The goal of MME should not be to get as much exposure to as many players as possible for each slate. We find the best plays, we create our player pool using the theories we’ve already discussed, and we allocate desired ownership percentages using a lineup optimizer (or if you’re a maniac like CubsFan: building by hand). Where these types of contests differ from SE/3-Max is that we can begin to introduce more variance (we can begin SPRINKLING in high ceiling, low floor players sparsely). For example, majority of our “core plays” should be 40+% owned in our optos, meaning we are building around our solid-in-a-vacuum plays, and reducing the variance associated with playing many different groupings of players. We now have the liberty, however, of playing players that have high ceilings, but low floors, getting small exposure to the “where did that come from” plays. We arrive at these plays primarily through our research for the week, finding the teams in the best situations, and instead of taking the primary benefactor of the likeliest scenario, we take a secondary or tertiary piece that has a chance at a slate-breaking game (albeit with a lower chance of hitting). Using terminology that JM has already introduced, we’re taking the “rifle” approach as opposed to the “shotgun” approach, meaning we’re still mostly concerned with the best on-paper plays for the week, but can leverage the unknowns and variance that are associated with a “one week season” to place ourselves in the optimal scenario to take down a large contest. 99% of the time this leads us to secondary/tertiary pass-catchers in games where we think there is a “larger-than-ownership” chance of beating Vegas scoring projections. My favorite places to look to uncover these plays are by looking at DvP (defense vs. position) metrics, team net pace of play, DVOA (defense-adjusted value over average), and matchup net-adjusted line yards, which can be found at Football Outsiders and Pro Football Reference for free.
