Orléans

Deep Dive for an iOS Roguelite Translation

Project context: Translate the core loop of Orléans (Reiner Stockhausen, dlp games, 2014) into a single-player iOS roguelite. The "Dominion -> Slay the Spire" pattern, but for bag-building with action-recipe placement. No standalone digital port exists; the only digital presence is BGA and a Tabletopia sandbox.

1. The mechanics, precisely

A round from the merchant's POV. The Hour Glass tile flips, revealing the round's event (Plague, Taxes, Pilgrimage, Trading Day, etc.) and confirming the round count toward the 18-round game. The Census phase rewards the leader on the white Farmer Track with a coin and penalises the trailer. You then reach into your felt bag and draw Character tiles equal to your current Knights Track value — typically 4 to 7 — and lay them face-up in your personal Market. In the simultaneous Planning phase you place those tiles onto the recipe slots of the action Places on your player board: Farm House (Boatman + Craftsman), Castle (Knight + Knight + Trader), University (Scholar + Boatman + Craftsman), Monastery (Trader + Farmer + Craftsman), Ship, Wagon, Town Hall, etc. In turn order during the Action phase, each player executes one activated Place per turn until everyone passes — used Character tiles return to the bag, but Technology tiles placed on slots stay forever (effectively a permanent half-cost discount). Finally the round's event resolves: Plague forces every player to permanently remove a tile from their bag; Taxes drains coins; Trading Day pays out for trading stations. Start player passes left and the next Hour Glass flips.

The bag-build economy. You start with five marked starter tiles — 1 Farmer, 1 Boatman, 1 Craftsman, 1 Trader, 1 Monk (Monks act as wild) — that can never be permanently removed. New tiles enter the bag almost exclusively by firing recipes: working the Farm House yields a Farmer added to your bag; working the Castle yields a Knight; the Monastery yields a Monk; the University yields a Scholar; the Village yields a Boatman, Craftsman, or Trader at your choice. Tiles also leave the bag, and this is where the strategic cuts live. The Town Hall action lets you exile 1–2 tiles to the Beneficial Deeds board, where they sit on themed recipe slots (Cathedral, Castle, Town, Trading Routes, Library) and are gone for the rest of the game in exchange for coins, Development Track progress, and a Citizen tile when a row is completed. The Plague event is the involuntary purge — random, painful, and the reason you cannot just bloat your bag. There is no shuffle/discard/draw cycle as in Dominion: the bag is literally the deck, and the only "shuffle" is you stirring with your hand.

The action-recipe pattern. Every Place is a labelled grid of 2–4 silhouettes — "this slot needs a Farmer, this slot needs a Boatman, this slot needs a Trader." A Place activates only when every slot is filled with the correct token (Monks substitute as wilds). Crucially, slots persist across rounds: you can place a Boatman on the University this round and the Scholar three rounds later — the recipe waits. This single rule is what converts Orléans from a draw-and-do game into a commitment-and-patience game. The constraint isn't "do I have enough tiles?" — it's "do I have the right tile types in the right combinations, and did I draw them this round, and are they more valuable elsewhere?" Knights Track raises your draw count (and is also the only way to scale how many Places you can fire per round); the Farmer Track scales food/income; Boatman/Trader/Craftsman tracks unlock board access and the Place tiles you can claim. Each track has a one-shot Technology bonus that permanently fills one slot of one Place, freeing that tile-type from your recipe forever — the closest thing to a Dominion-style upgrade purchase.

Pressure: the Beneficial Deeds track and events. The 18 Hour Glass tiles include three each of six event types. Plague is the iconic one — every player loses one tile of their choice from the bag, hard pruning. Taxes punishes hoarders. Pilgrimage blocks Monk recruitment. Trading Day rewards the player with the most trading stations. Income pays per Development Track step. Harvest demands Farmer Track resources or coin. The Beneficial Deeds board itself is the secondary clock: the rows fill up for everyone, and once a row's Citizen tile is claimed by another player, you've lost that race. So you have three nested clocks — round count, event timing, and Beneficial Deeds row depletion — all squeezing the same bag.

Sources: Orléans Rulebook PDF, BGA Gamehelp, Ultraboardgames rules, dlp games official.

