They simulate
the game.
We simulate their users.
Every commercial optimizer on the market hands its subscribers the same math — and on Sunday those subscribers hand the field the same lineups. WarPaint models what those tools tell their users to build, then builds the lineups the crowd structurally can't reach.
Built to beat the field, not the projection. Stop playing against the players. Start playing against their optimizers.
We turn the industry's consensus
into your leverage.
Point-max optimizers are why hundreds of people submit your exact roster. WarPaint starts from what the field is being told to build, then finds the lineups that live in its blind spots — with the ceiling to actually win when they land.
Your DraftKings slate is detected and loaded automatically. Vegas lines, prop-market projections, injuries, depth charts and weather are already attached to every player.
Load the projection and ownership exports from the sites the crowd subscribes to. Their consensus becomes the field we build against — where the chalk will be, how many copies each lineup will have, and where the sites disagree.
Pick the contest and your entry count. The engine builds your portfolio under the contest's real rules, keeps every lineup clear of the field's builds, and ranks each one by its chance to beat the field — not by projected points.
Download the DK-ready CSV. Sunday morning, one click fixes scratches and re-solves open slots after early games lock. Monday, upload the standings and the engine grades itself against what actually happened.
Built to find the blind spots
in commercial data.
Ten systems, one job: put your entries where the crowd's optimizers structurally won't, with the ceiling to win when the script breaks your way.
Load the projection and ownership exports from the sites the field subscribes to. Their median projection becomes the lineup set the optimizer steps around; their maximum ownership drives the duplicate estimate; where they disagree, a player's range of outcomes widens. Your own projections stay yours.
Milly Maker, 20-max, 3-max, single-entry, Showdown and cash each run their own rulebook: stack shapes, bring-backs, ownership windows, salary bands and exposure caps scale to the field size and entry limit you actually enter. Game scripts from the Vegas line upgrade or downgrade the stack per lineup.
Every portfolio is simulated through correlated game scripts — game total, team share, player noise, defenses inverse to the offense they face. Lineups are ranked by how often they beat the field's best lineup, divided by the copies they'd share the prize with. Points are the input, not the score.
A discrete-choice model of what the field sees: value, projection, recent form, implied totals, prop volume, injury status. Position totals sum the way DraftKings' actual %-drafted does, and the model re-fits its temperatures from real standings each week.
Every subscriber's builds start from their own randomized seed, applied as a small correlated jitter to projections. Same slate, same settings, different portfolio — no two WarPaint users submit the same lineups.
A 14.2 and a 13.9 are the same player. After the solve, projection-equivalent peers (±5% projection, ±$400 salary) rotate in to break shared cores without giving back expected value. QB and locked players never move.
Inactives and big projection movers surface on the Build screen; one click re-solves every affected lineup. After early games lock, the Late Swap solver keeps locked players, forces out scratches and re-optimizes open slots against live numbers — for Smart, Manual or DK-CSV portfolios.
A pre-lock snapshot freezes every projection and ownership number before kickoff. Upload the DK standings on Monday and the engine reports its own error by position, which competitor site was sharpest, and re-weights the field model accordingly.
Feed freshness, injury sync, depth charts, ownership sanity, field-model coverage and the pre-lock snapshot are checked in one panel, so a stale feed never reaches your entries.
A Claude Sonnet 5 strategist reads the slate — Vegas, implied totals, consensus, where our numbers diverge — and writes a plain-English attack plan for the week.
The engine grades itself
every week.
Most tools never tell you how wrong they were. WarPaint freezes its numbers before lock, then scores them against the real DraftKings standings: projection error by position, ownership error and chalk misses, which competitor site was sharpest, and how your best lineup actually finished.
Those errors feed back into the model — ownership temperatures, per-position volatility and the trust given to each competitor site all move with the evidence. The scoreboard lives on your Portfolio page.
Every projection, ownership and volatility number frozen automatically before kickoff — grading never peeks at the post-game pool.
Mean error and bias by position, ours versus the market consensus, worst misses and best calls.
Predicted versus actual %-drafted by position, with the chalk the field over- or under-rostered.
Each competitor site's error is tracked; the field model leans on whichever site has been sharpest.
One site. One ruleset.
Zero diluted math.
Multi-site tools average stack rules, blend scoring and smooth payout curves so one engine can serve six sites. WarPaint is written against DraftKings alone — the roster, the bonuses, the Milly Maker prize cliff — and carries separate rulebooks for cash, small-field, large-field and Showdown without compromising any of them.
| Dimension | Multi-site optimizers | WarPaint |
|---|---|---|
| Roster + scoring | One projection set stretched across several sites with different flex rules, PPR variants and salary caps. | DraftKings only: the 9-man roster, full-PPR with 100-yard and 300-yard bonuses, and the $50k cap are hard constraints in every solve. |
| Contest rules | The same max-entry output trimmed to fewer lineups. A 3-max, a 35k flea flicker and a 200k Milly Maker get the same construction. | Rules derive from the contest you enter: field size and entry limit set stack shapes, ownership windows, salary bands, exposure caps and the duplicate cap. |
| Ownership | Ownership averaged from wherever it's posted first — noisy, and position totals don't add up. | A share model that sums the way DraftKings' actual %-drafted does, re-fit from real standings weekly, with the competitor consensus as the pessimistic ceiling. |
| Duplicates | A vague 'uniqueness' slider. | An explicit cap on expected copies for your field size, plus a shadow field built from the sites the crowd uses so your lineups stay clear of theirs. |
| Ranking | Lineups sorted by projected points. | Lineups ranked by correlated-simulation Field-Win %: how often they beat the field's bar, divided by the copies they'd split with. |
| Late swap | Manual edits with no view of what's locked. | Lock-aware re-solve of open slots, scratches forced out, unique versus the rest of your book, for Smart, Manual or DK-CSV portfolios. |
| Accountability | Nothing. Last week's misses disappear. | Weekly grading against DraftKings standings, published on your Portfolio page. |
Duplicate cap, competitor shadow field, leverage anchors and contest-shaped stacks tuned for six-figure fields where copies kill your payout.
