You open the NFL optimizer for the first time. Stack shapes. Team allocation. Correlation rules. Flex options. Worlds. A diversity slider. Six presets.
It's a lot of knobs. Most of them you won't touch on week one, and that's fine — the presets exist so you don't have to.
But every one of them does something specific. Here's all of it.
THE SHORT VERSION
- Pick your slate at the top. Classic or Showdown.
- Dashboard — see which offenses the model wants you stacking.
- Players — lock, exclude, and set exposure on anybody you have an opinion about.
- Research — read the matchups and build a Matchup Plan. It carries into the Optimizer automatically.
- Sims — browse the ready-made pool of correlated lineups if you'd rather pick than build.
- Optimizer — pick a contest preset, set your stack shapes, hit Build.
- Saved — select your finals, export the CSV, upload to DraftKings.
You don't have to use every page. Dashboard → Optimizer → Saved is a complete workflow on its own.
CLASSIC VS SHOWDOWN
The slate selector at the top drives everything, and the tools change based on what you pick.
Classic is the normal format — QB, 2 RB, 3 WR, TE, FLEX, DST across all the games in that window. FLEX is any RB, WR, or TE. Stacking is built around your quarterback.
Showdown is a single game. One Captain at 1.5x points and 1.5x salary, plus five FLEX, with both teams eligible. Different settings appear for it, and the Stack Shape, Team Allocation, and Correlation Rules panels disappear because they don't apply.
DASHBOARD
This is your read on the slate before you touch a roster.
Projection Status tells you when projections were last generated. If it says Not yet generated, nothing else on the page is accurate yet.
Slate Strategy → Suggested Build shows what mix of stack types the model likes this week, with a one-line reason on each bar. Treat it as a starting point for your Stack Shape Allocation, not gospel.
Team Rankings lists every offense on the slate, ranked by stack score:
| Column | What it is |
|---|---|
| Score | Composite stack rating — the bar is relative to the top team |
| Imp | Vegas implied team total |
| Stack | Projected points for that team's best stack |
| Own% | Average projected ownership across that stack |
| vs Def | The defense they're facing, and that DST's projection |
| Stack size | Suggested number of pieces (4 or 5) |
Click any team row to expand it and see the actual players in that stack — position, salary, projection, ownership — plus stack total, salary cost, and ceiling.
The teams you want are high score with a low Own%. High score plus high ownership is the chalk stack. It wins plenty; you're just splitting the prize when it does.
Leverage Plays are players whose sim win rate sits well above their projected ownership — the guys the field is underrating. The Lev number is that gap.
Value Plays are points per dollar. They're how you afford the expensive pieces.
Weather & Game Environment gives temperature, wind, and conditions for every game with a live radar map. Wind is the one that actually moves NFL projections — under 8 mph it's noise, and somewhere north of 15 mph it starts pulling down the deep passing game and long field goals. Domes are flagged as such.
NFL News pulls the latest injury and beat reports for the slate.
PLAYERS
The full projection table, and where you express your opinions.
Filters: tap a position chip to filter — FLEX shows all RB, WR, and TE. Tap a game to filter to both teams, or a team to filter to one. Tap again to clear.
Sorting drives the whole list:
- Projected — median expected points
- Ceiling — the top end of the simulated range. This is the GPP number.
- Floor — the bottom end. This is the cash-game number.
- Value — points per $1,000 of salary
- Opt — how often a player appears in the optimal lineup across the sims. A ceiling-and-correlation signal, not a projection. A cheap player with a high Opt % is usually a live leverage play.
- Leverage — sim win rate minus projected ownership. Positive means underowned.
- Salary
- Ownership — projected field ownership
Lock forces a player into every lineup. Exclude removes him entirely. Setting one clears the other. Excluding a team excludes every player on it — except anyone you've locked, since a lock always wins.
Min / Max Exposure is the better tool most of the time. Max caps a player (40% means at most 40% of your lineups); min forces a floor. Once you're building more than a handful of lineups, exposure beats hard-locking almost every time.
A lock in all 20 lineups means one inactive designation ruins the entire portfolio. A 70% min exposure gets you the same conviction with an escape hatch.