2. What makes Orléans work as a tabletop game

The aesthetic — the dopamine moment. Shut Up & Sit Down nailed it: "the simple tactile pleasure of rattling around your little people, tumbling them through your fingers, then being delighted or dismayed by what you got." The peak hit is when you draw your tiles and the right Knight shows up exactly when you needed it — you fire the Castle action you've been priming for three rounds, your draw count permanently jumps from 5 to 6, and the rest of the game gets faster. It's the inverse of Quacks of Quedlinburg's bust moment: in Orléans the drama is resolution (the recipe completes) rather than escalation (the next chip explodes you). Justin Bell at Meeple Mountain describes the emotional curve: "For those first 2-3 rounds, I maybe took one action each round. But then as my bag began to fill..." — the engine accelerates from molasses to torrent over 18 rounds, and that acceleration is the game.

The tension of "stuck not knowing if you'll draw the right followers." Because recipe slots persist across rounds, every bag draw is a Bayesian update on a plan you committed to before the draw. You bought a Knight last round because you were going to fire the Castle this round; you draw 5 tiles and there are no Knights, only the one you just bought is somewhere in a bag of 14. Now you must decide: place the partial recipe and wait, hope the Knight surfaces next round (committing one of your 4–7 draws to be useless), or pivot the placement to a different Place entirely. This is the central Orléans feeling — managed disappointment. You are never wiped out by a single bad draw, but you are constantly re-planning. Compare to:

Dominion: deterministic — you will see every card in your deck once per shuffle. Draws are independent across cycles. Orléans removes that guarantee — your Knight might sit at the bottom of an unstirred bag for three rounds.
Quacks of Quedlinburg: bag-pull, but pure push-your-luck — you either bust or you don't. There is no recipe; the chips combine additively. Orléans gives you combinatorial tile demand (the right mix, not the right count) which deepens the planning surface dramatically.
Hyperborea / Roll for the Galaxy: closer cousins. Hyperborea is the most direct precedent (bag of cubes onto action tracks) but lacks Orléans' Beneficial Deeds purge mechanism and event clock.

Why bag-building beats deck-building here. Stockhausen built a system where the player meaningfully shapes the bag composition — unlike Quacks where chip purchases are constrained by what's offered, in Orléans you choose which Place to fire, which directly chooses which tile enters your bag. So the bag is deliberate, but the order of draw is random. That split — deliberate composition, stochastic surface order — is the pure form of bag-building, and it's why fans say Orléans is the genre's defining game. dlp's own tagline captures it: "You always want to carry out more actions than you can" — the bag never gives you enough, and that scarcity is the engine.

Designer commentary worth carrying forward. Reiner Stockhausen has been less interview-active than Granerud (Heat), but the dlp publisher description is signal: Orléans is described as Stockhausen's "most complex and comprehensive work to date," and the design intent is strategic variety — "Depending on your strategy, you should consciously choose and coordinate your followers." The praise pattern across BGG and reviewer sites converges on one phrase: "all the Place tiles are viable" — meaning Stockhausen succeeded at building a system without a dominant strategy, where Farmer-rush, Knight-rush, Scholar-rush, and Trade-rush all win at roughly equal rates in the hands of skilled players. This is the design property to protect in the iOS port — multiple viable build paths is the heart of the replay loop.

Sources: Shut Up & Sit Down review, Meeple Mountain — Games We Love: Orléans, Opinionated Gamers — Dan Blum review, Beyond Solitaire — Orléans solo review.

3. Existing digital attempts

Official digital Orléans: none, in the sense of a standalone branded app. As of April 2026, dlp has not licensed a dedicated iOS or Steam port. The two existing digital surfaces are:

Board Game Arena. Full implementation, browser-based, asynchronous-friendly, multiplayer-only, freemium. Faithful port; no single-player automa, no roguelite reframing, no mobile-native UI.
Tabletopia (Steam + iOS + Android). Sandbox-only — the engine is generic, you must manipulate every token by hand. Acceptable for groups familiar with the rules, hostile to onboarding. Tabletopia gates Orléans behind the base game purchase and a Tabletopia subscription/pass.

There is no native iOS app, no AAA single-player port à la Twilight Struggle or Through the Ages. The Orléans: Invasion expansion ships official solo scenarios ("The Dignitary," "Capital Vierzon," "Traveling Salesman") plus a co-op mode, and Beyond Solitaire reports they work but eventually become predictable because "the solo scenarios have set events that occur in a specific order." That predictability is exactly the gap a roguelite fixes.

