How Many Lineups Should You Enter in a GPP?
There is no magic GPP entry count — entry volume controls variance, not expected value. The right number is the largest portfolio of genuinely distinct, individually strong lineups your bankroll can absorb, which for most players is far fewer than the 150-entry cap.
"How many lineups should I run?" is usually asked as a strategy question, but it decomposes into three separate ones: how much should I risk on this slate (bankroll), how should that risk be shaped (variance), and can I actually build that many good lineups (edge dilution)? Answering them in that order keeps the decision honest.
Bankroll first: the number that keeps you playing
GPPs are extreme-variance instruments. Even a genuinely skilled player — someone with a real long-term edge — endures long streaks where nothing binks. If a bad month can erase your bankroll, your strategy question is moot. The old discipline holds: risk a single-digit percentage of bankroll per slate, keep most of it in lower-variance formats, and treat GPP allocation as the part you can lose twenty slates running without changing your behavior. That's convention from the long-term-player community, not a law of nature — but the players still around after years follow something like it.
One entry vs. many: what actually changes
Imagine your edge produces lineups that each turn $1 of entry into $1.15 of long-run expectation. Whether you enter one lineup for $20 or twenty lineups for $1 each, the expectation is the same $23. What changes is the distribution around it:
- One entry is maximum concentration — your night is decided by a single lineup's draw from its distribution.
- Twenty distinct entries sample twenty different regions of the outcome space. The portfolio's bad nights are less bad, and the chance that at least one lineup lands a tail outcome rises substantially.
The catch is the word distinct. Twenty near-clones of the same build aren't twenty samples — they're one sample wearing twenty hats. If the shared game script dies at 7:05pm, the whole portfolio dies with it.
According to DFS Simulator's portfolio analysis, the practical test of lineup distinctness isn't how many players two lineups share — it's whether their simulated wins arrive in the same iterations. Two lineups can differ by six players and still win in identical universes; the sim sees through the cosmetic difference.
Edge dilution: the argument against volume
Every additional lineup you build is, by construction, a lineup you liked less than the previous one. Somewhere in the sequence — entry 8, entry 40, entry 120 — your marginal lineup stops carrying an edge and starts being filler that pays rake for the privilege. The right entry count for you is where that line sits, and it moves with your tooling: constraint-driven generation with stacking rules and exposure caps pushes the dilution point much further out than hand-building does.
A practical framework by bankroll posture
| Profile | Reasonable GPP posture | Why |
|---|---|---|
| Recreational, small bankroll | 1–3 entries, single-entry contests preferred | Single-entry fields cap the tooling disadvantage against max-entry professionals |
| Regular player, defined bankroll | 5–20 distinct entries across 2–3 contests | Enough portfolio breadth to diversify scripts without outrunning lineup quality |
| High-volume grinder | 20–150 entries, portfolio-simulated, exposure-capped | Variance smoothing only works if distinctness is enforced by tooling, not vibes |
Framework, not prescription: contest selection, slate size, and your actual demonstrated edge move these bands. The constant is that entry count should be a decision you make before the slate, not an impulse at lock.
Where simulation fits
Portfolio thinking is exactly the workload simulation is built for. Run the slate, generate candidates under your constraints, then read the portfolio-level output: combined exposure to each player, how correlated your entries' winning universes are, and whether your 150th lineup is actually adding new outcome coverage or just rake. The iteration budget matters here too — portfolio tail analysis is a tail statistic, and tails need samples. Every plan on DFS Simulator caps runs at 150 lineups, matching the standard max-entry contest limit.