The dynamic game of Daily Fantasy Sports (DFS) requires much more than simply knowing the sport for which we’re entering contests to be successful. We must be adaptable, precise, and open to learning from previous endeavors, the latter of which will be the primary focus of this weekly written piece. Game Theoretic methodologies will allow us to analyze and dissect the previous week’s winner of the largest and most prestigious Guaranteed Prize Pool (GPP) tournament on DraftKings – the Millionaire Maker. These same tenets of Game Theory, which can most simply be explained as the development of decision-making processes given our own skill and knowledge, assumptions of the field based on the cumulative skill and knowledge of others playing the same game, and the rules and structure of the game itself, will allow us to further train our minds to see beyond the antiquated techniques of roster building being employed by a large portion of the field. Approaching improvement through these methods will give us insight into the anatomy of successful rosters and will help us develop repeatably profitable habit patterns for the coming weeks. We’ll start by looking at the previous week’s winning roster, extract any pertinent lessons for future utilization, and finish with a look ahead towards the coming main slate.
Winning Roster
Lessons Learned
Primary Stack
User abedollars applied a primary stack of Lamar Jackson and Mark Andrews, most notably with no correlated bring-back. The strongly-differentiated stack scored 72.32 combined DraftKings points, good for 34.29% of the total points on the roster – from just two spots! The powerful lesson here has to do with the decision to not utilize a “primary DFS stack” of a quarterback plus one of his pass-catchers plus a correlated bring-back. As we have discussed in previous weeks, our decision-making processes must also take the fields’ decision-making processes into account to generate leverage (put simply, doing things differently from the field so we aren’t fighting against so much of the field in our path to first place).
Looking at the data from last year’s Millionaire Maker winners, 27.8% of regular season winners (Weeks 1-18) utilized a quarterback plus one pass-catcher and no correlated bring-back as their primary stack, the highest occurrence of a single roster construction amongst the 18 winners. Now consider the fact that the “standard DFS stack” of a quarterback plus one pass-catcher plus a correlated bring-back is the most common roster construction starting point in Guaranteed Prize Pool (GPP) tournaments each week, and we’re left with a beautiful piece of leverage starting us in the face.
Let us also take a quick look at the theory of roster construction, which is a pertinent discussion point here. The reason we stack is to increase the variance on individual rosters while simultaneously reducing the number of variables that must “go right” to be successful. That said, the field is stuck in parrot mode, seeing that stacking helps take down tournaments and blindly following those processes. By understanding why we stack, we can make leveraged tweaks to those methodologies to generate additional expected value (EV) away from the most commonly used practices of the field – this learning point falls under that category.
Roster Construction
I’ll preface this discussion with the assumption that abedollars likely doesn’t care what I say about his roster because he has a million dollars. This isn’t a personal attack. This is learning. From a roster construction standpoint, the winning roster has a lot of individual pieces that are thrown together in a seemingly random fashion. That is not a repeatably profitable habit pattern that can remain consistent over time. Let’s break it down through a theoretical example.
Think about a roster as a path we must walk towards first place, with each roster spot a decision node along that path. Assuming a slate with 13 games, that means the quarterback decision node has 26 possible paths leading from it (each team playing in a game has one quarterback). Then we would come to the running back decision node, which would theoretically have more than 26 possible paths leading from it, but we’ll use 26 to keep the math relatively simple. Then we would come to the wide receiver position, which would theoretically have 78 paths leading from it (three starting wide receivers times 26 teams). This would continue until we reach the end, having walked the correct path at each decision point.
That would mathematically mean our chances of guessing correctly at each decision point to find the optimal path would be (1/26)*(1/26)*(1/25)*(1/78)*(1/77)*(1/76)*(1/26)*(1/99)*(1/26).
