentive to confessing in this game.
Let’s now consider this simple exercise in a DFS setting by looking at the two extremes in payouts. You stand to win the exact amount of your buy-in in double-ups, while you stand to win 50,000 times your buy-in in the Milly Maker. In which contest would we want to take on more risk? The Milly Maker. The same reasoning applies to field size, the number of entries you can place in a contest, and the maximum allowable entries in a contest. The smaller the field and the fewer the maximum allowable entries, the less risk we should adopt. The larger the field or maximum allowable entries, the more risk we should adopt. In this example, risk is equivalent to variance, which we will define and discuss below.
Risk Dominance Versus Payoff Dominance
DFS is not a game about knowing football (or baseball, or basketball, or golf, or tennis, or whatever DFS sport you are playing). You can’t consistently win by simply understanding the nuances of offensive coordinator and defensive coordinator tendencies, route participation rates, yards per carry, broken tackles per rush attempt, or any of the other vast descriptive statistics we use in the fantasy football industry. Why? Variance. These tools help us reduce the player pool for a given week by removing players who project poorly, but they cannot build a roster that is profitable for you. You must do that work, and you cannot do that well without a solid understanding of variance.
Simply put, variance is all the aspects of a game you can’t control. You can’t control the decisions of other players, the way the cards are dealt in a hand of poker, the decisions of a coaching staff, mistakes from players on the field, weather, the way the ball bounces, busted coverage – the list goes on. And to best harness variance, we need to understand dominant theory. More precisely, risk dominance versus payoff dominance.
Risk dominance refers to the tendency of players in a game to choose strategies that minimize potential losses. In contrast, payoff dominance refers to a criterion for choosing between multiple equilibrium points in games of competing agents. Both are considered deviations from Nash equilibrium, which is a condition in a game when no single player can generate a meaningful edge over all others. First off, we know DFS is not “solved,” being a game far from equilibrium. Taking what we’ve already learned in the areas of human psychology, which deviation would you expect to be the percentage solution for the field, of the two deviations provided? Risk dominance.
To put this another way, the field is generally, and unknowingly, executing a deviation from Nash equilibrium in DFS, and the percentage solution is that they are adopting a risk-averse strategy aimed at minimizing their potential for loss. This isn’t entirely their fault; it’s a byproduct of the human condition paired with a game that is far from solved. Simply put, the game of DFS carries too much variance for us to focus on what we stand to lose. The theoretics explored to this point would optimally shift our focus towards what we stand to gain due to the doctrine of payoff dominance, particularly if the players in a given game are focused more on what they stand to lose.
And there’s the “Aha!” moment. You, the person reading this, are too focused on what you stand to lose in DFS – a game with borderline infinite variance.
A Game of Borderline Infinite Variance
You can control only what you can control. My pops raised me with this mantra in mind, something that I have passed on to my kids. As kids grow, they constantly seek approval for their actions and thoughts. That is their reward. This leads to increased anxiety in our youth, which is elevated by uncertainty (they don’t know if they’ve done something you will approve of or something that will make you mad). Reminding them to only worry about things they can control, instead focusing on making the best decision they can at the time given the situation, can help them remove some of that uncertainty. It’s the same thing in DFS (or other games, or life, for that matter).
A starting NFL DFS roster typically contains nine positions – a quarterback, two running backs, three wide receivers, a tight end, a flex, and a defense. Suppose we remove salary constraints and talent on the field. In that case, the odds of landing an optimal roster on a full slate, with 13 games being played, is on the order of one in three trillion (26 quarterbacks, 26 starting running backs, 39 wide receivers, and so on). Salary constraints, situation, and talent allow us to reduce that number slightly, but you get the picture. This is the reason there has never been an optimal roster win the Milly Maker on DraftKings in its 17 years of existence. We simply don’t have a large enough sample. Statistically speaking, we’ve had about 72 million rosters played in the Milly Maker, assuming an average of 250,000 entries per slate and 17 slates per year (there are some with more and some with less). On that pace, again, statistically speaking, it would take 41,667 years to have a large enough sample size to make the odds of an optimal roster being hit 1:1. We’re talking an extremely small sample size of data relative to current odds.
Sample size is inversely proportionate to variance in that variance is going to be higher the smaller the sample, and that’s before we consider the variance associated with a game played between 11 flawed human beings on each side, the way an oblong ball bounces or flies through the air or is influenced by the wind, the variance in coaching tendencies and decisions, situational variance regarding fourth down tendencies, weather, or injuries. The list goes on. All these things are present in DFS – things we can’t control – and all of them add to variance. So, if you think your knowledge of football is enough to be profitable in DFS, you are sorely mistaken.
Variance Manipulation
This game that we are playing thusly becomes a game of variance manipulation. We have tools that we use when constructing a roster to influence variance. Have you ever considered why we stack and correlate, or is it something you’ve simply seen done before or discussed before and are now blindly repeating? Let’s use an example to drive this one home.
