What percentage of my DFS bankroll should I put in play each night?

Between 5% and 15% depending on your win rate and confidence. Sharp cash-game players with a proven positive ROI can push toward 15–20% of bankroll per slate; GPP-heavy portfolios should stay at the low end (5–10%) because tournament variance is punishing over even 100-slate samples. Above 20% per slate, a bad-luck week can wipe out your ability to keep playing entirely — which is the definition of losing the game whether or not your projections were right.

How much should go to cash games vs. GPPs?

The industry-standard split for a positive-ROI player is roughly 70% cash / 30% GPP. Cash games are your income; GPPs are your upside. Reversing that ratio is how people go broke chasing million-dollar tops with 20% ROI cash-game skill. If you're genuinely edge-negative in cash but breakeven in GPPs, you should stop playing cash entirely rather than treat GPPs as your primary income source — variance is not a strategy.

What is Kelly criterion and does it apply to DFS?

The Kelly criterion is a bet-sizing formula that maximizes long-term bankroll growth given a known edge and payout. Full Kelly assumes you can precisely quantify your win probability, which you can't in DFS — so most sharp DFS players use fractional Kelly (0.25 to 0.5 Kelly) as a sanity check on nightly exposure. The practical takeaway: as your edge shrinks, your position sizes should shrink too. Same slate, worse projections → smaller bankroll bet, not larger to compensate.

How big does my DFS bankroll need to be to play seriously?

At least 30x your typical nightly buy-in for cash-heavy play and 100x for GPP-heavy play. Under 30x cash buy-ins, normal variance can wipe you out during a bad week regardless of skill. Under 100x GPP buy-ins, a single dry stretch (which is inevitable — tournament variance is enormous) can end your bankroll before your skill compounds. If you have $500 and you want to play $10 GPPs, you have 50x — you should be playing $2s until your bankroll doubles.

How do I know when to move up in stakes in DFS?

When your rolling 90-day ROI is positive AND your bankroll is at least 30x cash / 100x GPP of the NEW target stake — not the old one. The temptation is to move up when your bankroll is 30x of your current stake, but that immediately puts you at 15x of the higher stake, one bad week from a wipeout. The correct move: hit 60x of your current stake, then move up to a level where you're 30x of the new one.

DFS Bankroll Management: How to Not Go Broke Playing Daily Fantasy

By DFS Simulator TeamPublished July 28, 202610 min read

Most DFS players who lose money don't lose because their projections are wrong. They lose because they risk too much of their bankroll on a single slate, chase losses after a bad night, and blow their entire roll before their edge gets a chance to compound. Bankroll management is the discipline that keeps you in the game long enough for your skill to matter — and the arithmetic behind it is non-negotiable.

Variance in DFS is not a metaphor. A cash-game player with a real 10% ROI edge still loses money in roughly 35% of individual weeks. A GPP-focused player with a real 20% ROI edge still endures 8-week dry stretches multiple times per year. Those numbers come from binomial arithmetic on typical contest sample sizes, not from being unlucky — they are the floor of the distribution for every skilled player.

Bankroll management exists to survive that variance. Get it right and a modest edge compounds into a real income over a season. Get it wrong and it doesn't matter how sharp your projections are — you'll be broke by month three.

The 30x rule (cash) and 100x rule (GPP)

The simplest bankroll rule with real predictive power:

Cash games: your bankroll should be at least 30 times your typical nightly buy-in. For a $20 nightly cash player, that means holding at least $600 at all times.
GPPs: your bankroll should be at least 100 times your typical nightly tournament entry. For a $10 GPP player, that means holding at least $1,000.

The delta between the two rules is variance. Cash-game outcomes cluster tightly around expected value; GPPs pay 0.1% of the field the majority of the prize pool, so your income arrives in occasional lumps. To smooth those lumps into something you can plan around, you need enough capital to survive the gaps.

