Can I use the same lineup for cash games and GPPs?

You can, but you're leaking money in one of them. In our simulations, ranking 50,000 lineups by cash rate and by tournament ROI produced top-100 lists that shared zero lineups. The builds that cash most reliably and the builds that win tournaments are entirely different lineups.

Why doesn't the highest-projected lineup win GPPs?

Because projection is a median-style estimate and tournaments pay extremes. In a 50,000-simulation run, the highest-projected lineup on the slate cashed 48.8% of the time but never finished first once — its safe, high-floor construction had no path to the ceiling scores that top a 50,000-entry leaderboard.

What should I do differently in cash games?

In cash you're paid for beating the median, so ownership is irrelevant and floor is king: high-floor players, no forced stacks, no leverage plays. Everything that makes a GPP lineup good — correlation, ceiling-chasing, being different from the crowd — adds variance you aren't paid to carry in a 50/50.

Is variance good or bad in DFS?

It depends entirely on the payout curve. In cash games (flat, half the field doubles up) variance is a cost. In top-heavy GPPs variance is the asset: a lineup that folds most weeks but owns the weeks everything hits together outperforms a steady lineup that never spikes. Same slate, opposite incentives.

GPP vs Cash Strategy: Why They're Different Games (50,000-Sim Proof)

Two payout curves, two opposite questions

A cash game (50/50s, double-ups) pays roughly half the field the same modest prize. The only question that matters is: does this lineup beat the median score more often than not? Nothing above the cash line earns an extra cent, so ceiling is literally worthless there.

A GPP pays a top-heavy curve — the top 1% takes most of the pool, first place alone can take a fifth of it. The question inverts: on the small number of Sundays this lineup spikes, is it standing above everyone? Median weeks are worthless there; you're betting entirely on the extremes.

Same slate, same players, same salary cap — and the two formats price every decision in opposite directions. Here's how big that gap actually is.

The receipt

50,000 sims · 50,000-entry $20 GPP · real slate

Ranking all 50,000 field lineups by cash rate and by tournament ROI: the two top-100 lists share zero lineups. The field's best cash lineup (48.8% cash rate) ranked around 15,000th as a tournament play. The best GPP lineup (+572% ROI) cashed less often than the chalk — 38.5% — and finished first in just 0.03% of sims. Those ~15 first-place Sundays out of 50,000 were its entire edge.

If you enter the same builds in your double-ups and your tournaments — and most casual players do — one of those entries is mispriced every single week. The ten most reliable cash lineups on the slate ranked as deep as 15,253rd of 50,000 by tournament ROI — the safest builds weren't just absent from the GPP elite, some sat in the bottom two-thirds of the field.

The lineup that proves it

The cleanest illustration is the "optimal" lineup — the single highest-projected legal roster on the slate, the one every optimizer spits out first. Ours projected 144.7 against a field median around 126. Across 50,000 simulated Sundays it cashed 48.8% of the time — it was literally the best cash lineup in the entire 50,000-entry field. Its first-place finishes: zero. Not one in fifty thousand. Its GPP ROI ranked 15,025th out of 50,000.

Nothing was wrong with the projections. The lineup was built from safe, high-floor scoring concentrated in chalk — a construction with no live path to the 200-point ceiling games that top a 50,000-entry leaderboard. Projection measures the average Sunday; tournaments only pay the extreme ones.

What actually changes between formats

Variance flips sign. In cash, variance is a cost — you're paid for reliability. In GPPs, variance is the asset you're buying. A player's right-skewed tail (the boom games) is exactly what wins tournaments, and it never shows up in the projection column.
Correlation flips too. Stacking a QB with his receivers adds no expected points — it synchronizes them. Worthless in cash, decisive in GPPs: matched for projection, our simulated QB+2 stacks finished first 1.9× as often as naked-QB builds (+39% vs −13% ROI).
Ownership only exists in GPPs. In cash, who else has your players is irrelevant. In tournaments it's decisive: among equally-projected lineups in our run's top projection tier, the low-owned third earned +69% ROI while the high-owned third earned −29% — a 98-point swing explained by nothing but the crowd (shared upside, plus duplicated prizes).
The floor/ceiling tradeoff inverts. Cash wants the 90th-percentile floor; GPPs want the 95th-percentile ceiling, and will happily accept ugly median weeks to get it.

Practical translation

Cash: high-floor players, no forced stacks, ignore ownership, take the boring points. The "optimal" projection lineup is genuinely close to correct here.

GPP: build for the week everything lands at once — correlated stacks, live ceilings, exposure the field underprices — and accept that most weeks fold. The math of the payout curve does the rest.

The hard part is that GPP quality is invisible on paper: two lineups with the same projection can differ by nearly 100 points of ROI. That's the entire reason SimSlate exists — it plays out the actual tournament (field, correlation, duplication, payout curve) and hands you the ROI ranking directly. The free plan runs a real 1,000-lineup field simulation, no card required; the methodology is public.

Numbers above are from a single 50,000-simulation run on a real last-season main slate (50,000-entry, $20, top-heavy payout). Different slates and seeds move the exact figures; the structure — near-zero overlap, leverage gap, stack multiplier — holds run after run.

GPP vs cash: we ranked 50,000 lineups both ways. Not one made both lists.