NFL DFS Strategy for Beginners — Cash, GPP, Value and Bankroll

The two contest families: cash and GPP

Cash games (50/50s, double-ups, head-to-heads) pay roughly half the entrants double-ish money. The winning score is modest and predictable, so you build lineups that reliably clear it: high-floor players, no weird correlation, minimal risk. GPPs (guaranteed prize pools, tournaments) pay the top ~20% steeply and the top 0.1% life-changingly. The winning score is an outlier, so you build lineups that can produce an outlier: ceiling plays, correlation, low-owned punts. The single most common beginner mistake is bringing a cash lineup to a GPP — it finishes 40th percentile-ish, which pays nothing, over and over.

The one formula: points per $1,000

Every lineup decision compresses into value arithmetic: a player projecting P points at salary S delivers P ÷ (S ÷ 1,000) points per $1,000. The classic NFL cash-game rule of thumb says a winning lineup needs roughly 2x — about 100.0 total points on DraftKings, 120.0 on FanDuel. That makes cheap players the swing factor: a $3,300 back projecting 6.9 (2.09 pts/$1k) funds a stud elsewhere; the same price bracket projecting 1.4x quietly caps your whole build. The value calculator runs this arithmetic on any player; the optimizer runs it on all nine at once.

Correlation: why GPP lineups stack

Football scores are not independent events: a QB's big game statistically feeds his receivers' big games. Stacking — rostering a QB with his own pass-catchers — turns one long touchdown into points for two or three of your slots at once. In cash this raises variance without raising the median much (bad trade); in GPP it fattens the right tail (exactly what a top-heavy payout needs). The Stack Finder prices the standard shapes; the optimizer lets you lock a stack in and builds the best legal remainder around it.

Ownership: the invisible opponent

In GPPs you're not just picking players — you're picking players relative to how many entrants picked them. The default-projections optimal lineup is, by construction, close to what every other optimizer user found; rostering it means sharing any prize with the whole herd. Practical counter: once your build is optimal, ask of each alternative 'would the field find this?' — modest projection cost for meaningfully lower ownership is often the best trade in DFS.

Bankroll rules (the section you'll skip and shouldn't)

Standard discipline: entry fees under ~10% of bankroll per slate, single contest type while learning, and a hard stop-loss per week. The reason isn't Puritanism — it's variance math: even a genuinely good process hits losing streaks measured in weeks, and the players who survive them are the ones whose bankroll did. Track results per contest type for a season before believing any strategy is working.

A sane first month

• Week 1: play one 50/50 entry. Goal: build a legal, high-floor lineup and watch how scores actually distribute.
• Week 2: same entry, but write down your players' pts/$1k beforehand and compare after — calibrate your value intuition.
• Week 3: add a 3-max GPP with a stack you priced in the finder; feel the difference in how the lineup sweats.
• Week 4: review all three weeks honestly. If the hobby is fun at the price of a movie ticket, continue; if it isn't, cash out with your lessons.

More tools used in this guide

More guides