100 Fantasy Football Tips in 100 Days, Day 98: How Many Lineups to Play
At DraftKings, I discussed why I believe it’s acceptable to play multiple lineups in cash games:
There’s not too much discussion about the merits of multi-entry in DraftKings tournaments—most people agree that you can enter a GPP a bunch of times and still maintain a positive expected value—but there’s a fierce debate over how many lineups you should construct for cash games (head-to-heads, 50/50s, and three-man leagues).
Popular wisdom suggests that daily fantasy players should enter just a single lineup into cash games—their “optimal” lineup with all of their top values, or as many of the top values as they can fit under the salary cap. I disagree with the idea that there is one true “optimal” lineup, as well as the notion that playing just a single lineup in cash games is a smart strategy.
The Problem with Counting Lineups
I first want to address an important related idea, which is that what determines risk is the number of lineups you play on DraftKings. Instead, risk depends on the number of players you use, the way you combine those players, and the amount of money you have on each player/combination.
Player A might use two different lineups that have completely different players from each other—18 different players in an NFL contest—while Player B might have two lineups with every player the same except one (10 different players across the two lineups). Who is taking on the greater amount of risk? Clearly Player B.
When you frame the issue in this way, it becomes clear that what matters shouldn’t be the total number of lineups, but the total player exposure. How much exposure do I have to this player and how much money do I have on him? In the case of Player B, the total risk and player exposure would be almost exactly the same as with someone who used just a single lineup, while Player A’s diversification would make his risk/reward much different.
The key word there is ‘diversification’: the risk you assume has little to do with how many lineups you play and more to do with how much you diversify those lineups, i.e. how many players you use and what type of exposure you have to them. In a world with perfect knowledge, we should indeed use a single cash lineup optimized around our values. There’s indeed only one true set of players that will optimize a lineup’s projected points.
The problem is that we don’t have perfect knowledge of outcomes. We’re fallible, our projections are fallible, and our player values are fallible. One of the biggest mistakes a daily fantasy owner can make is not being aware of their own fallibility.
We see this all the time in the NFL when teams give up a king’s ransom to trade up in the draft to snag “their guy.” Many times, the only way those moves work out for the team trading up is if they get exactly what they think they’re getting. They fail to realize that the draft is a game of probabilities, and no one has perfect knowledge of how future events will unfold. They don’t say to themselves, “This is what we think and we’re pretty confident we’re right, but what happens if we aren’t?”
When we assume that our projections and values are flawless, we create an extremely fragile system for creating lineups. Every week, there are numerous players all ranked near one another in terms of value. To use one player because he’s projected to score 0.1 points more than another requires a level of confidence in your predictions that is simply unattainable.
Once we start to account for our own fallibility—once we acknowledge the fact that, hey, we could just be wrong—it opens up the door for greater diversification.