We can say whatever we want about our individual skill at DFS or ability to pick the right players (paths), but those odds are akin to playing the lottery. Correlating, stacking, and targeting game environments is a way to reduce the number of decision points encountered along our roster’s journey, as we can theoretically capture two (or three, or four, or more) high-end outcomes through one decision (or decision node). That would reduce the number of decision nodes dramatically, thusly reducing the variables and increasing our odds of winning. The assertion for the utilization of these methodologies is strengthened through the realization that we don’t need optimal to win; we simply need to beat other imperfect human beings.
Correlated Pairings
User abedollars did well in identifying low-owned upside with DeVonta Smith, Mack Hollins, and the primary stack of Lamar Jackson and Mark Andrews. The good here is realizing a rushing quarterback’s true ceiling is unlocked through a pass-catcher, as was the case with both Lamar Jackson and Jalen Hurts this week. This roster could have been improved by thinking through the correlated pairings associated with the other two upside hits – Mack Hollins and DeVonta Smith.
Mack Hollins’ upside was theoretically unlocked through the Titans leading the game throughout, which correlates well to increased Derrick Henry usage. We know what Henry can do with increased usage. He ended up scoring 25.3 fantasy points on 26 running back opportunities.
DeVonta Smith succeeding likely came from one of three game environments – a shootout, a Philadelphia beat down, or a Washington beat down. The former two environments were the most likely based on each team, meaning Smith’s production was likeliest to directly correlate to increased pass attempts from Carson Wentz and the Washington Commanders. Washington alpha wide receiver Terry McLaurin put up a 6/102 line on nine targets, which was good for 19.2 DraftKings points.
Looking Ahead
DK Metcalf (or Tyler Lockett) + Josh Reynolds (Correlated Pairing)
Somewhat surprisingly, the Seahawks/Lions game opened with the second-highest game total on the slate with a massive 50.0 line, likely drawing heavy influence from the Lions and their pace and poor defense.
D’Andre Swift has been playing through an ankle injury, seeing only 10 and 11 running back opportunities over the previous two weeks. Amon-Ra St. Brown rolled his ankle in the second quarter of Week 3, showing a lack of explosiveness for the rest of the game. Enter Josh Reynolds, who has played only nine fewer snaps than St. Brown and is priced at only $4,600 on DraftKings this week.
Seattle surprisingly ranks in the middle of the pack in pass attempts per game at 34.3, tied with Tampa Bay, Miami, and Pittsburgh. Pete Carroll talked up the necessity to get Metcalf more involved, which he made good on in Week 3. Metcalf saw 12 targets and found the end zone against a stronger Atlanta defense than he’ll face this week in Detroit. Lockett has seen 11 targets in back-to-back games. The concentration of the Seattle pass offense means we can capture increased upside at likely low ownership in game environments where expect them to be passing more.
Josh Allen + Stefon Diggs + Mark Andrews (Primary Stack)
Not to directly contradict what we explored above, but this situation is a bit different. The Bills are coming off a loss, Diggs currently has a nice, big, red “Q” next to his name (which should serve to lower his ownership), and the Ravens are a concentrated, yet low volume, pass offense, meaning we would want to target their primary pieces in game environments where we can expect a boost to volume. All of that comes together to provide a situation where the upside from the game environment directly correlates to increased upside for these three players, specifically.
CeeDee Lamb + Terry McLaurin + Curtis Samuel (Roster Construction)
Lamb has seen 11 targets in each of his first two games this year (has not played in Week 3 yet) and backup quarterback Cooper Rush has targeted him heavily when he has filled in for injured starter Dak Prescott. Carson Wentz and the Commanders check in fourth in pass attempts per game and hold a top-five pass rate over expectation value, indicating it’s likely we see the high pass rates continue. Pairing the schemed usage of Samuel with the per-target upside of McLaurin provides a solid cost-considered range of outcomes, while Lamb can be added to further leverage the upside brought to the table through the game environment. Feel free to add Carson Wentz coming off a disappointing game against the dynamic defense of the Eagles.
