Tigris & Euphrates

Deep Dive for an iOS Roguelite Translation

Project context: Translate the core loop of Tigris & Euphrates (Reiner Knizia, Hans im Glück, 1997) into a single-player iOS roguelite. The "Dominion -> Slay the Spire" pattern, but for Knizia's tile-laying / area-control / lowest-of-N masterpiece. The Codito iOS port was delisted on Dec 31, 2020 when the license expired; no live mobile version exists today, and no roguelite reframing has ever been attempted. This is the most structurally unique candidate in the project's top-100 short-list because the lowest-color scoring mechanic is essentially nowhere else in the digital catalog.

1. The mechanics, precisely

A turn from the priest's seat. You hold a hand of 6 face-down civilization tiles drawn from a shared bag (~150 tiles in the four colors). On your turn, you take up to two actions, in any order, freely chosen from this menu:

Place a tile from your hand onto an empty land space (rivers only accept blue farm tiles). The tile may join an existing kingdom, found a new one, or — when it bridges two kingdoms each containing a like-colored leader — trigger a war.
Place, move, or withdraw a leader. A leader (the four wooden disks belonging to your dynasty: black king, red priest, blue farmer, green trader) must be placed on an empty land space adjacent to a red temple and may not unite two kingdoms. If you place a leader of color X into a kingdom that already has a leader of color X, you trigger a revolt.
Play a catastrophe tile (you start with 2 of these, can never replenish them) onto an empty space or on top of any non-leader, non-monument tile. The tile underneath is destroyed; the catastrophe square is dead ground for the rest of the game. Catastrophes can split kingdoms, sever leaders from their power base, and turn a treasure into salt.
Discard and redraw any number of tiles from your hand (up to all 6).

After your two actions, you draw back up to 6 from the bag and pass.

The four colors and their leaders. Each tile is one of four colors, and each color has a matching leader who scores that color's victory cubes:

Color	Tile	Leader	Earns a cube when…
Red	Temple	Priest (red)	a red temple is added to (or scored in) the kingdom with your priest, AND a temple is the most-respected color in the kingdom (always counts in revolts)
Blue	Farm	Farmer (blue)	a blue farm tile (river-only) is added to the kingdom; blue tiles also count in revolts/wars only when the king is absent
Green	Market	Trader (green)	a green market tile is added; greens also award treasures (see below)
Black	Settlement	King (black)	a black settlement tile is added; the king also collects red cubes when no red priest is present in his kingdom

When a tile is placed inside a kingdom, the leader of that color in that kingdom gets the cube. If no matching leader exists, the cube goes to the king (if any), else to no one. Cubes are kept face-down in front of you until end-game.

The lowest-color scoring — the singular mechanic. Your final score is the count of cubes in your smallest color. A player with 12 red, 10 green, 6 blue, 2 black scores 2. A player with 5/5/5/5 scores 5 and beats them. Treasures (the wild tan cubes that begin the game on each of the 10 starting temples) act as jokers and can be assigned to any color at scoring — they are the release valve for inevitable bottlenecks. Ties are broken by the next-lowest, then next, then next.

That single rule — min, not sum — does almost all the strategic work in the game. Every tactical decision becomes "do I extend my lead or shore up my deficit?" The player with 12 red has zero marginal incentive to play another red tile; the player with 2 black is desperate for any kingdom with a king. This converts the spatial map into a multi-dimensional pressure puzzle, and it forces every fight to ask "which color does the loser actually need?"

Internal conflict — the revolt. When you drop a same-color leader into a rival's kingdom, you (the attacker) and they (the defender) compare temple strength. Each side counts the red temples orthogonally adjacent to their leader, then attacker secretly commits any number of red tiles from hand as reinforcements; defender sees the bid and may counter with their own red tiles. Highest sum wins; ties go to the defender. The losing leader returns to its owner's hand. Red tiles bid are discarded. The winner gets one red cube. Revolts are cheap, brutal, and decisively about red — meaning red tiles in hand are always potential ammunition, never just scoring fodder.

External conflict — the war. When a tile placement merges two kingdoms each containing a leader of the same color, every shared color triggers a war (one at a time). Now the tally is the count of all tiles of the warring color in each kingdom (not just adjacent — the whole kingdom). Both sides reinforce by committing tiles of that color from hand. Loser's leader returns home AND every tile of the warring color in the loser's former kingdom is removed from the board. The winner takes one cube of the warring color per tile removed. Wars are massive — a 6-tile vs 4-tile blue war can dump 6 blue cubes into one player and erase a continent's farmland. Wars are the campaign-deciding moments; revolts are the quiet sabotage.

