NFL DFS Stacking Strategy

Why NFL stacking isn't optional

A quarterback and his own pass-catcher score the same touchdown twice: the QB gets the passing yards and the TD pass, the receiver gets the reception, the yards, and the same TD. That is not two independent good outcomes you're hoping both hit — it is one event paying out twice into your lineup. Because of that, a QB-alone build without any of his pass-catchers is leaving the strongest correlation in the entire slate untouched. That's why the optimizer stacks by default rather than treating it as an optional toggle.

Shaping the stack

Stack the quarterback sets how many teammates (any mix of WR, TE, RB) ride with your QB. Leave the position breakdown at 0 and the solver picks whichever teammates project best; set WR, TE, or RB counts to force a specific shape — a WR-count of 2, for instance, forces a double-WR stack instead of letting the solver substitute in a tight end.

Bring-back adds a player from the team your QB is playing against. A shootout lifts both offenses at once, so owning a piece of the opposing team turns one good game into a bigger ceiling for your whole lineup. It's off by default for a reason: it's a per-QB conditional constraint (only applies to whichever QB you end up rostering), and turning it on measurably slows batch generation since the solver has to satisfy it lineup by lineup.

Cash games vs. GPPs

Cash: a light stack (QB + 1), no bring-back, exposure capped at 100%, no cooldown — you want your best-projected core repeated with minor variation, spreading risk rather than chasing a ceiling.

GPP: a real stack (QB + 2, one specified as a WR by default), bring-back left off by default even here, player exposure capped near 40%, stack-team exposure near 35%, and cooldowns of 1 lineup on both the stack team and the QB itself. That last one matters more in NFL than baseball: there is exactly one quarterback per lineup, so a QB cooldown is the single strongest lever you have for forcing real variety across a lineup set — without it, 20 "different" lineups can all quietly be the same QB with the depth chart shuffled underneath him.

Exposure and diversity, precisely

Max player exposure % caps how many lineups any one player can appear in. Max stack-team exposure % applies the same cap to whichever team anchors your QB stack. Stack-team cooldown and QB cooldown both stop last-K-lineups repeats — the QB cooldown is the sharper tool since it's a hard 1-per-lineup slot. Min unique players is the overall diversity floor: every lineup must differ from every other by at least this many players, independent of the stack rules.

A worked example

Say two teams are projected for a high-scoring Sunday night game. For a GPP: stack your QB with 2 teammates (say WR count forced to 1, leave the second open), turn bring-back on since this is exactly the shootout scenario it's built for, cap stack-team exposure near 35% so your 20-lineup set isn't entirely dependent on one game, and set QB cooldown to 1 so the same quarterback can't anchor back-to-back lineups. For cash: light stack (QB + 1), bring-back off, exposure left at 100%, and let the solver reuse your highest-projected plays — correlation upside isn't the goal when you only need to clear the middle of the field.

Also free: MLB lineup optimizer and its own stacking strategy — same solver, batting-order correlation instead of the quarterback throw.