The lineup gets ninety percent of your attention and it's the smaller edge. Where you drop that lineup β the contest, the field size, the payout curve, the entry limit β is the bigger, more durable one, and it's a decision you make before a single projection matters. This is the full playbook: how contests actually pay, how to match one to your real edge and your bankroll, and an exact contest menu for every roll from $50 to $5,000 β NFL and MLB included.
Article by James at CourtEdge · CourtEdge DFS Strategy Series · proprietary CourtEdge framework · π EVERGREEN GUIDE · ~30-MIN READ · UPDATED FOR THE MULTI-SPORT GRIND
Everything below the fold: rake & payout curves, cash vs. GPP, SE β MME field dynamics, the duplication problem, overlay & +EV, the bankroll math (just enough), the decision framework, and a bankroll-tiered contest menu for NFL & MLB.
See live picks & track record at courtedge.net/picks
You cannot choose a contest intelligently until you can read its payout structure at a glance. Three numbers tell you almost everything: the rake, the shape of the curve, and where the money stops. Learn to state those three before you enter anything and you're already ahead of most of the field.
Every contest skims a percentage off the top before a dollar is paid out. If 11,000 people enter an $11 contest, that's $121,000 in entries β but the prize pool is not $121,000. After the operator's cut (the rake, typically 10β15%), the pool might be ~$103,000. That missing ~$18,000 is the house's fee, and it is the single most important number almost nobody checks.
Rake matters more than any lineup decision because it is certain and it compounds. A player grinding 15% rake has to be dramatically better than the field just to break even; the same player grinding well-structured 8β10% contests keeps far more of their edge. Over a season, the gap between the high-rake and low-rake versions of the "same" contest can be the entire difference between a green year and a red one.
Prefer the lower-rake version of any contest. Bigger-buy-in contests usually carry lower rake percentages than micro-stakes ones β which is why moving up in stakes (within your bankroll) can quietly improve your ROI even as the competition sharpens.
Two contests with the same buy-in and field can pay in opposite shapes:
These shapes demand opposite lineups and opposite mindsets. In a flat contest you want floor and safety β you're clearing a line, not soaring. In a top-heavy contest you want ceiling and uniqueness β min-cashing is nearly worthless, so you build to finish first. Playing the wrong shape for your lineup style, or your bankroll, is one of the most common and expensive mistakes in DFS.
Every contest has a line where the money stops. In a double-up it's ~44% of the field; in a top-heavy GPP the paid line might be ~20β25%, but with most of the prize concentrated far above it. In cash: you don't need to finish first, you just need to beat the first loser β the highest-scoring lineup that didn't cash. In a tournament the opposite holds: beating the first loser by a point earns a rounding error, so you build to beat the field's best lineup, not its bubble.
(1) Rake % β what the house keeps. (2) Curve shape β flat cliff or top-heavy ski jump. (3) The paid line β where the money stops. If you can state those three before you enter, you understand the contest. If you can't, you're guessing.
"Cash" and "GPP" get lumped together as "DFS," but they are as different as index-fund investing and buying lottery tickets. Confusing them is the fastest way to blow a bankroll.
Cash is the equivalent of grinding small, consistent edges. The goal isn't to make a ton of money β it's to make a little, reliably, and let the edge compound. You're trying to beat roughly the top-half line, so you build the highest-floor lineup you can: safe, high-median players, minimal dart-throws, ownership be damned. Being "different" is worthless here; being right and safe is everything. A good cash player wins a modest percentage above break-even and does it week after week.
A GPP is a leverage instrument. The prize is enormous, the odds are long, and the correct build is one that can finish first in a huge field β which means ceiling, correlation (stacking), and uniqueness. The larger the field and the more top-heavy the payout, the harder it is to win and the more you can win. You will lose most of your GPP entries. That is not a bug; it's the entire design. GPPs pay for the rare night your ceiling build hits and almost nobody else has it.
The most common leak is building a cash lineup (safe chalk) and firing it into a tournament, or a tournament lineup (boom-or-bust darts) into a cash game. Pick the game first. Then build for that game. Never let a lineup wander into the wrong room.
Most winning grinders put the majority of their volume in cash (to grow the roll steadily and survive variance) and a smaller slice in GPPs (for the top-heavy upside). A common starting split is roughly 70β80% cash / 20β30% GPP, tightening toward cash when your roll is small and loosening toward GPPs as it grows. We'll make this concrete in the bankroll section.