Monuments. Place the fourth tile of one color into a 2×2 square and you may flip the 4 tiles and erect a monument that color (they come pre-printed with two colors; pick one matching). Each turn the monument's owner passively receives a cube of each of the monument's two colors as long as a matching leader is in the same kingdom. Monuments are engine pieces — slow, durable, and disproportionately swingy because they trickle a color you might be starving for.

Treasures. The 10 starting red temples each carry a tan treasure cube. Whenever a kingdom contains 2+ treasures and a green trader, that trader's owner takes all but one of those treasures (corner-temple treasures must be taken first). Treasures are wild cubes. The trader is therefore the most strategically important leader for managing your bottleneck color.

End game. Triggers when fewer than 2 treasures remain on the board OR a player cannot draw because the bag is empty. Then everyone scores their lowest color and the highest of those wins.

Sources: Z-Man rulebook PDF, UltraBoardGames rules, UltraBoardGames conflicts page, edp.org rulebook PDF.

2. What makes T&E work as a tabletop game

The "lowest is your score" aesthetic. Almost every other Eurogame says "specialize and snowball." Knizia inverts the gradient. The single best tactical move is often not to take the cube your kingdom just generated — because if you're already heavy in that color, the cube is worthless to you and merely concentrates your opponent's attention on a fight that doesn't hurt you. This produces a constant low-grade anxiety: your bottleneck is visible to everyone, and the table can see exactly where to hurt you. There is a famous board-state moment — common enough that BGG threads name it — where you realize "I haven't been red for five turns, I have 1 red cube, and someone just dropped a priest next to my temple stack." That sentence summarizes the entire dramatic arc of the game.

Conflict pacing — the slow simmer, the sudden volcano. The first ten minutes of a four-player game look almost peaceful: small kingdoms forming on opposite sides of the board, each player tending their own colors. Then someone places a black tile that connects two kingdoms and the whole table inhales. Wars in T&E are not skirmishes — they are cubes per casualty, scoring 4-8 cubes in one bid, plus board-clearing erasure. The revolt is a needle, the war is a meteor. Knizia's tempo design — that you can only act twice per turn, and reinforcements come from a hand of 6 — means wars are committed moves: spend your blue tiles on a blue war and you have nothing left to score blue cubes for the next two turns. Every conflict is paid in the currency you most need.

The "cold euro" reputation — and why it's wrong, but feels right. T&E famously has zero theme veneer. Mesopotamia is a label; the rivers are just lines that constrain blue placement. There is no narrative beat, no flavor text, no decision that says "you are building a civilization." Reviewers routinely call it "abstract," and Knizia himself explicitly designs to invoke the decision-process of a theme rather than simulate the theme. ("Knizia's thematic game designs tend not to try to model a specific environment, but instead try to invoke the thought and decision-making processes that are key to the theme." — Wikipedia / Knizia interviews) The result feels cold to new players and reads warm to veterans, who recognize that the empires-rising-and-falling narrative is fully embedded in the cube totals — you just have to learn to read them.

Knizia's design philosophy — the puzzle-y nature. Knizia is a math PhD; his stated aim is mechanical elegance in the chess sense — minimal rules generating maximal emergent decision-space. His most-quoted line is "When playing a game the goal is to win, but it is the goal that is important, not the winning." The implication runs through T&E: the game is about navigating the chase of the lowest-color goal, not about feeling clever for executing a combo. You cannot "build an engine" in T&E in the Dominion sense — every turn you must re-evaluate which color is bleeding and which is overflowing. It is, essentially, a real-time multi-objective optimization rendered as wood.

Why it's beloved despite intimidating new players. BGG places it in the top 50 all-time (rating 7.6+, weight 3.4/5). Reviewers consistently flag two things: (1) the "no two games feel alike" emergent property of the lowest-color rule, and (2) the tactical denseness of every single tile placement. You are never bored on your turn and rarely bored off it (because watching someone walk into a war is theater). The retention rate is unusually high for a 28-year-old design.

Sources: Meeple Mountain review, Talking Shelf Space review, Wikipedia — Reiner Knizia, BGG — Tigris & Euphrates, Wargamer reprint coverage.