Expand any player row for his season game log. Columns change by position — completions, passing yards and EPA for QBs; snap share, carries and targets for RBs; targets, target share, air yards and WOPR for WRs and TEs; points and yards allowed for DSTs; made field goals by distance for kickers.
Showdown only: each player row has a captain toggle. Setting a captain also locks him.
RESEARCH
Four tabs: Matchups, Film Notes, Standings, and Formations.
Matchups is the heart of it. Every game is broken down by side with favorable (▲), unfavorable (▼), and mixed (◆) flags. Click any flag to read the reasoning behind it. You also get the model's positional multipliers, defensive rates allowed by position, and a DFS angle line suggesting which side to stack.
Click a position tag on any flag to open a popover and immediately Lock, Target, or Fade that team's players at that position.
Target game tags the core of both teams at once — QB, RB1–2, WR1–3, TE1. Fade game removes every player in that matchup. Fade game is the fastest way to shrink the pool when you're confident a game is a dud.
The Matchup Plan
Everything you tag collects into the Matchup Plan bar at the top. It saves automatically per slate and loads into the Optimizer on top of your manual settings.
- Lock → a hard lock in the optimizer
- Target → a minimum exposure, 25% by default, editable on the chip
- Fade → an exclude
- Stack TEAM → pre-fills that team's stack allocation, partners, and bring-backs
A Plan badge appears on the Players page and the Optimizer showing how many items came from the plan. The Clear button next to it removes only those items — your manual settings stay untouched.
SIMS
Every time a simulation runs, the optimizer builds a large pool of correlated lineups in the background across a range of stack and exposure settings. The Sims page is that pool, ready to browse. If you'd rather pick lineups than build them, work here.
Sorting:
- Ceiling — highest simulated top end. The default, and the right one for GPPs.
- Top 10% rate — how often this lineup finishes in the top 10% of sims. The most honest tournament metric in the list.
- Floor — safest
- Projection — highest median
- Salary, high or low
- Sim Rank
- Actual — real scores, once games are final
Filters work by team, by individual player, or by stack pattern. The stack label is a real structure — something like CIN QB+3+1BB — so you can pull exactly the shape you want out of thousands of lineups.
The Game Scripts panel lists the distinct worlds the simulation produced, what percent of sims each represented, and the average winning score inside it.
OPTIMIZER
The big one. Top to bottom.
Build Settings
| Setting | Range | Default | What it does |
|---|---|---|---|
| Salary Floor | 40,000–50,000 | 49,000 | Minimum salary a lineup must spend. Raising it forces expensive builds; lowering it opens up punt plays. |
| Uniques | 1–8 | 2 | How many players must differ between any two lineups. NFL runs higher than MLB because a game stack already eats 5–6 of your 9 slots. Showdown caps at 5. |
| Max Exposure % | 10–100 | 100 | Global cap on any single player. A per-player limit set on the Players page overrides it for that player. |
| Projection Floor | 0–30 | 0 | Filters out anyone projected below this before building. Set it too high and you kill your punt options. |
| Max / Team | Off–5 | Off | Caps players from one team in a lineup. Careful — set it to 3 and you've banned every QB+3. |
| Max / Game | Off–7 | Off | Caps players from a single game. Set it to 5 and QB+3+2BB is impossible. |
| Max Stack % | Off–50% | 20% | Classic only. Caps how many of your lineups any one team's QB can anchor. Your anti-overexposure lever at the team level. |
Reset at the top right restores every build setting and clears your shapes, team allocation, partners, and selected worlds. It does not touch locks and excludes — those have their own reset on the Players page.
Single Lineup
Build one lineup at a time. You need at least two locks or excludes before Optimize turns on — it's a manual build tool, and with zero input it would just hand you the same chalk lineup every time. Portfolio settings like exposure, max stack %, and uniques are ignored here because they're meaningless for one lineup.
The salary, points, and ownership row at the top updates live as you lock players in.
Stack Shape Allocation (Classic)
Every lineup gets exactly one shape, and you're deciding what percentage of your portfolio takes each. Whatever you leave unallocated is built free-form by the sims — allocate 60% and the other 40% is auto.
