Home/CFB DFS Optimizer

CFB DFS Optimizer

A CFB DFS optimizer solves a constraint problem. It does not solve the hard part, and knowing where that boundary sits saves a lot of misplaced trust.

Updated 2026-10-27 · rebuilt daily

Lineup optimizer

DraftKings college football classic: eight roster spots under a $50,000 cap. FLEX takes a runner or receiver; S-FLEX also accepts a second quarterback.

Loaded with the current DraftKings slate: 120 players at today’s real salaries, across 12 games, highest-priced first. The projection column is blank on purpose, and you have to fill it before this can solve anything. We publish no projections for this sport, and inventing numbers to make the button work would produce a lineup that looks authoritative and means nothing. Type your own projected score for the 8+ players you are considering, delete the rows you are not, then press Re-solve. The solver is exact — for a roster this size it returns the true optimum, not a greedy approximation.

The problem it solves is a known one

Fill every roster slot, stay under the salary cap, maximise projected points. This is a bounded knapsack problem — a standard exercise in integer programming with well-understood solutions. Given a set of projections, an optimizer returns the mathematically best lineup in milliseconds.

There is no proprietary advantage in this step. Two optimizers fed identical projections return identical lineups.

Garbage in, confidently optimal garbage out

Because the solver is exact, it will faithfully maximise whatever numbers you hand it. Weak projections do not produce obviously weak lineups — they produce lineups that look rigorous and are wrong. The output's confidence is a property of the solver, not evidence about the inputs.

The parts an optimizer cannot do for you

Correlation between teammates, exposure across a portfolio of entries, and expected ownership are all decisions you supply as constraints. An optimizer run without them returns the single highest-projected lineup, which in a large tournament is usually the most common lineup on the board — the worst possible place to be.