Bag-builder roguelites on Steam/iOS. The genre is strikingly sparse compared to deck-builder roguelites. The deckbuilder-roguelite slot is saturated (Slay the Spire, Monster Train, Balatro, Vault of the Void, Wildfrost, Dicey Dungeons, Pirates Outlaws, Griftlands, Backpack Hero, StarVaders). The bag-builder corner contains essentially:

Backpack Hero — inventory-grid placement, not bag-pull; spiritually adjacent (placement-on-recipe is the design language) but mechanically distinct.
Inscryption — card-based, not bag.
A handful of itch.io experiments using "bag" tag — none with traction.

The clear nearest neighbor in mobile is Backpack Hero, which proves that a placement-puzzle roguelite can work on touchscreens with the right grid metaphor. There is no shipping bag-builder roguelite that uses the draw-N-tokens-then-place-on-recipes loop. That slot is empty.

Worker-placement-on-recipes mobile games. The category is thin. Lords of Waterdeep has a polished iOS port (worker placement, no bag), and a Viticulture iOS port shipped from Stonemaier with strong solo automa support — both are faithful adaptations rather than roguelite reframings. No mobile game I can find combines (a) worker-placement-on-action-recipes, (b) bag-pull randomness, and (c) roguelite run structure. Orléans-as-roguelite is genuinely an empty design slot.

Sources: Tabletopia Orléans on Steam, BGA Orléans, Backpack Hero on Steam, MiniReview — best mobile roguelike deckbuilders, D&D Lords of Waterdeep iOS.

4. The translation problem

What gets stripped.

Multiplayer simultaneous Planning. SU&SD called this Orléans' "greatest design strength" — everyone draws and plans at once, killing downtime. In single-player there is no downtime to kill; this is a non-loss in solo.
Position-rivalry on the map. Trading station races, the Trader leader bonus, the Citizen-tile race for Beneficial Deeds — these all assume opponents racing the same shared resources. Replace with AI rivals competing for shared Beneficial Deeds rows and a shared map, but cut the Knights/movement subgame: the Loire-region map is texture, not core to the bag loop.
18-round runtime (60–90 min tabletop). A mobile run must compress to 8–12 minutes per "race" so you can fit 3–5 races in a roguelite arc. Cut to 8–10 rounds per run with a faster Beneficial Deeds clock.
Tactile bag rummage. Lossy but unavoidable. Mitigate with sound design (cloth-rattle, tile-clack) and an animated draw moment where tokens visibly tumble out.

What gets added (the digital meta-layer — pick 2–3 for v1).

Roguelite expedition. A run is 4 "chapters" played on increasingly hostile maps. After each chapter you draft 1 of 3 Follower Pack upgrades (e.g., "Add 2 Knights and 1 Scholar to your bag"; "Replace 3 starter tiles with a custom Trader-heavy pack") and 1 of 2 Relics (passive bag/draw modifiers).
Relic system that modifies bag draw rules. Concrete examples: Loaded Dice (always draw 1 extra Monk per round); Plague Mask (immune to Plague, but lose 1 coin per round); Apothecary (when a Plague hits, choose between two random tiles); Stirring Rod (peek at top 2 tiles before drawing); Heretic's Ledger (gain 1 VP each time you exile a tile to Beneficial Deeds). These are the Slay-the-Spire-relic equivalent and the primary build-identity vector.
Asymmetric "Patron" characters — pick a starting bag composition + starting board: the Abbot starts with 2 Monks and a discounted Monastery; the Marshal starts with 2 Knights and pre-built first Castle slot; the Steward starts with 4 Farmers and a Beneficial-Deeds discount. Three to five Patrons at v1, each with unique Place tiles unlocked.
Daily Plague seed. Fixed map + fixed event sequence + fixed starting bag, daily leaderboard by VP at end of run. The streamer/Twitch surface.
"Boss event" final round. Last round of each run is a souped-up Plague that strips multiple tiles unless you've completed certain Beneficial Deeds rows. The dramatic climax replacement for "the last round of a Euro."

Recommendation: ship v1 with #1 + #2 + #3. Roguelite arc is the spine; relics are the build identity (and the differentiator from the tabletop game); Patrons unlock the asymmetric-start dimension that gives the meta-progression layer something to hand back. Daily and Boss event are post-MVP.