A player with $500 who enters $50 GPPs every night is playing with 10x bankroll of the buy-in. Even at a real 20% edge, the probability of going broke inside 30 nights is above 40%. The same player, same edge, playing $5 GPPs (100x bankroll) has a bust probability under 5% over the same 30 nights and the same total capital risked over the month.

Nightly exposure: 5% to 15% of bankroll

Once you have adequate total bankroll, the second rule is how much of it you put in play in a single slate. The standard band:

5–8%: GPP-heavy portfolios, unfamiliar sport, or when you're less confident in your reads.
8–12%: Balanced night with normal slate confidence.
12–15%: Cash-heavy portfolio, favorable conditions, high-edge slate (early-week NBA, main-slate NFL Sunday).

Above 15% you're treating your bankroll as if a single night's variance can't hurt you. It can — even the best players have "fat zero" nights where nothing cashes. If a 15% exposure night going to zero would meaningfully change your bankroll trajectory, you're over-levered.

The cash/GPP split

Cash games (50/50s, double-ups) pay the top ~44% of entries a flat payout — usually about 1.8x the buy-in. GPPs pay the top 20% of entries, but the payout structure is severely top-heavy: first place often gets 15–25% of the entire prize pool.

The consequence for bankroll management is fundamentally different variance profiles:

Cash: your ROI stabilizes over ~200 slates. Skill visible in ~50 slates. This is your income surface.
GPP: your ROI barely stabilizes over <500 slates. Skill invisible in individual nights. This is your upside surface.

The 70/30 cash/GPP split most sharp players use is a consequence of this — you want 70% of your action on the surface where variance is bounded and skill compounds quickly, and 30% on the surface where you buy occasional tournament tickets that pay for the compounding.

The Kelly-lite framework

The Kelly criterion says the optimal fraction of your bankroll to bet is (edge × odds − 1) / odds. It assumes you know your edge precisely, which in DFS you don't. Full Kelly is aggressive; even a small overestimate of your edge leads to overexposure and busts.

The workable version: assume your true edge is HALF of what your recent ROI suggests, and size your positions accordingly. This is Kelly-lite. A player who thinks they have a 20% edge should size positions as if they had a 10% edge, which corresponds to nightly exposure at the lower end of the 5–15% band.

Moving up in stakes

The temptation after a hot streak is to move up quickly. The arithmetic argues against it:

At $500 bankroll playing $10 games, you have 50x. A hot week takes you to $800, and moving up to $20 games seems justified.
At $800 playing $20 games, you have 40x — worse bankroll coverage than before the streak.
The right rule: don't move up until you have 60x of your CURRENT stake AND you'd still have 30x of the new higher stake after the transition. That gives you double the safety margin the old stakes needed.

Moving DOWN is symmetric — if you drop to 20x bankroll of your current stake, drop stakes immediately. The alternative is called "chasing" and it has ruined more winning-EV DFS players than bad projections ever will.

Contest selection compounds bankroll rules

A player following every rule above still loses money if they're entering unwinnable contests. The most costly example: the Millionaire Maker on DraftKings. Fields of 200,000+ entries, top-heavy payouts. To beat the average entrant's EV, you need to be in the top ~0.5% of constructors. Sharp players enter these, but they enter with 20+ lineups constructed to span different game scripts, not with one hopeful lineup.

Single-entry GPPs, contests where you can only enter once, are structurally friendlier to bankroll-managed players: smaller fields, sharper opponents, but every entry counts. Cash games are the friendliest of all — flat payouts, no tail dependence, ROI stabilizes fastest.

The DFS Simulator role

Simulation doesn't reduce variance — it quantifies it. The DFS Simulator run summary shows floor, mean, and ceiling per lineup precisely because those three numbers correspond to cash / expected-outcome / GPP-tournament strategies. Bankroll rules tell you HOW MUCH to enter; the simulator tells you WHICH lineups to enter for the contest type you chose.

The two disciplines compound. A sharp bankroll manager with weak projections goes broke slowly; a sharp projector with no bankroll management goes broke quickly. The winning combination is disciplined on both axes — and the DFS surface has enough dead money that the compounding is real over a season.