An incremental is about numbers: rebuilding our sushi game's core loop after one playtest
A playtest of Neta said it plays like a simulation: guests with tastes, free ingredients and packs that feel like nothing. One Claude Code session rebuilt the loop around price: ingredients that run out, one try per card, duplicates that level up, and a price that counts up like a Balatro hand. Part 10.
What we asked“A playtest found Neta's core loop is designed like a simulation game, not an incremental. In the player's words: the player only cares about numbers; customer preference doesn't fit; free ingredients might work in Minecraft but not here; packs have no special feeling; making sushi should use up ingredients that take a while to refill, with a menu of three and the goal of making sushi as expensive as possible; a repeated card should upgrade the ingredient; and players shouldn't experiment infinitely, only once for each card they get. DESIGN.md is updated to v3. Rework the game to it and re-balance the whole route.”
What happened
- Read the new brief and the three updated skills, then rebuilt the shared rules: stock that refills, batches, ideas, card levels and the price build-up
- Removed the guests' wishes from all six cities and made each city's twist work with stock and price
- Re-balanced the route city by city with the bots, then played the first 10 minutes in Chromium
After part 9 Neta played from Tokyo to the credits. Then someone played it, and said what nine parts of measurements had missed:
"This is an incremental game. The player only cares about numbers. Customer preference doesn't fit. The ingredients are not consumables: that might work in Minecraft, but it doesn't work here. The packs have no special feeling."
He was right. We had built a restaurant simulation and added an economy on top. This part rebuilds the core loop around one question: how do I make more money per second?
Play the new loop
Follow the chef: invent your first dish, watch the price count up, run out of salmon, open the free pack. Open full screen (saves your game).
What changed
| Before | Now |
|---|---|
| Guests ask for "something fresh and sour" | Guests eat and pay |
| Flavour rules give a grade | Flavour rules multiply the price |
| Every card is an endless supply | Every ingredient is a stock that runs out and refills |
| Try any combination for free | One try (an "idea") for each card you pull |
| Duplicates turn into pearls | Duplicates level up the ingredient |
| Five-card packs | Eight-card packs that end with your income jump |
1. A price that builds up
Balatro is the best recent example of making numbers feel good: every hand is chips × mult, and the flames around the score grow until they shake the numbers (GamesRadar). Neta now does the same for a dish. The topping sets the base, then each rule that fires adds to the multiplier, one line at a time: a nigiri that fits the topping +0.2, "acid cuts fat" +1.0. Salmon at 4.2 with yuzu becomes 4.2 × 2.2 = 9 coins a plate.
2. One idea per card
Free experiments turn discovery into brute force. Now inventing a new recipe costs one idea, and every card you pull gives one, up to 8 saved. Making a recipe you already know costs no idea, only stock. So a pack of 8 cards is also 8 tries.
With so few tries, a blind guess would feel like a wasted card. So before you spend an idea, the board shows the price predicted from the rules you already know. Rules you haven't found yet still fire, as a surprise bonus on top: above, the player expected 5 and got 9.
3. Stock that runs out
A sink should eat something, not only give something (dev.to). Now each topping holds 12 portions and refills one every 7 seconds, and rarer fish refill more slowly. A batch is 6 plates. The menu holds three dishes and sells the most expensive one first, so your best dish runs out and your second best earns while it refills. Variety comes from the numbers, not from guests' tastes.
Three new upgrade lines work on stock: refill speed, stock size and batch size. Later, a manager re-makes sold-out batches for you.
4. Packs that move your numbers
In Clash Royale, extra copies of a card are how it levels up, and each level makes it about 10% stronger (Game Developer). Neta now does the same: a duplicate levels up the ingredient, each topping level is ×1.3 on every dish that uses it, and it adds stock and refill. Levels need 1, 2, 4… copies, so common fish keep levelling while rare ones level slowly.
Packs got bigger and rarer: 8 cards, the first always a level-up for the fish of your best dish, the last Rare or better. And every pack ends on what it did to your numbers:
Balance
Bots played the whole route 8 times. Every run reached the credits.
| City | Median time to its goal | Top dish's share of plates |
|---|---|---|
| Tokyo | 34:35 | 36% |
| Osaka | 35:56 | 25% |
| Toyama | 30:29 | 53% |
| Kanazawa | 26:55 | 44% |
| Sapporo | 31:44 | 42% |
| Fukuoka | 31:50 | 57% |
| The whole route | 3:12 |
The target was 40–60% for the top dish: one dish clearly best, but not the only one that sells. Four cities hit it. In Tokyo and Osaka the menu spreads wider than that.
What it cost
One session: 3 hours 12 minutes and $37.04, the longest and most expensive part so far, because it re-balanced all six cities one after another. It ran out of time before its last full bot run, so we ran that ourselves.