QB+N = N pass catchers from the QB's own team. BB = a bring-back from the opposing team.
| Shape | What it is |
|---|---|
| QB+0 | Naked QB — no correlation |
| QB+1 | QB + 1 pass catcher |
| QB+1+1BB | QB + 1, run it back with 1 opponent |
| QB+1+2BB | QB + 1, run it back with 2 |
| QB+2 | QB + 2 pass catchers |
| QB+2+1BB | QB + 2, run it back with 1 |
| QB+2+2BB | QB + 2, run it back with 2 |
| QB+3 | QB + 3 pass catchers |
| QB+3+1BB | QB + 3, run it back with 1 |
| QB+3+2BB | QB + 3, run it back with 2 — six players from one game |
| QB+4 | QB + 4 pass catchers |
| QB+4+1BB | QB + 4, run it back with 1 — six players from one game |
Go over 100% and the extra shapes get dropped. The running total under the panel tells you where you stand.
Rule of thumb: bigger shapes win bigger, but they need the game to actually go off. Small field with few entries, concentrate on two or three shapes. Large field with many entries, spread across five or six.
Team Allocation (Classic)
Stack % is how often that team's QB anchors your lineups. The run-back team is derived from the matchup automatically. Impl is the Vegas implied total, and Vs shows the opponent and their implied total. Leave it all blank and the game scripts pick the anchors for you.
The Sim button fills these numbers straight from the simulation — the percentage of sims in which each offense finished as a top-2 scoring team. Teams under 3% are left blank. It's an informed starting point, not a final answer.
Correlation Rules (Classic)
One rule is always on and can't be turned off: a DST is never rostered against any player you own. Not just against your QB — against anyone. Bring-backs are wide receivers, or an elite pass-catching back.
Two optional toggles:
- Allow any RB from the QB's own team — off by default. Receiving backs are already allowed; this opens it to all backs. Turning it on creates lineups where your QB throwing for 350 and your RB running for 120 have to happen in the same game, which is a real correlation cost.
- A rostered DST must bring an RB from its own team — the game-script play. If the defense is winning, their back is running out the clock.
Flex Options (Classic)
Which positions are allowed in the FLEX slot and what share of your lineups each takes. Defaults are WR 60 / RB 25 / TE 15. Checked positions must total exactly 100% — the build button stays disabled until they do, and the error under the panel tells you what you're at.
Stack Partners (Classic)
Where you name who fills the stack instead of letting the optimizer pick. Choose a QB team, a side (With QB or Bring-back), the player, a percentage, then Add.
The percentage is of that QB's lineups, not your whole portfolio. If the team anchors 30% of your builds and you set a receiver to 60%, he lands in roughly 18% of your lineups. The shape still decides how many pass catchers and bring-backs ride along.
After a build, each partner row shows a hit count — how many of that QB's lineups actually got him. Ask for 60%, get 20%, and something else in your settings is fighting it.
Showdown Settings
- Captain Positions — restrict the captain slot to specific positions. None selected means any position is eligible.
- Team Splits — 5-1, 4-2, or 3-3. How the six spots divide between the two teams. None selected is unconstrained; selecting several builds across all of them.
- Majority Team % — of the lineups using an uneven split, what share gives the bigger half to each side.
Game Scripts
The distinct worlds the simulation produced, each with what percent of sims it represented and the average winning score inside it. All Worlds is the default and builds across every scenario. Select specific worlds to build only from those game scripts — this is how you aim a portfolio at an exact scenario, like the shootout world, or the one where the favorite blows them out.
Multi-Lineup Builder
A contest preset sets build count, diversity, max exposure, uniques, and max stack % at once, then auto-selects the top lineups after the build. Everything stays editable afterward.
| Preset | Builds | Keeps | Diversity | Max Exp | Uniques | Max Stack |
|---|---|---|---|---|---|---|
| 150-Max | 150 | 150 | 100% | 45% | 3 | 25% |
| 20-Max | 60 | top 20 | 85% | 55% | 3 | 35% |
| 5-Max | 30 | top 5 | 70% | 60% | 4 | 50% |
| 3-Max | 25 | top 3 | 60% | 70% | 4 | 50% |
| Single | 25 | top 1 | 40% | 100% | 4 | 100% |
The smaller presets deliberately over-build and keep only the best, with a cap on how many lineups can share the same QB stack. That's why 5-Max builds 30 and hands you five.