3. Existing digital attempts

The PC version (Sierra, ~2005). The often-cited "Sierra 2005 Tigris" is a small piece of fog. What actually shipped was a Java/web implementation (later hosted on gametableonline.com) and an early-2000s effort by Dartmoor Softworks that was distributed as shareware — not a Sierra-published retail box. Sierra's name attaches because of confusion with the broader Sierra/Vivendi-era licensing of European designer titles. There is no surviving, supported, native PC Tigris & Euphrates today; the Java version is dead.

The Codito iOS / Android port (2011 iOS, 2013 Android). Codito Development released a polished single-player + pass-and-play adaptation: full rules, three AI difficulty levels, in-app hints, decent visuals. Reviewers (Liquid Hip, Pixelated Cardboard) praised it as one of the best digital Knizia adaptations ever made. It was delisted from the App Store on December 31, 2020 when the Knizia license expired, alongside Codito's Medici and Ra. The Android version lingered briefly afterward but is effectively abandonware. There is no live Tigris & Euphrates app on iOS in 2026.

Board Game Arena. The browser/asynchronous-only port on BGA is excellent and heavily played, but it's a faithful multiplayer port — no AI, no campaign, no roguelite, no native mobile experience.

Tabletopia. Has a digital edition built for online multiplayer; same caveats as BGA.

Lowest-of-N scoring on mobile — anywhere? Essentially nowhere. Highest-Lowest scoring is a recognized BGG mechanism (BGG mechanic #2889) and Tigris is widely credited as the first significant implementation. Beyond Lost Cities (also Knizia, where each suit can score negative), almost no digital-native mobile game uses this scoring shape. The closest cousin is Balatro's "lowest hand chip count" jokers — but that's a single tweak inside a sums-based engine. The primary scoring rule of "min across N categories" is a wide-open design space on the App Store.

Spatial conflict resolution as a pattern on mobile? Civ-style 4X exists (Polytopia, Civ Revolution mobile), area-control exists (Risk-likes), and tile-laying exists (Carcassonne, Patchwork). What is missing on mobile is the specific combination of tile placement + leader pieces with adjacency-based conflict bidding + shared-bag resource draw. Polytopia's combat is unit-vs-unit; Carcassonne's meeple is non-combative; even Through the Ages on mobile is card-based. The "tile-driven kingdoms with leaders triggering bidding wars" pattern is uniquely T&E.

Verdict: Tigris-the-IP is digitally unclaimed at the licensed level (since 2020) AND the genre slot — lowest-of-N tile placement with conflict bidding — is a true vacuum on mobile.

Sources: VideoGameGeek "iOS app disappearing soon" thread, VideoGameGeek "Android app abandoned" thread, Codito Facebook announcement, BGG Highest-Lowest scoring mechanic, BGA Tigris page, Old PC Version BGG thread.

4. The translation problem

Solo with AI vs. fierce 4-player. T&E shines because four humans see and exploit each other's bottleneck colors socially — there's eye contact, quiet table-talk, the moment everyone realizes who is short on red. Stripping to single-player loses that table-read entirely. The fix is not "smarter AI" — it's a structural reframe: instead of three opponents trying to bottleneck you, the board itself becomes adversarial. Procedural events (catastrophes that fire on a timer, a "raiders" mechanic, kingdoms that "decay" if neglected) make the map the antagonist. Two named AI dynasties at most, each with a clear personality (the warmonger, the monument-stacker), provide enough legible threat without needing four-way social dynamics.

The lowest-color-scoring UI problem. On a table you see everyone's cube piles. On a phone, you have one player and need to surface (a) your own four totals, (b) the AI's four totals, (c) which color is in critical short supply right now, and (d) what color the next tile placed will award. The spec needs a permanent color-bottleneck heads-up display at the top of the screen — four colored chips with your current count, your lowest highlighted in red, and a "predicted next cube" indicator that shows which color will benefit from the tile you're about to drop. Without that HUD, the lowest-color mechanic becomes invisible and the strategic core collapses.