Still weak: the cards show finished sushi, so a card looks just like the dish you make from it, and the pack screen still has a leftover "pearls" panel.
Next: cards that show the fish
A card should show the whole tuna, and sushi should only appear when you make it. The fish is cut onto the rice, and the finished piece is the reward.
Make your incremental about numbers
- Copy the skill below into
.claude/skills/sushi-crafting/, or download it. - Have someone play your game, and write their words into the prompt as they said them.
- Ask for a second limit besides coins (stock, energy) and check that the best option doesn't take every sale.
The skill that made it work
The crafting skill from part 3, rewritten after the playtest: stock instead of free ingredients, one idea per card, a price that builds up, and duplicates that level up. Copy the rules below, or download it with its references.
- Rules, not a table. Each card has a few axes (umami, fat, acid, salt, sweet, aroma, delicacy, texture) and 2–4 tags from a fixed list of about 12. Every synergy and clash refers only to tags and axes, so the player can learn it once and use it everywhere. At most about 5 hand-written special combos, as rare secrets.
- A dish is small. Form + rice + one topping + up to two accents (a preparation, a sauce or a garnish). Like Cafeteria Nipponica's base plus one or two ingredients, or Atelier's three trait slots: few slots, many cards.
- Bounded score. Quality is 0–100: Taste (40) + Harmony (40) + Form fit (20). Price multipliers (dish level, customers, shops) live outside it, in the economy. Grades read like letters (D to S) and stars.
- Clashes cost, empty slots don't. An empty accent is 0; a clash is −10 to −20. Diminishing returns: the second copy of a tag counts half, the third nothing. Fat and umami past their cap turn "heavy" unless something cuts them.
- Ingredients are stock, not unlocks. Each owned card gives a stock of portions that refills over time (its level raises both the cap and the refill rate). Making a batch of a dish uses one portion of each ingredient per plate; the dish sells until its batch runs out. Refill speed, stock size and batch size are upgrade lines. Infinite free ingredients make the choice meaningless (Minecraft can do it; an incremental can't): variety must come from the numbers, so when the best ingredient runs dry, the second-best dish earns while it refills. A bad dish wastes stock and sells for 30–50% of the price (Breath of the Wild's Dubious Food).
- Every attempt teaches a rule. After making a dish, show the three sub-grades, a green/yellow/red chip on each card (helps, neutral, hurts) and one line naming the main reason ("Acid cuts fat", "The sauce drowns the delicate fish"). Experiments are limited, not free: a new recipe costs one idea, each card pulled gives one idea (capped at about 5), and repeating a known recipe costs only stock. Infinite free tries turn discovery into brute force; ideas tied to cards make every pack a set of tries. Before committing, show the predicted price from the rules the player already knows; undiscovered rules that fire are a surprise bonus on top, so no try is blind. See [references/feedback.md](references/feedback.md).
- A journal of rules. The first time a synergy or clash fires, add a rule card to the journal ("Acid cuts fat: kabosu, yuzu or sudachi on fatty fish"). Discovered dishes are saved to the menu with their recipe, so repeating one is one tap.
- In an incremental, the player cares about the number. No customer preferences or wishes to satisfy: customers are the crowd that eats and pays, and their reactions scale with the price (a gasp and a big coin for an expensive plate). The goal of inventing is a dish as expensive as possible. Show the price building up like a Balatro hand: base, then each multiplier (ingredient level, each rule that fires, the city bonus) counted up one at a time with a rising pitch, bigger effects for bigger results.
- Names that sound real. `[best-synergy word] [preparation] [topping] [form] (with [accent])`, like "Aburi Salmon Nigiri with Yuzu" or "Zuke Maguro Nigiri". Bad dishes get joke names from a table ("Soggy Regret Roll"). Names are deterministic: the same recipe always has the same name.
- Duplicates level the ingredient. A second copy of a card raises that ingredient's level (Clash Royale, Marvel Snap): more value per plate, more stock, faster refill. Every dish using it gets more expensive at once, so every pack pull raises a number.
- Ground it in real sushi. Each rule cites the chef practice behind it in the data file (Sushi University, Michelin, chef interviews), and the tooltip says it in plain words. Don't invent fake Japanese words.
- Measure it. Score every possible dish and report: the best 10, the share of dishes per grade, that every accent helps somewhere (+10 or more) and hurts somewhere (−10 or less), that every topping has at least two good preparations, and that no single dish beats the second best by more than about 15%. Fail the check if any card is useless or dominant. Check the bench UI in a real browser at 720 px wide and on a phone.
Save it as .claude/skills/sushi-crafting/SKILL.md in your project.