Two tournaments with the same buy-in can be worlds apart depending on how many entries each player is allowed. This is the single most under-appreciated dial in contest selection, and getting it right is a real edge over the recreational field.
| Format | What it is | Who it favors | Best when⦠|
|---|---|---|---|
| Single-Entry (SE) | One lineup per person, period | The disciplined solo player β no one can out-volume you | You have a real read but limited bankroll/time; you want a level field |
| 3-Max | Up to 3 lineups each | Players who can build a small, sharp, diversified set | You want a little coverage without a mass-entry arms race |
| 20-Max | Up to 20 each | Mid-volume players with a build process | You can produce 10β20 genuinely different lineups around a thesis |
| 150-Max | Up to 150 each (the classic "milly" cap) | Sharp mass-entrants with tooling | You have a simulator/optimizer and can attack duplication at scale |
| MME (large-max) | Hundreds to 1,000+ each | Pros with pipelines, exposure rules & bankroll to match | Rarely the rec player's friend β you're funding the sharks here |
In an SE contest, the pro across the table is limited to the same one bullet you have. They cannot bury you under 150 lineups. Your one well-constructed lineup competes on equal footing with everyone else's one lineup. For a player with a genuine read but a modest roll, SE tournaments are frequently the single best +EV room on the board β you get top-heavy upside without the mass-entry arms race. This is exactly why the massive single-entry NFL major (see the NFL playbook below) is such a prized target: a huge prize pool, one bullet each, no volume advantage for anyone.
In a 150-max or MME field, a well-tooled pro enters a diversified fleet of lineups engineered for uniqueness and exposure. Their portfolio covers outcomes your single lineup can't. You can still win β someone with one great lineup does β but you're structurally disadvantaged, and the rake is skimming you the whole time. Entering big-max fields without the tooling and bankroll to play the volume game is one of the most common ways good players quietly bleed out.
Small roll or limited time β lean single-entry and small-field. Growing roll with a build process β add 3-max and 20-max. Only step into 150-max / MME when you have a simulator, exposure discipline, and a bankroll that can absorb the variance of firing many bullets.
Here's the counter-intuitive engine of large-field tournaments: your prize is split among everyone who plays your exact lineup. If you and 40 other people submit an identical lineup and it wins, you don't win the top prize β you split it 41 ways. In a top-heavy contest, that turns a life-changing score into a nice one.
This is why uniqueness is a real, dollar-denominated edge in big fields, not a stylistic preference. The most-obvious build β all chalk, the popular stack, the "correct" lineup everyone's projection tools spit out β is also the most-duplicated. When it hits, you share it. When you leverage off the chalk with a lower-owned pivot that still has a real path to ceiling, you're buying uniqueness: the same win, split fewer ways.
In a massive-field GPP, deliberately leaving $200β$1,400 of cap unused is a legitimate uniqueness lever β most of the field spends to the dollar, so an off-the-grid salary total lands you a less-duplicated lineup. It only makes sense in top-heavy, large-max contests; never do it in cash.
Overlay is the closest thing to free money in DFS. A guaranteed prize pool (a "GPP" is a Guaranteed Prize Pool) pays out its advertised total whether or not it fills. When a contest guarantees $100,000 but only draws $85,000 in entries, the operator covers the $15,000 gap. That $15,000 is injected into the pool for free β spread across the players who did enter. Overlay instantly raises everyone's expected value, sometimes turning a slightly βEV contest sharply +EV.
A contest is +EV when your expected score, against that specific field and payout curve, returns more than your entry after rake β and overlay is a thumb pressing the scale in your favor. Learn to spot under-filled guarantees near lock and you're collecting edge that has nothing to do with your lineup.
Every contest choice should scale to your honest, evidenced edge on that specific slate β not your hope. Some nights you have a genuine read (a clear game environment, a pricing mistake, a data advantage). Some nights you're guessing. The mistake is playing the same contest menu regardless.
Your edge also has to be real. Be honest about time, tools, and data. If you can't build differentiated lineups, don't enter 150-max fields against people who can. If you don't have a projection edge, lean on contest edges β overlay, single-entry discipline, low rake β which don't require you to out-project anyone.
This is the "with your bankroll" half of the title, and it's the guardrail that keeps the rest of this guide from blowing up on you. You don't need a finance degree β you need a few rules that keep you in the game long enough for your edge to pay off.
Define one unit as a fixed fraction of your total DFS bankroll and size everything in units. A common, survivable rule: risk no more than ~2β3% of your bankroll on a single slate, and keep any one tournament entry to a small fraction of that. On a $500 roll, that's ~$10β15 in play per slate β not one $50 milly-maker bullet. Sizing in units means a cold streak shrinks your bets automatically and a heater grows them; you can't go broke fast if every bet is a small slice of the whole.