Showdown presets use the same idea with higher diversity across the board and a $48,000 salary floor — Showdown winners' captains tend to be a top-3 actual scorer but only around sixth in projection, so you have to build inside the worlds rather than pull everything back toward the projection.
The Diversity slider runs 0 to 100, Chalk on the left, Unique on the right. It tilts every player toward his ceiling or floor based on how he actually finished in a sampled game script. At 0 you're building straight off projections and every lineup looks the same. At 100 you're building inside individual simulated worlds.
Diversity does nothing if no simulation has run for the slate. Check Projection Status on the Dashboard first.
Built Lineups
Sort results by build order, stack type, salary, projection, ownership, ceiling, or actual points. The order you set here carries into your saves and exports, so sort before you save.
Each lineup carries a stack badge. A badge containing +Nx is highlighted — that flags opposing-team players who rode along but don't count as a true bring-back, which is usually worth a second look. A badge marked auto means the sims chose the depth rather than you.
SAVED & EXPORT
Your saved lineups, grouped by build. Select what you want and export the DraftKings CSV. In DraftKings, open the contest, choose bulk lineup upload, and select that file.
Whatever sort order you set on the build carries through into the export.
THE SETTINGS DON'T WIN THE WEEK. THE STRUCTURE DOES.
RECOMMENDED STARTING SETUPS
SMALL-FIELD GPP
Small fields are the best place to learn. Less duplication means you don't need extreme leverage to get paid — you need correlation and a real ceiling.
1. Start on the 20-Max preset, or 5-Max if you're entering fewer.
2. Check the Dashboard's top five teams and pick three or four offenses you actually like.
3. In Team Allocation, give those teams 25–30% each. Leave the rest blank so the sims fill in around you.
4. In Stack Shape Allocation, concentrate: 40% QB+2+1BB, 30% QB+3+1BB, 20% QB+1+1BB, leave 10% auto. In a small field, always run the bring-back — you're not fading correlation to be different from 200 people.
5. Leave Max Stack % at the preset's 35%. A small field lets you be more concentrated than a 100,000-entry tournament would.
6. Diversity 80–90%.
7. Use min exposure on the two or three players you genuinely believe in rather than locking them.
8. Build, sort by Ceiling, keep the top N, save, export.
CASH
Use the Single preset with diversity at 40%. Sort the Players page by Floor instead of Ceiling. Keep the stack small — QB+1 or QB+2, no bring-back. Don't chase leverage; ownership doesn't matter when you only need to beat half the field.
SHOWDOWN
Start with the preset matching your entry count. Restrict Captain Positions if you have a strong read — QB and WR captains are the most common winners, while a kicker or DST captain is a real but rare outcome.
Add 4-2 and 3-3 to Team Splits if you expect a close game, 5-1 if you expect a blowout.
TROUBLESHOOTING
| Symptom | Fix |
|---|---|
| A setting got ignored | Conflicting rules made it impossible. Usual culprits: Max/Team at 3 or lower while asking for QB+3, Max/Game below your stack size, exposure caps under what your locks require, or more locks than a stack allows. Errors appear in a red panel at the top of the Optimizer. |
| Build button won't turn on | Flex Options has to total exactly 100%. |
| "2 more lock/exclude needed" | That's Single Lineup. Lock or exclude at least two players first. |
| Fewer lineups than requested | The constraints ran out of legal combinations. Lower uniques, raise max exposure, or widen your stack shapes. |
| Diversity isn't doing anything | It needs a simulation to have run for that slate. |
| Every lineup looks the same | Raise Uniques, raise Diversity, lower Max Exposure, spread the shape allocation wider. |
Still stuck? The in-app Help page has a searchable version of all of this, plus FAQs for MLB and NFL. Or email support@onlydfs.com.
// Put it to work
Everything in this guide can be set in one place. The NFL DFS optimizer builds DraftKings NFL lineups inside correlated game sims with your stack shapes, bring-backs, and exposure caps.
BUILD YOUR FIRST
NFL PORTFOLIO.
Simulations, projections, QB stacking, leverage scoring. One subscription, all three sports.
START NOW