What cannot translate cleanly.

The "all Place tiles viable" property is hard to preserve when you remove the multiplayer race for shared resources, because much of what balances Orléans is opponent pressure. Without it, dominant strategies will emerge faster. Mitigation: vary the chapter map and Beneficial Deeds composition each run so no single Place is always optimal.
The 90-minute commitment-and-patience arc. Mobile attention demands that bag growth be visible within 90 seconds of opening the app. The deep slow-burn engine acceleration over 18 rounds compresses awkwardly. Consider front-loading bag expansion (start with 8 tiles instead of 5) and accepting a slightly less dramatic acceleration curve.
Negotiated table-talk plays (sniping a Citizen, blocking a Place tile from a known leader). Replace with AI rivals that have visible intent — show a thought bubble icon over the AI's player area, "going for the Cathedral row," so the player has a target to race.

5. Concrete iOS prototype spec

Target: iOS portrait orientation, single-developer SwiftUI + SpriteKit (or Unity), 4–6 month MVP.

60-second core loop. Tap "Draw" — your bag visibly shakes, 5–7 tokens tumble into the Market strip across the middle of the screen. Each token is a colored chip with a profession icon (Farmer green, Boatman blue, Craftsman brown, Trader black, Knight red, Scholar grey, Monk yellow). Drag tokens onto recipe slots of the Place tiles arranged on your board (bottom 50% of screen). Filled recipes glow gold. Tap "Resolve Round" — activated Places fire in sequence (animated, ~3s total), tokens return to bag, new tokens earned drop visibly into the bag. Event tile flips with a card-flip animation showing the round's event. AI rivals' actions resolve in a side panel as a 2-line text log ("Marshal fired Castle, drew Knight"). Repeat 8–10 times per chapter, 3–4 chapters per run, ~10 minutes per run.

Single-screen layout (phone portrait, 6.1" reference).

Top 20%: Hour Glass / round counter, current event icon with hover-tooltip, Beneficial Deeds row with shared progress bars (you vs. AI). Compact.
Middle 15%: The Market — horizontal strip of drawn tokens this round. Drag handles for placement.
Middle 50%: Player board — 6–8 Place tiles arranged in a 2×4 grid, each showing its recipe slots and yield. Tap-and-hold to inspect a Place's effect.
Bottom 15%: Bag display (count + breakdown by color), Tracks bar (Knight/Farmer/Trader/Scholar progress as small icons), action buttons (Draw, Resolve Round, Inspect).
Drag-up gesture on the bag: open a full bag inspector showing every tile (essential for purge planning into Beneficial Deeds).

Token system.

Starter bag: 1 Farmer, 1 Boatman, 1 Craftsman, 1 Trader, 1 Monk (mirror tabletop) plus 2 extra Farmers to front-load engine acceleration for mobile pacing.
New tokens enter via firing Places: Farm House -> +1 Farmer; Castle -> +1 Knight; University -> +1 Scholar; Monastery -> +1 Monk; Village -> +1 Boatman/Craftsman/Trader (player choice).
Tokens leave via Town Hall -> Beneficial Deeds (permanent exile, yields VP/track progress) and via Plague event (forced loss).
Tap-to-place and drag-to-place both work; on-screen ghost shows valid recipe slots when a token is held.

Run structure.

4 chapters per run, each 8–10 rounds. Each chapter has its own map (Beneficial Deeds row composition + available Place tiles) drawn from a procedural deck of ~20 chapter templates.
1–3 AI rivals per chapter, drawn from a roster of ~6 personalities (Knight Rush, Farmer Engine, Scholar Tech, Trade Network, Monastery Wild, Balanced). Rivals compete for Beneficial Deeds rows visibly.
Chapter win condition: most VP at chapter end. Top finisher draws first in next chapter's upgrade pool.
Run win condition: survive 4 chapters without being eliminated (last place 2x = elimination), final score = sum of chapter VPs, leaderboard rank determines unlock currency.

Meta-progression.

Between chapters: pick 1 of 3 Follower Pack upgrades + 1 of 2 Relics.
Between runs: unlock new Patrons (asymmetric starting bags), new Relics added to the pool, new Place tiles, new chapter templates. Roughly 40–60 unlockables for a 25–35 hour completion arc.
A "Codex" tracks which Patrons have been won-with, which Relics seen, which events survived — Slay-the-Spire-style achievement progression.