Tighter correlation, higher floors, relaxed ownership windows and concentrated exposure when you only get a few bullets.
Game-script routing from the spread and total, captain exposure budgets, K/DST captain limits and bring-back logic for one-game slates.
Floor objective on median outcomes — no leverage forcing, no ceiling chasing, the highest-floor legal lineup.
Stop playing against the players.
Start playing against their optimizers.
The tools below are good at what they do — which is exactly the problem. Tens of thousands of entries are built from the same three or four projection sets every week. WarPaint doesn't try to out-project them; it ingests their numbers and builds where they aren't. One tier, every feature included.
| What it optimizes | WarPaint DFS | Traditional Pro-Tier Sites | Token-Gated Simulators | Mass-Market Content Tiers |
|---|---|---|---|---|
| Objective | Beat the field's lineups | Maximize projected points | Maximize simulated points | Maximize projected points |
| Field model | Their consensus, ingested | None | Ownership-weighted random | None |
| Duplicates | Explicit copies cap | Uniqueness slider | Estimated | None |
| Accountability | Graded vs DK weekly | None | None | None |
| Price | $99/week · all-in | $150–$299/mo · gated | $200–$350/mo · credits | $29–$149/mo · bundle |
| Feature | WarPaint DFS | Traditional Pro-Tier Sites projection-first subscriptions | Token-Gated Simulators pay-per-simulation tools | Mass-Market Content Tiers multi-sport bundles |
|---|---|---|---|---|
| Field model built from commercial-optimizer consensus (their builds become the shadow field) | ✓ | — | — | — |
| Explicit expected-copies cap for your field size | ✓ | — | PARTIAL | — |
| Correlated Monte Carlo (game → team → player) with Field-Win % ranking | ✓ | PARTIAL | ✓ | — |
| Ownership as a share model that sums like real %-drafted, re-fit weekly | ✓ | — | — | — |
| Contest rules derived from field size + entry limit (Milly / 20-max / 3-max / SE / Showdown / cash) | ✓ | PARTIAL | PARTIAL | — |
| Vegas game-script stack routing per lineup | ✓ | — | — | — |
| Per-user variance seed — no two subscribers ship the same portfolio | ✓ | — | — | — |
| Post-solve peer swap for uniqueness without EV loss | ✓ | — | — | — |
| Sunday-morning inactive fix across the whole portfolio in one click | ✓ | — | PARTIAL | — |
| Lock-aware Late Swap re-solve (open slots re-optimized, locked players kept) | ✓ | PARTIAL | ✓ | — |
| Weekly self-grading against DraftKings standings (projection + ownership error) | ✓ | — | — | — |
| Competitor site trust weights learned from graded results | ✓ | — | — | — |
| Pre-lock preflight checklist (feeds, injuries, ownership sanity, coverage) | ✓ | — | — | — |
| Manual optimizer with group rules, exposure caps, peer swap and Field-Win ranking | ✓ | ✓ | ✓ | ✓ |
| Row-level projection & ownership overrides | ✓ | ✓ | ✓ | ✓ |
| Bankroll-aware entry sizing on the Build screen | ✓ | — | — | — |
| AI slate brief (Claude Sonnet 5) | ✓ | — | — | — |
| Live Vegas · prop-market projections · injuries · depth charts · weather included | ✓ | ✓ | ✓ | PARTIAL |
| Price | $99/week · everything | $97–$299/mo by tier | $110–$349/mo + sim credits | $29–$149/mo bundle |
| Commitment | Weekly · cancel any week | Monthly / annual | Monthly + credit packs | Monthly / annual |
| Pricing model | One plan, every feature | Feature-gated tiers | Token-gated simulations | Multi-sport content bundle |
| DraftKings-native rules, scoring and payout curves | Purpose-built | Multi-site generic | Multi-site generic | Multi-site generic |
Category descriptions reflect publicly advertised plans as of the 2026 preseason. Prices vary by promotion.
$99 a week. Every feature.
Cancel any week.
DFS is played week to week, so that's how we bill. No tiers, no add-ons, no season-long commitment — pay for the Sundays you play and keep every feature from the first build.
- ✓ Smart Optimizer with contest-shaped rules
- ✓ Field model from the sites the crowd uses
- ✓ Correlated Monte Carlo + Field-Win % ranking
- ✓ Manual Optimizer with group rules and peer swap
- ✓ Sunday-morning fix + lock-aware Late Swap
- ✓ Weekly self-grading against DK standings
- ✓ AI slate brief
- ✓ Unlimited slates · DK-ready CSV