Open the app and pick a slate
Head to the simulator and sign in (the free plan needs no card — it runs the same engine on a 1,000-lineup field). The top bar is your command center for the whole session:
- The Slate dropdown switches between Main (Classic) and Showdown single-game contests.
- Next to it, the classic slate selector narrows Main to the Early or Afternoon slate when those are available for the week.
- The two numbered buttons — ① Generate and ② Simulate — are the two phases of every run, in order.
Everything below the top bar is one full-height table that switches between four tabs: Projections → Pool → Lineups → Exposures. That's also the order you'll use them in.
Review the projections
The app opens on the Projections tab: every player on the slate with salary, projection, projected ownership, optimal rate, and points per $1k. Position filter chips and the search box narrow the list; clicking a column header sorts.
The Projections tab — the session starts here, and every number is editable.
Both projection and own % are editable. Have a different opinion on a player than the feed does? Click the value and type yours — the engine refits that player's outcome distribution around your number, and the ownership you enter becomes the field-building target. Overrides apply to this session only; the saved database is untouched, and ↺ Reset puts everything back.
You don't have to touch anything here. The defaults are a full, coherent slate. Editing is for when you want to press a personal read — fade a player's ownership, boost a projection — and see what the field does with it.
Check your settings
The ⚙ Settings popover controls the contest you're simulating and how hard the engine works:
Settings — contest shape on top, simulation depth in the middle, field size at the bottom.
- % TO 1ST / CASH LINE
- The payout curve's shape: how top-heavy the tournament is and what share of entries cash. Match these to the real contest you're entering.
- SIMULATION
- Quick (10k sims) while you iterate, Full (50k) for real runs, Max (100k) when you want the deepest read.
- FIELD
- How many opponent lineups to build — 10k is the standard; go bigger to model large-field GPPs.
- FIELD STACK %
- How much of the field stacks QBs with pass catchers, i.e. how sophisticated your opponents are.
Field size and stack % apply on the next Generate; sim mode, copula, and uncertainty apply on the next Simulate — so you can re-simulate the same field at different depths without rebuilding it.
Generate the field
Click ① Generate. The engine builds the full contest field — thousands of opponent lineups, constructed so that every player's exposure lands within a tight band of their projected ownership. A progress readout runs in the top bar; on the standard 10,000-lineup field it takes under a minute.
Generating — the progress readout sits in the top-right corner.
When it finishes, the app switches to the Pool tab: the receipt for the field it just built. For each player you see projected ownership, actual ownership in the generated field, and the difference — proof the field you're about to sim against is the field the ownership projections describe, not a caricature of it.
The Pool tab — projected vs generated ownership, per player. Diffs stay small by construction.
Simulate the contest
Click ② Simulate. The engine now plays out the slate tens of thousands of times — every player drawn from their own skewed outcome distribution, correlated so stacks spike together — and scores the entire field in each one, paying the contest's actual prize curve with duplicate lineups splitting their prizes.
Simulating 50,000 contests — same top-bar progress readout.
Generate once, simulate freely. The field is fixed after Generate, so you can re-run Simulate with different sim counts or uncertainty settings and compare boards against the identical field.
Read the lineup board
The Lineups tab is the product of everything above: every lineup in the field, ranked by simulated ROI.
Top ROI lineups after 50,000 simulated contests.
- ROI %
- Average return on the $20 entry across every simulated contest — the ranking stat, and the number that already accounts for ownership and duplication.
- WIN %
- How often the lineup takes first place outright.
- OWN Σ
- Total projected ownership of the nine players — a quick chalk-vs-contrarian gauge.
- DUPES
- Expected copies of this exact lineup in the field. A duplicated lineup splits its prizes, which is why chalk builds sink on the board even when they score well.
- STACK
- The lineup's construction — e.g. HOU QB+2 is a Texans quarterback with two of his own pass catchers.
Search filters the board by player or stack; ⊞ Columns toggles extra columns; the Diagnostics strip at the bottom opens ownership-vs-ROI and projection-vs-ROI charts for the whole field.
Boost players and set exposure caps
The board ranks lineups by the engine's math — 🚀 Player Boost is where you overlay your own convictions without re-running anything.
Player Boost — a +25 boost on one player, a −15 fade on another, a 40% cap on a chalk RB, and a 60% global cap.
- Boost (%pts) adds or subtracts ROI percentage points from every lineup containing that player, and the board re-ranks instantly. It's display-only — the simulation itself is untouched, so Clear all boosts restores the pure board.
- Max Exp % caps how large a share of your favorited lineups a player may occupy. The global cap applies to everyone; a per-player value overrides it; blank inherits the global.
- The Your % / Opt % / Diff columns show your current exposure against the player's optimal rate, so you can see who you're over- or under-weight on while you tune.
Hit Apply & close and the board re-ranks with a boosted ranking badge so you always know when you're looking at an adjusted view.
Favorite the lineups you want to play
Only favorited lineups export. Two ways to build the set:
- Hand-pick: tick the ★ checkbox on any row. Hand-picked lineups are never dropped by anything you do later, and they export first, in the order you clicked them.
- Bulk: ★ Favorite N… selects the top 10, 20, 150, or a custom count by ROI — respecting your exposure caps. If taking the next row would push a capped player over their share, the picker skips it and takes the best lineup that fits.
Favorite 20 with caps active — note the skipped rows where the capped RB would have exceeded 40% of the set.
You can re-open Player Boost after favoriting: change a boost or a cap and Apply re-selects the same count under the new rules, keeping your hand-picks.
Review your exposures
The Exposures tab turns your favorited lineups into a portfolio summary: how much of each player you actually hold, and how that compares to the field.
Exposures for a 20-lineup set — every capped player sits at or under their cap.
Leverage is the column to read: your exposure minus the field's projected ownership. Positive leverage on players the sim likes, negative on the chalk you're fading — that's the shape of a GPP portfolio with a reason to exist. If something looks wrong, go back a step: adjust caps or boosts and re-favorite.
Export to DraftKings
⬇ Export N to DK CSV (on both the Lineups and Exposures tabs) downloads your favorited lineups as a DraftKings-formatted CSV — players in DK's slot order with their DraftKings IDs attached, ready for the bulk lineup upload on the DK entry page. Hand-picked lineups lead the file in the order you selected them.
Weekly rhythm: once you've done this loop once, a slate takes minutes — open, spot-check projections, Generate, Simulate, cap the chalk, Favorite N, export. The engine does the heavy thinking between clicks.
Where to go next
- Showdown: flip the slate dropdown to a single-game Captain Mode contest — same loop, with captain-vs-flex ownership and exposures tracked separately.
- The Lineup Builder: hand-build a lineup, sim it against the field, and get an AI critique of its structure, dupes, and leverage.
- Methodology: the statistics behind every number you just used — distributions, correlations, duplication, and the invariants the engine must pass.
- The Edge: why ranking by ROI beats optimizing for projection.