What gets ADDED as the digital meta-layer (the roguelite spine). Pick 2-3:

Civilization picks (run-start). You begin each run by choosing 1 of 4 starting civilizations, each bending the rules:
- Sumer — your trader (green leader) collects 2 treasures per trigger instead of 1 (better wild generation).
- Akkad — your priest (red) wins all revolt ties (aggression-friendly).
- Babylon — when you build a monument, it grants cubes of 3 colors instead of 2 (engine-friendly).
- Assyria — your lowest-color score is doubled but your highest is halved (high-variance bottleneck-or-bust).
Relics (between-game drafts). After each completed game you draft 1 of 3 relics — passive modifiers like:
- Hanging Gardens: catastrophe tiles return to your hand after each round.
- Cuneiform Tablet: see the top tile of the bag before drawing.
- Ziggurat of Ur: your monuments grant a treasure cube on construction.
- Code of Hammurabi: your scoring rule becomes "lowest of 3 colors instead of 4" (you choose which to drop).
Map seeds & boss kingdoms. Each run is a 5-game arc on procedurally seeded boards (river layout, starting temple positions, treasure density). The 5th game is a "boss board" — e.g., a hostile rival dynasty that starts with double monuments, or a famine map with no blue farms in the bag for the first 10 turns.
The Code of Hammurabi pivot (mid-run choice). Once per run, you may change what your lowest-color rule means — drop a color from your scoring entirely, or swap to "lowest two summed." This is the meta-defining choice that makes each run feel different.

Recommendation: ship v1 with #1 + #2 + #3. Civilizations give run-to-run variety, relics give game-to-game variety, the boss board gives each run a climax. The Code of Hammurabi pivot (#4) is the most heretical-to-fans mechanic and should be added in v1.1 after validation.

What CANNOT translate.

The mid-game social staring contest where everyone realizes player 3 is short on green and silently agrees to attack their trader. AI cannot replicate this without feeling either telegraphed (you see the conspiracy coming) or unfair (the AI cheats with hidden info). Replace with explicit "rival has marked you" notifications — it's worse but legible.
The 90-minute slow-burn pacing. A solo session needs to be 8-15 minutes per game, 30-60 minutes per run. That means cutting the bag (~80 tiles instead of 150), cutting the hand to 5, and capping turn count.
Kingmaker dynamics (where a losing player decides who wins by attacking the leader). Solo strips this entirely; not a loss.

5. Concrete iOS prototype spec

Target: iOS portrait orientation (with optional landscape for tablets), single-developer Unity or SwiftUI+SpriteKit, 5-7 month MVP. Premium one-time purchase.

60-second core loop. Tap tile in hand -> board ghosts highlight legal placements -> tap a hex to commit (action 1 of 2). Cube animates from board to your tally. Tap a leader from your tray -> highlight legal placements -> tap to commit (action 2 of 2). If a placement triggers a revolt or war, a modal pops with a slick "bid your reds" interaction (drag tiles from hand into a bid slot, AI counter-bids with face-down tiles flipped on resolve). End turn auto-draws to 5. AI rivals take their turn with a 1.5-second animation showing their move on the same board. Repeat. Game ends when bag empties (~30-40 player turns ≈ 8-12 min).

Single-screen layout (6.1" portrait reference):

Top 12% — bottleneck HUD. Four color chips with your cube count, lowest pulsing red. Toggle button to overlay opponent's totals.
Top middle 8% — bag/turn counter + monument & treasure tally on board.
Middle 55% — the board. A roughly 11×16 hex grid (the original is 11×16 squares — hex reads cleaner on mobile and avoids 4-way ambiguity). Pinch to zoom, two-finger pan; default zoom auto-frames the active kingdom. Rivers are pre-painted blue lines along the lower third; only blue tiles snap into river hexes.
Bottom 25% — hand + leader tray. Your 5 face-up tiles in a fan; your 4 leaders in a compact row beside them; "End Turn" button in the bottom-right. Long-press a tile or leader for tooltip.

Tile system.

Starter hand of 5 tiles drawn from a per-run bag of ~80-100 tiles in the four colors (proportions tunable per civilization).
4 leaders in your tray (place once, then they live on the board until killed/withdrawn).
2 catastrophe tiles per game (they sit in a special slot, not in your hand).
Drag-up on a tile shows possible placements; commit is a second tap on a hex.

Run structure.

1 "run" = 5 successive games on linked procedural boards.
Each game ≈ 8-12 minutes; full run ≈ 45-60 minutes (sessionable in 2-3 sittings via auto-save).
Win condition per game: highest lowest-color score after bag empties. Lose a game and lose 1 of 3 "tablets"; 0 tablets = run over.
Win condition per run: clear all 5 games. Final game is a boss board with a named rival dynasty.
Failed runs unlock new civilizations, relics, and rival dynasties — classic roguelite meta.