"Risk of ruin" is the probability you lose your whole bankroll before your edge compounds. It rises sharply with two things: bigger bets relative to your roll, and higher-variance contests. Here's the part players miss β you can have a real edge and still go broke if your bets are too big or your contests too top-heavy for your roll. A profitable GPP player betting 20% of their roll per night will bust with uncomfortable frequency simply from variance. The fix is never "get better picks." It's bet smaller and lean flatter until the roll can absorb the swings.
You don't need the full Kelly formula. The usable idea: the size of your bet should track the size of your edge. Big, well-evidenced edge β size up (toward the top of your unit range). Thin or speculative edge β size down. Most players do the opposite β they bet big to chase losses on nights they have no read. Inverting that instinct is worth more than any single lineup.
| Bankroll temperature | Contest lean | Why |
|---|---|---|
| Small / building | Mostly cash + single-entry; tiny GPP dabble | Survive variance, grow steadily, learn the ropes cheaply |
| Healthy / stable | 70/30 cash-to-GPP; add 3-max & 20-max | Compound the cash edge, take real top-heavy shots |
| Large / pro | Wider GPP allocation, selective MME with rules | Roll can absorb swings; volume + uniqueness game unlocks |
Never let one slate risk more than ~2β3% of your roll, size your bets to your edge, and lean flatter (more cash, smaller GPPs) the smaller your bankroll is. Do that and no cold streak can end you β which is the only way to still be here when your edge pays.
Put it all together into a repeatable sequence you run before you build anything:
Run that seven-step loop every slate and you've institutionalized the single biggest edge available to a non-professional. The build is the last step, not the first.
This framework scales with you. The rules don't change as you improve β the dials do. Find yourself below and know which levers to turn next.
Live almost entirely in cash and single-entry at micro stakes ($1β$3). Your goal isn't profit yet β it's reps and floor. Build safe, clear the bubble, and read your ROI by contest type. Skip every 150-max milly; they'll teach you nothing but variance. One good habit at this stage β always pick the game before the lineup β is worth more than any pick.
Now the small-max lanes open up: 3-max and 20-max GPPs and the single-entry major. You can build a few genuinely different lineups around a thesis and press a real edge without an arms race. Start tracking RAROI, not just ROI, and let it steer you toward the rooms that pay you steadily. Cash is still the backbone; tournaments are the upside on top.
Selective 150-max and MME unlock, because you can attack duplication with a simulator, run exposure rules across a fleet, and absorb the variance. You hunt overlay deliberately, move up in stakes for lower rake, and audit every room by RAROI monthly β pruning the weak, feeding the strong. Even here, the shape holds: cash-heavy, single-entry-favored, millies a controlled slice. The sharks didn't abandon the fundamentals; they just turned every dial to its optimal setting.
NFL is a once-a-week decision, which makes contest selection more important, not less β you get ~17 quality swings a season, so wasting entries in the wrong rooms is expensive. Here's how the main NFL contest types map to the framework.
The DraftKings Fantasy Football Millionaire and its cousins are the biggest, most top-heavy, most-duplicated rooms in the sport β huge guarantees, a millionaire up top, 150 entries per player. They are a pro's mass-entry battlefield. For a rec player, one or a handful of bullets here is a lottery ticket, not a strategy: play it for fun with a tiny slice of your roll, lean hard into uniqueness (contrarian stacks, a low-owned bring-back, leave a little salary), and never mistake it for +EV grinding. If you can't build genuinely differentiated lineups at volume, most of your NFL bankroll should be elsewhere.
This is the sweet spot for a sharp player without a mass-entry pipeline: a huge single-entry NFL tournament β enormous prize pool, top-heavy, but one bullet each. No one can out-volume you. Your one well-built, appropriately-leveraged lineup competes dead-even with every pro in the room. If you have a genuine weekly read, the big single-entry major is often the best top-heavy value on the NFL board β you get milly-maker upside without funding the volume sharks. Build it like a tournament (ceiling + a leverage angle), but you can play a touch closer to the "correct" build than in a 150-max because duplication is far lower.
The DraftKings Power Sweep is a 3-entry-max GPP β a middle lane between single-entry and the mass-entry milly. Three bullets let you express a small, diversified set (say, three different game-stacks you like) without an arms race against 150-max fleets. Small-max GPPs (3-max, 20-max) are where a mid-volume player with a build process can press a real edge: top-heavy upside, but a field that's far less saturated with pro fleets than the milly.