Monetization. Premium one-time, $6.99–$9.99. Optional expansion pack at $3.99 adding Plague-style themed run (mirroring the Orléans: Plague tabletop expansion's identity, with permission/licensing) and 2 new Patrons. No energy mechanics. A run is a session; gating sessions kills the loop.

Aesthetic direction. Lean into the medieval-illuminated-manuscript look — Wes Anderson Beasts of Burden colour palette (muted ochres, deep indigos, gold leaf accents), flat-vector silhouettes for tokens (Mini Metro readability), parchment textures for the board. The tabletop Orléans uses cartoony illustrated tiles; the iOS version should reference period-romance without ye-olde kitsch. Audio: cloth-bag rattle on draw, parchment crinkle on round end, monastery bell on Beneficial Deed completion, plague-doctor crow caw on Plague event. Music: low ambient lute/recorder during planning, swell on round resolve.

6. Risks and unknowns

Top design risks.

The slow-burn engine acceleration may not survive compression. Tabletop Orléans builds across 18 rounds; mobile demands 8–10. If the engine doesn't feel like it's snowballing by round 4, the dopamine arc collapses. Mitigation: front-load starter bag (7 tiles not 5), make the first 2–3 Place activations cheaper (2-slot recipes instead of 3), tune draw counts upward.
"All Places viable" balance breaks without opponent pressure. Multiplayer Orléans has a built-in metagame where opponents punish dominant strategies; solo doesn't. Mitigation: per-chapter Place availability (not every Place is on every map), Relic-driven build-identity steers different runs into different Place choices, AI rivals that target the same Beneficial Deeds rows you go for.
Bag-pull frustration in single-player. In multiplayer, drawing the wrong tiles is funny because everyone's suffering; solo, it's just punishment. Mitigation: a "stir" relic that lets you peek; soft-pity (after N rounds without drawing tile-type X, increase its draw weight); never let the player be unable to fire any recipe in a given round.
Action-recipe placement is spatially demanding on phone. 6+ Place tiles each with 2–4 slots is dense for a 6.1" screen. Mitigation: large recipe slots, drag-snap with magnetic targeting, optional "auto-place obvious moves" toggle for advanced players.
Plague event is a tone-killer in solo. Random tile loss feels great when it hits everyone; brutal when it's just you. Mitigation: always let the player choose which tile to lose (the tabletop rule), and reward Plague survival with a small VP bonus to flip the emotional valence.

What needs validation, in order.

Tabletop solo playthrough first. Play the Orléans: Invasion Dignitary scenario five times before writing code. Does the bag loop carry the experience without opponents? If yes -> the iOS version has a chance. If no -> the AI design becomes the riskiest part of the project.
Paper prototype of the compressed 8-round chapter. Hand-draw a board, halve the Place count to 4, run through with 5 round-events. Does the engine snowball arc still fire? If acceleration doesn't happen by round 4, redesign the starter bag.
Greybox digital prototype of one chapter, one AI rival, no meta-layer. Validate that drag-to-place feels good on a 6.1" screen, that the 60-second core loop is actually 60 seconds (not 3 minutes per round), and that the AI rival creates legible competitive pressure on Beneficial Deeds rows.

Three questions for the prototype builder.

How precious is the bag-as-physical-object aesthetic? The tactile pleasure SU&SD identifies is the Orléans dopamine source. If you can't nail the cloth-rattle draw animation and the visceral "tumble" of tokens into the Market, the digital version is a strictly worse experience than the cardboard. How much of your art/audio budget can you dedicate to selling the draw moment?
Are you willing to deviate from the action-recipe constraint to make placement frictionless on mobile? E.g., "any Monk fills any slot" (already true) extended to "Boatmen and Craftsmen are interchangeable in the Farm House" — heretical, but maybe necessary for thumb-driven play. What's your purity budget?
Solo automa or full opponent simulation? The cheap path is a Viticulture-style automa (simple rules, stateless, fast). The expensive path is a behavior-tree AI that simulates Knight-Rush vs. Trade-Network vs. Scholar-Tech as legible personalities. The latter is what makes the game feel like a roguelite race; the former is what ships in 6 months. Pick now — it shapes the entire codebase.