Meta-progression.

4 starting civilizations -> unlock to ~8.
Relic pool of ~30, draft 1 of 3 between games.
Rival dynasty roster of ~8 named AI personalities (the warmonger, the monument-stacker, the treasure hoarder, the catastrophist).
~20-40 hours to unlock everything; daily seeded run with global leaderboard for retention.

Monetization. Premium one-time, $7.99-$9.99. No energy, no ads, no IAP. Optional $4.99 expansion pack DLC at month 6 (new civilizations + new relics + new boss boards). The mental price-anchor is "Slay the Spire mobile" ($9.99), not "Carcassonne" ($4.99).

Aesthetic direction. Lean into Mesopotamian historical illustration — terracotta/lapis/ochre palette, cuneiform-inspired UI iconography, stylized but not cartoony tile art (think the watercolor feel of Through the Ages mobile, not the chunky meeples of Carcassonne). Music: low ambient drones with hand-drum percussion that picks up tempo during conflicts. Diegetic sounds: tile snap on placement, distant horn on war trigger, treasure chime on collection. Avoid the trap of going modern-abstract — the historical romance is the only warmth this otherwise-dry game has, and the digital version should lean into it harder than the physical one.

6. Risks and unknowns

Top design risks.

Lowest-color scoring is hard to teach. First-time players will end their first game with a high cube total in three colors and zero in the fourth and feel cheated by the "score = 0" reveal. Mitigation: a forced 5-minute interactive tutorial that makes the player score a 0 in the third tutorial game so they internalize the rule before it costs them. Permanent bottleneck HUD (see §5). End-of-turn "your lowest is red, +1 red would be your most valuable cube" coaching toast — toggleable off for veterans.
Conflicts are intimidating. Revolts and wars have cascading rules (count adjacents, then the whole kingdom, then bid, then resolve, then potentially trigger another war from the merge). On a phone, walls of UI text are poison. Mitigation: every conflict resolves through a single dedicated modal with animated diagrams ("you have 4 adjacent temples, opponent has 3, bid more reds to overwhelm"). No text the player has to parse during a fight.
AI quality is the project. A bad T&E AI just dumps tiles randomly and the player wins by accident. A great AI bottleneck-attacks the player's lowest color, builds toward monuments, and feels personally hostile. This is the single biggest engineering risk. Mitigation: invest in heuristic-driven AI personalities early, validate at greybox stage, do not ship until the AI can beat a competent human ~30% of the time on default difficulty.
Procedural boards may produce degenerate states. A board with no rivers in one quadrant starves blue; a board with all temples clustered makes treasures trivially harvestable. Mitigation: constraint-solver-based procedural generation with hand-validated archetypes.
Run length vs. mobile attention span. A 45-60 minute run is long for mobile. Mitigation: aggressive auto-save, "single-game mode" outside the roguelite for short sessions, fast-forward toggle on AI animations.

What needs validation, in order.

Tabletop solo first. Play the physical T&E with 1 human + 2 dummy hands following a simple heuristic for 5+ games. Confirm the core feel survives without three live opponents. If it doesn't, you need an even more aggressive structural reframe than the spec proposes.
Paper prototype the bottleneck HUD. Mock the four-cube heads-up on an index card and play a normal game with it visible. Does it preserve the "I'm short on red" anxiety, or does it make decisions trivially obvious?
Greybox digital prototype: one game, one AI rival, no roguelite. Validate that 8-12 min/game pacing holds on a phone, and that the conflict modal doesn't break the flow. Only then build the meta-layer.

Three questions for the prototype builder.

How comfortable are you bending the lowest-color rule? The Code of Hammurabi mechanic ("score lowest-of-3") is the single most controversial proposal in this spec — Knizia purists will revolt. But it may be the lever that makes solo runs feel meaningfully variable. Do you treat the original rule as sacred or as a starting point?
Do you build the AI yourself or license/clone the Codito heuristics? Codito's defunct AI is the existing best-in-class and (since it's delisted) potentially studyable behavior for inspiration. A from-scratch AI is months of work; mimicking Codito's style is a force multiplier. Which appetite do you have?
Premium $8 or live-service? A premium one-shot ships in 6 months and earns $40-100k year 1 if it lands. A live-service model (seasonal civilizations, daily seed leaderboards, cosmetic dynasties) is 14-18 months and a different financial bet. The spec above fits premium cleanly; pivoting later is hard. Pick a side now.