Say each player we place on a roster is one “bet” we’re making. Assuming nine spots on a DFS roster and a simple 50/50 chance of getting that “bet” right, the odds of hitting what amounts to a nine-leg parlay are about 0.195% (0.5^9 x 100), or about one in 513 (we know the “bet” on a player in DFS is much lower than 50/50, but we’re simplifying this for a reason). Say, then, that you can place a “bet” on a team succeeding by combining three players from that team, making it now a 50/50 chance of getting three places on a roster right with one “bet.”. We now have placed seven “bets” instead of nine, increasing our odds to about 0.781% (0.5^7 x 100). And say this team succeeding increases the chances for their opponent remaining aggressive throughout the game, increasing the odds of a player from the other team succeeding in the process. That brings our number of “bets” down to six, increasing the odds of a roster’s success to 1.56% (0.5^6 x 100). A simple four-player game stack effectively increases our chances of being right by almost 10x. Again, this is an oversimplified example, statistically speaking, to drive this point home.
That is the reason we stack and correlate – it reduces the variance associated with DFS! Each correlated “bet” you make on a DFS roster reduces the variance by reducing the number of “bets” you must get right to win.
We can also look to historical data on hit rates for certain roster constructions and compare those to field utilization rates to influence variance. Over a five-year sample from 2020-2024, a tight end was included on a primary stack on about 22% of the winning rosters in the Milly Maker. Similarly, a running back was included on a primary stack on about 22% as well. Utilizing roster construction tendencies that carry hit rates higher than the field’s utilization rate (the field is not including a running back in their primary stacks at a 22% frequency, nor are they with tight ends) is another way of manipulating variance in our favor.
Our observations of field tendencies are paramount to these edges. Observing total combined ownership on the recent data we have, we can see that the highest percentage of broad roster construction used by the field is the skinny stack with a bring-back (a quarterback plus one of his skill position players and a skill position player from the other team). Other simple “bets” we can make to influence variance, that the field is not utilizing at appropriate rates, are skinny correlations without a quarterback (a player from each side of a game without either of their quarterbacks), team over-stacks (three players from the same team with their quarterback), and skinny stacks (a quarterback with one skill position player) without a bring-back from the other team.
Leverage and Ownership
Ownership projections are an invaluable tool in DFS, but the field generally approaches ownership the wrong way. The field generally looks to ownership to guide its decision-making process, which is theoretically a good thing under the idea of “what you win when you’re right.” But how can I say they are doing this wrong? Allow me to elaborate.
A broad term that gets thrown around the industry by people who think they are manipulating the variance in DFS is “leverage,” when the fact of the matter is these same people are simply referring to a “pivot.” A pivot is executed by removing a highly owned player and replacing him with a player of minimal ownership at the same position. Except, theoretically speaking, this is simply a one-for-one bet on one player outperforming another. You do gain a slight edge on the field if your player outperforms the highly owned one, but it is much less than you think. Let’s illustrate this with a simplified example.
Say a player’s 50 percent outcome (median, an outcome on their range of outcomes that is better than 50 percent of the other outcomes) is 15 fantasy points, and a highly owned player’s 50 percent outcome is 17 fantasy points. Mathematically, Player B will outscore Player A 88 percent of the time if we assume the range of outcomes of each player to be perfectly linear and balanced. Now include ownership. Say Player A carries two percent expected ownership while player B carries 30 percent expected ownership. How often would Player A need to outscore Player B to make this a profitable bet in the long run? 6.67 percent. Mathematically speaking, this is a profitable bet to make in the long run, given the significant disparity in ownership levels. Your true odds of this being a profitable bet are 12 percent, while the required implied odds to be profitable are only 6.67 percent. But what most people fail to consider is that we’re still talking about a player who has a relatively small 12 percent chance of outscoring the other. This influences the odds of your “bets” if we return to a previous example.
I prefer to utilize the chalk build and expected positional salary allocation to “flip the build,” which is a more sustainable approach to fantasy success. We’ll use the same working, simplified example to illustrate this process.
Say Running Back B has a salary of $7,000 on DraftKings, and we know about 30 percent of the field will be utilizing his services on a given slate. Instead of allocating the same salary to the position by choosing Running Back A and his $6,900 salary at two percent ownership, why not hunt in the $7,000 range for Wide Receiver Z that also carries minimal ownership but could have a range of outcomes that more closely aligns with Running Back B? Say Wide Receiver Z has a range of outcomes that is the exact same as Running Back B and the same ownership expectation as Running Back A. How often would Wide Receiver Z need to outscore Running Back B for this to be a profitable bet to make? 50 percent. The implied odds of being profitable are still only 6.67 percent. As you can see, this is a more profitable bet in the long run. What we just described is leverage, while the field is generally executing pivots – taking Player A at a low ownership level instead of the high ownership level of Player B.
Game Plan Development – Play the Game, Stop Letting it Play You
“But I know ball,” he screamed into the computer as his rosters narrowly missed the cash line for the fourth consecutive week. How can we optimize our rosters to increase our chances of winning the contests we’re entering?
The first thing we need to realize when we’re developing a game plan for DFS is that we don’t need the optimal roster to win. We aren’t playing against a supercomputer with an infinite sample size, we’re playing against other human beings with a defined number of maximum entries. With that in mind, we can start considering the amount of variance we are willing to accept based on three key factors: field size, the maximum allowable number of entries, and the payout structure.