Double-ups and 50/50s are the steady core. Build the highest-floor lineup you can β bell-cow backs, high-target receivers, a safe game environment β and clear the ~top-half line. This is where a disciplined NFL player compounds a roll across a season; the tournaments are the upside on top.
| NFL contest | Format | Play it when⦠| Build |
|---|---|---|---|
| Cash (DU / 50-50 / H2H) | Flat | Always β the core | Highest floor, safe, ownership-blind |
| Single-Entry Major | SE, top-heavy | You have a weekly read; want fair top-heavy upside | Ceiling + one leverage angle; near-correct OK |
| Power Sweep / small-max | 3-max, top-heavy | You can build a few diversified sharp lineups | Ceiling + uniqueness across your 3 |
| Milly Maker | 150-max, very top-heavy | You mass-enter with tooling β or want a lotto dart | Max uniqueness; contrarian stacks; tiny roll slice |
MLB is a daily sport, which flips the math: you get a slate almost every day, so discipline and rake control matter more than any single night. The grind rewards steady, correctly-sized play and punishes the urge to fire the biggest tournament every night. Baseball is also the highest-variance major sport for DFS (one swing changes everything), so bankroll discipline is doubly important.
Because you play so often, a small, repeatable cash edge compounds fast. Build high-floor: a safe ace with strikeout upside, hitters in strong park/pitcher environments, and avoid the volatility darts. Double-ups and 50/50s on the main slate are where a disciplined MLB player quietly grows a roll over a month.
As in NFL, MLB single-entry tournaments level the field β one bullet each, no fleet advantage. A single-entry MLB GPP with a real prize pool is often the best top-heavy value for a grinder: you're building one sharp, stacked, appropriately-leveraged lineup against a field that can't out-volume you. If you're going to take a tournament shot most nights, the SE GPP is usually the highest-EV place to take it.
MLB tournaments are won by team stacks (correlated hitters who score together). In a big-field 150-max milly, the popular stack is heavily duplicated β leverage into a lower-owned stack with a real ceiling. In single-entry and small-field, you can play the sharper "correct" stack because you won't be sharing it. Always let the field size decide how contrarian your stack needs to be.
| MLB contest | Format | Play it when⦠| Build |
|---|---|---|---|
| Cash (DU / 50-50) | Flat | Daily core, any slate | High floor; safe arm; strong-environment bats |
| Single-Entry GPP | SE, top-heavy | You want a fair tournament shot most nights | One sharp stack + leverage; near-correct OK |
| Small-field / 3-max GPP | Low-max | Mid-volume, a couple of stacks you like | 2β3 diversified stacks; moderate uniqueness |
| 150-max Milly | Very top-heavy | You mass-enter with a sim β or lotto dart | Low-owned stack; contrarian; small roll slice |
Whatever MLB room you pick, put the lineups through a contest simulator that models that field and payout curve, and rank by ROI and RAROI. The right build for a single-entry double-up is not the right build for a 150-max milly β the sim is how you prove your lineups fit the contest, not just the slate.
The framework travels. Only the texture changes:
This is where it gets concrete. Here's a sane, survivable contest menu at three bankroll temperatures. Adjust the dollars to your own roll; the shape is what matters. (Every number below assumes the ~2β3%-per-slate guardrail.)
At every roll, cash is the backbone, single-entry is the rec player's best tournament value, and the mass-entry millies get a small, deliberate slice β never the bulk. The bigger the roll, the wider the GPP allocation, but the shape β cash-heavy, SE-favored, milly-light β holds all the way up.
You can't fix what you don't measure. The players who improve fastest track results by contest type, not just overall. "I'm down this month" is useless; "I'm +12% in cash and β40% in 150-max millies" tells you exactly what to cut.
Once a month, sort every contest type by RAROI. Cut or shrink the bottom room. Add volume to the top. That single habit β pruning your worst room and feeding your best β will out-perform almost any lineup improvement, because it's compounding a structural edge instead of a noisy one.
The whole guide in one sentence:
You don't win DFS by being right β you win by being right in a room that pays you enough for it, often enough that your bankroll survives the nights you're wrong. Contest selection is how you build that room, on purpose, every slate.
Every CourtEdge playbook tells you the contest-by-contest attack for the slate, and the Contest ROI Simulator lets you build (or import) your lineups and rank them by ROI and RAROI against the exact contest you're entering β cash, single-entry, small-max, or a 150-max milly. The framework in this guide plus the tools is the whole loop: choose the right room, build the right lineup for that room, and prove it in the sim before you fire.