3,944,085 lineups crunched | A user crunched an AFL lineup for DraftStars 31 minutes ago

Superbowl DFS Strategy: Dissecting a one-game contest

Posted {{post.formattedDate}} by Darryl Woodford

We don’t normally cover the NFL here at Fantasy Insider - particulary given there’s so much US-based coverage that applies given the similarity in scoring systems, but we’ll make an exception for the Superbowl, as it also gives us an interesting opporunity to look at how to enter one game contests.


A one-game contest, in DFS terms, can really only have three outcomes: Team A blowout, close game, or Team B blowout. For NFL, with an average total of 47, a rule of thumb is that a 2 TD win is a ‘blowout’ sufficient to impact on scoring (28-14; 31-17 etc). Which team wins a close contest means little to us, as it gives players on both teams - and, depending on the sport, both sides of the ball - a chance to score.

So the first thing I do in a one-game contest is try and put %s on the likelihood of each of those outcomes, and here bookmakers come in handy. At the time of writing, the superbowl markets at Betfair put the chance of Carolina winning at 1.49 (66%), and winning by more than 2 TD at around 3.3, or 30%. The odds of Denver winning at all are 3.0 or 33%, so we can roughly guess the chance of a Denver blowout is around 10%. Putting these together, we get:

Carolina Blowout: 30%
Denver Blowout: 10%
Close Game: 60%

We don’t have to be exact here, we’re just trying to get a feel for the game. Those numbers then dictate what a multi-lineup strategy should look like; i.e. if I enter 10 lineups, I’ll try to have 1 that assumes a Denver blowout, 3 a Carolina blowout, and 6 a close game. It may not always work like that, and in a big tournament I might enter a couple more that assume a Denver blowout, on the theory that most other people won’t be doing the same..


The next thing to do is to predict each player’s performance. For the NFL, like the NBA, there’s a whole heap of American sources for this; personally I like http://www.footballguys.com and http://www.4for4.com. For Aussie sports, you can do this yourself, use the basic averages, or (soon) subscribe to our weekly predictions. For today’s purposes though, I’ve used a mix of the projections from the past two sites. Here’s the key info:

That’s enough to calculate the ‘optimal’ lineup; here’s what it would look like on Moneyball:

We can pretty much put that lineup in the ‘close game’ category. While it is predominantly Panthers players, the Denver defense isn’t going to fare well in a blowout, and in fact for a Panthers blowout you’d probably go with a few more Broncos receivers (and not a RB), on the theory teams are forced to throw when they’re behind.

Using the projections and our desired lineup distribution, the next job is to go through the players and build those lineups; using a teams defence, kicker and RB’s in blowout lineups, and varying the players to give yourself a maximum chance of hitting the lower owned players who might have a great game (e.g. Brown & Funchess) in the above lineups.


The twist in the tale is the captain at Top8. You want this to be a high scoring player, and in a future article I’ll talk about using Monte Carlo simulations to estimate the percentage of the time each player is highest scoring. For the superbowl, in most lineups I’ll likely have Cam Newton, Jonathan Stewart or Greg Olsen as captain. In lineups favoring Denver, C.J. Anderson will feature prominantly, with Manning, Thomas and Sanders also warranting consideration. Generally you want a stable player (QB/RB) in case things go as expected, but also a couple of lower chances (WR/TE) who if they hit form will pretty much guarantee you a place in the money.

Whoever you pick this Sunday/Monday, good luck!

Photo credit: Rajiv Patel