As we uncovered in the above sections, we want to embrace additional variance. The bigger the field size, the larger the maximum allowable number of entries, and the more top-heavy the payout structure. Conversely, roster construction will be fundamentally different for your small field single-entry and three-max contests. But what does this look like when building rosters?
The answer is twofold and combines individual range of outcomes as far as player selection is concerned, and roster construction manipulation away from field tendencies. In smaller field single-entry and three-max contests, look for players with combined risk-reward profiles, or elevated floors that also carry ceiling. There are numerous tools available around the industry that include range-of-outcomes projections. I recommend subscribing to a platform that includes these projections as opposed to utilizing player rankings, median projections, or “feel.” On that note, median projections mean next to nothing when constructing a DFS roster. We can, and should, embrace more variance with individual player selection the bigger the field, the more entries are allowed per player, and the more top-heavy the payout structure is.
The second part of that answer deals with field tendencies. I spend a lot of time each week diagnosing what I call the “chalk build.” Picture this as the way salary comes together most easily on a given slate, given ownership expectations. I do this because it provides the blueprint for deviations. We can utilize ownership projections to piece together where the field is likeliest to allocate their salary, allowing us to stray from chalky field tendencies in the process. We can then utilize known field roster construction tendencies to return this chalk build, which is far more important than expected ownership levels on their own.
As we touched on earlier, the most common generalized roster construction utilized by the field is a quarterback, one of his pass-catchers, and a corresponding bring-back from the other team. Did you know that the winning Milly Maker roster contained this roster construction only four times in 2024? Whereas double stacks were present on eight winning rosters, a skinny stack with no bring-back was present on five winning rosters, a team over-stack (quarterback plus three skill position players) was present on one winning roster, and a game over-stack (4+ skill position players from a game) was present on one winning roster. That is a small glimpse into the boost to EV we can gain by simply straying from generalized field tendencies in DFS.
The data from last season also shows that the top five quarterbacks are in a league of their own. A quarterback scored more than 30 fantasy points 30 times last season. Of those 30 instances, Josh Allen (four), Lamar Jackson (four), Joe Burrow (four), Jalen Hurts (two), and Jayden Daniels (three) were responsible for 17 spike weeks. Jared Goff was the only other quarterback to score 30 or more fantasy points more than once last year. Also of note, there were six instances in 2024 of a quarterback finishing the week as a top-two at the position in raw points that also failed to bring a skill position player into the top six at their respective position. All six of those instances were provided by Josh Allen (three), Jalen Hurts (two), and Jayden Daniels (one), or the top mobile quarterbacks in the league. In other words, mobility at quarterback is a cheat code in fantasy and does not require a correlated bet from his team.
Also consider the inclusion of a running back and/or a tight end on a primary stack moving forward. A roughly 22% hit rate (Milly Maker win rate) for each of those positions over the previous five seasons dwarfs the utilization rates by the field. Let’s look at the data over the previous two seasons to hammer this point home. We’ll only look at quarterbacks finishing in the top two on a given week. Of the 68 instances of a quarterback finishing in the top two at their position from Week 1 to Week 17 over the previous two seasons, a running back from that quarterback’s team finished in the top six on the same week 16 times (23.5%). That number was 28 times at the tight end position (41.1%). Those hit rates are much higher than the field’s utilization of the generalized roster construction tendencies.
Finally, continue to utilize projected ownership levels to guide your decision-making process, but also include the use of true leverage instead of assuming you have it while executing a pivot. If you get nothing else from this piece, let this be the thing that sticks with you the most.
Contest selection is paramount to sustainability in DFS due to the variance at play. But beyond that, we’ve proven human beings to be inherently bad at DFS due to innate psychological conditions. We want to chase the highest payouts to get that dopamine hit, but we’re afraid of what we might lose at the same time. Those two mental blocks cannot coexist in DFS. You must be okay with what you might lose to chase the highest payouts – optimal theory demands it to embrace the appropriate levels of variance. Conversely, many overexpose themselves to variance in contests that demand higher probability outcomes. Plugging this hole will have a significant impact on your bottom line.
I personally also adopt what I call “bubble building” as part of my weekly process. I don’t look at a single outside source for information on the DFS slate until I’ve gone through my entire process for the week first, which includes analyzing every game from top to bottom. I do realize not everyone will have the time or weekly mental capacity to do this themselves, but it is supremely beneficial to my own process. This allows me to reach my own conclusions without outside influence. I then use those outside influences as a checks and balances process against my own convictions. I look at projected ownership only after both steps are complete. This helps me fight the influence of groupthink and crowd psychology.
The game of DFS does not return a one-size-fits-all solution. I can’t give you all the answers to the test each week, but I can position you better to win the most money when you’re right – because that’s what it’s all about. A single hit can change your season, your year, and your life, but you absolutely must be optimizing your rosters each week against the field size, the payout structure, and the maximum allowable entries. Stop letting this game play you.
