DFS Lineup Construction: Salary Cap and Rake

The best lineup is not the one stuffed with the highest projections. It is the best one that still fits under the cap. DFS lineup construction treats the roster as a puzzle: positions to fill, a budget to respect and contest rules to read. You will see the problem written as a simple choice for each player, then work a small example where two lineups cost different amounts yet project the same points. The lesson explains why salary efficiency is a clue rather than the goal, how rake raises the break-even point you need to clear, and what to check before assuming any contest format is standard. Everything is practiced on paper.

Learning objectives

• Explain the method, its assumptions, and when it fails.
• Recreate all examples and interpret net outcomes.
• Complete a paper-only workflow using contest rules and data timestamped before lock.

The problem a DFS lineup solves

A salary-cap contest asks you to select an eligible roster while respecting a budget and roster rules. The objective is not simply to collect the players with the highest projections. A combination must fit positional slots, salary restrictions, contest-specific limits, and the scoring system. In some contests players may have multipliers or captain designations; always read current rules rather than assuming a standard format.

Build the mathematics

Define a binary decision xᵢ = 1 when player i is selected and 0 otherwise. Let sᵢ be salary and μᵢ the player's projected fantasy score. A baseline optimization maximizes Σ μᵢxᵢ, subject to Σ sᵢxᵢ ≤ C and valid roster constraints. This describes projected points, not certain winnings and not tournament-optimal play. If roster rules require exactly one quarterback, write Σ xᵢ over quarterbacks = 1. You can also exclude a player who is ruled out.

Example: A fictional three-player roster needs one player from each of three positions and a $25 budget. Position A offers Player A1: $10, projected 20; or A2: $7, projected 15. Position B offers B1: $9, projected 18; or B2: $7, projected 13. Position C offers C1: $8, projected 16; or C2: $5, projected 9. Selecting A1+B2+C1 costs $25 and projects 49 points. A2+B1+C1 costs $24 and projects 49. A1+B1+C2 costs $24 and projects 47. The first two tie on projection but may differ greatly in volatility and outcome dependence. A full optimizer compares all eligible rosters, not just these examples.

Salary efficiency is a clue, not the objective

Points per dollar can help identify value but should not be maximized blindly: a cheap player who projects for 8 points per $1,000 may still cost you ceiling needed for a tournament. Opportunities also change when starters are inactive. Document the time and source of each injury and lineup update; late news can invalidate earlier projections.

Contest rake changes break-even

A contest with 100 $10 entries collects $1,000. If the posted prize pool is $900, the implied fee or rake is $100, or 10% of entry fees. A player who thinks their opponents are equal skill should not expect zero net return in a contest with rake. Exact rake and prize accounting may vary by platform, overlays and promotional rules.

Guided lab

1. Write down scoring, roster slots, max salary, team limits and late swap settings for a fictional contest. 2. Estimate projections and uncertainty bands. 3. Optimize a simple valid lineup. 4. Compare two tied projection lineups on correlations and downside. 5. Measure effective entry cost and contest payout terms. 6. Decide whether your projections are sufficiently informed to evaluate the contest at all.

False shortcuts

Do not confuse top raw projection with highest chance to win a tournament. Do not trust points-per-dollar rankings without checking position scarcity. Do not assume rake is returned when a contest fills or does not fill. Do not copy projected ownership values as certainties.

Practice and answer key

Q1: With a $25 salary cap, can a $26 roster be used if its projection is higher? Answer: No. It violates the cap. Q2: A contest receives $1,000 in entries and pays $900 in prizes. What's implied rake? Answer: $100, equal to 10% of entries. Q3: Two lineups tie at 49 points projected. Must they have equal tournament win chances? Answer: No; joint distributions, ownership and scoring variance may differ.

Worked example

Dana plays a fictional salary-cap contest with a $50,000 cap. She picks five players with salaries of $11,000, $10,000, $9,500, $9,000 and $8,500. Total: 11000 + 10000 + 9500 + 9000 + 8500 = 48000. That leaves $2,000 under the cap.

Her projections for the five are 24, 21, 19, 17 and 15 points. Total: 24 + 21 + 19 + 17 + 15 = 96 projected points.

Now the rake. The contest has 50 entries at $20 each, so entries bring in 50 x 20 = 1000 dollars. The posted prize pool is $900. The rake is 1000 - 900 = 100 dollars, which is 100 / 1000 = 10% of entry fees.

If every player were equally skilled, the average payout per entry is 900 / 50 = 18 dollars on a $20 entry. That is a $2 average loss before any skill edge.

Recap

1. A salary-cap lineup is a constrained choice: maximize projected points within the budget and roster rules, but a high projection is not the same as a high chance of winning a tournament.
2. Two lineups can tie on projection (49 points in the example) and still differ in variance and correlation.
3. Rake (10% when $1,000 of entries pays $900) means even evenly skilled players do not break even, so check the rules and timestamp your news before lock.

Frequently asked questions

How do you build a DFS lineup under a salary cap?

Pick the roster that maximizes projected points while staying under the cap and meeting every position and team rule. A high projection alone is not enough, since the lineup has to be valid and fit the contest's scoring.

What is rake in a DFS contest?

Rake is the fee the platform keeps. If $1,000 in entries pays $900 in prizes, the rake is $100, or 10%. Even evenly skilled players should not expect to break even, since part of the entry money is kept.

Is points per dollar the best way to pick DFS players?

It can flag value, but it should not be maximized blindly. A cheap player can cost you the ceiling you need, and position scarcity and late injury news change the picture. Treat it as a clue.

Related lessons

• DFS Stacking and Correlation: Late Swap Basics
• DFS Ownership Projections and Leverage Explained
• Cash Games vs GPP: Contest Structure and Lineups

Sources and further reading

• About Sharp App.

Try it in the Pick’em screen. Examples in this lesson are fictional, built to show the method. Work the numbers by hand first, then repeat the same steps on a real board. You can also read how the Pick’em Tool works on the Sharp App site.

Questions? Ask ARBAI about this lesson · Ask our Sharps in Discord →

Sharp App surfaces +EV opportunities from public sportsbook data. 21+ where legal. Betting carries risk, so only stake what you can afford to lose. If betting stops being fun, call 1-800-GAMBLER or visit ncpgambling.org. More at Sharp App responsible gaming.