NBA Lineup Optimizer
Build a DraftKings Classic lineup off our own projections, or let it generate a set. Every number here is the same one the prop boards publish.
No priced DraftKings slate yet. DK posts a slate before it prices it, and this tool waits for real salaries — a slate where every player costs the same would hand you the top projection at each seat and call it optimal.
The fantasy board has tonight's projected points in the meantime.
Other optimizersMLB lineup optimizer DraftKingsNFL lineup optimizer DraftKings + FanDuelPGA lineup optimizer DraftKingsNHL lineup optimizer DraftKings
How this optimizer works
It solves the whole roster at once. DraftKings Classic seats 8 players — PG, SG, SF, PF, C, G, F, UTIL — under a $50,000 cap, and the hard part isn't the cap. It's that a player can fill several seats: a guard listed PG/SG can take the PG seat, the SG seat, the G seat or UTIL, and which one he takes changes who else fits. Treating each position separately finds legal lineups it can't see and reports the best of what it could — which looks exactly like a working optimizer. So the solve is an assignment over every seat at once, and it's exact: no greedy swapping, no sampling.
Where the projections come from
The same place the prop boards get them. Every player's points, rebounds, assists, steals, blocks, threes and turnovers are projected per game, then scored under DraftKings' own table — a rebound is 1.25, an assist 1.5, a steal or block 2, a turnover −0.5 — plus the double-double and triple-double bonuses. Scoring a projection and scoring a finished box score run through one function, which is why a lineup's number here reconciles with the fantasy board rather than drifting from it.
A projection isn't a promise. Minutes move on a late scratch, and the bonus terms are the softest part of the number. Read the pool, not just the total.
Cash or tournament
Maximising projected points is the cash-game answer. It is not the tournament answer: a points-max lineup is the chalk, so it's the lineup you share with the field. That's what the auto-generate tab and the exposure cap are for — several different lineups, with a limit on how many of them any one player appears in, so a single bust can't sink the set. We don't claim an edge on the field here; we claim the arithmetic is right.
Narrowing the pool
A main slate runs past a hundred players and you probably like about fifteen of them. Star the ones you want and every build draws from that list instead of the whole board. Lock a player and he's in all of them — the exposure cap steps aside for him, because you asked for him on purpose. Star too few to fill eight seats and the rest comes from the slate rather than nothing; a pool of three isn't a lineup, and refusing to build one reads like a broken tool. The × still takes a player off the board entirely, by name, one at a time.
DraftKings and FanDuel are different rosters, not one at two caps
DraftKings seats eight under $50,000: PG, SG, SF, PF, C, a G, an F and a UTIL. FanDuel changed its NBA roster for 2026-27 and seats nine under $60,000: one of each position, a G/F flex, an F/C flex and two UTIL. FanDuel also caps you at four players from one team and makes you span three. The solver carries each book's seats, cap and club rule, so a lineup built on one tab is legal on that site and only that site. Each tab also scores its own way — FanDuel pays 3 a steal and block to DraftKings' 2, and pays nothing for a three or a double-double — so the same player ranks differently on the two boards.
One honest limit on the FanDuel side: our salaries come from a public mirror of FanDuel's numbers, which doesn't carry FanDuel's own player IDs. The CSV you download from that tab lists players by name, so it reads as a lineup sheet rather than a file FanDuel's bulk uploader will take as-is.