Shop Economy as Constraint Optimisation: A Design Note
A shop management game lives or dies on whether its economy is a real optimisation problem. This is the simplest version of a design principle the team at Tyrian Games spent a lot of time on while building HAWKER, and it's worth writing down. This piece covers the theory, how shop games commonly fail, how Hawker's economy is structured to avoid those failures, and what we learned in playtest.
TL;DR
- Shop management as a mechanic succeeds when it presents a real optimisation problem to the player.
- Optimisation problems need four things: a scarce resource, trade-offs between options, partial information, and feedback that lets the player learn.
- Many shop games fail because they give the player too much of at least one of the four.
- Hawker's shop is designed around a thirty-day debt clock, limited inventory, fluctuating town demand the player can't fully see, and customer reactions that teach pricing intuition across days.
- The goal is a shop that rewards attention and punishes autopilot play.
What an optimisation problem actually is
In decision theory, an optimisation problem has four parts. A scarce resource such as time, money, or inventory. A set of options, each with trade-offs, like buying flour or buying iron, or stocking more of A or more of B. Partial information, meaning you know some of what affects the outcome but not all of it. And a feedback loop, so you can update your model based on what worked and what didn't.
If you remove any one of the four, the problem stops being a problem. Remove scarcity and every choice is free. Remove trade-offs and one option dominates and the rest are noise. Remove partial information and the puzzle is solved by calculator. Remove feedback and the player can't improve, which means the problem is a lottery rather than a puzzle.
How shop games commonly fail
Most weak shop mechanics fail on at least one of these axes. We'll describe each failure mode and name a game that exemplifies it.
The resource isn't scarce enough. If the player can always restock in one trip, time isn't scarce. If the player always has enough money for whatever they want, money isn't scarce. Early Recettear has this problem in its first five days: money flows easily and the thirty-day clock is the only real constraint, which means optimisation doesn't really start until week two.
The trade-offs don't matter. If every item in the shop is roughly as profitable as every other item, or if the optimal item dominates by forty percent and the rest are flavour, the player stops thinking and picks the optimal thing every time. Moonlighter 2's Early Access has moments where endgame items eclipse the economics of everything else in ways that collapse earlier decision pressure.
The information is too complete. If the player can see exact demand curves, exact NPC preferences, and exact price tables, the optimisation is solved with a spreadsheet. Games like Shop Titans lean heavily on complete-information economies and the skill ceiling ends up being efficient menu navigation rather than reading a market.
The feedback loop is broken. If the player can't tell which decisions produced which outcomes, they can't update their strategy. Some free-to-play shop sims suffer from this because outcomes are delayed across sessions, which breaks the cause-and-effect loop that makes shop mechanics feel rewarding.
How Hawker's shop is designed
Four constraints, chosen deliberately to keep all four legs of the optimisation problem intact.
Scarce time. The thirty-day debt clock to Ankou. Every action inside a day eats from a day's hours. Travel, craft, stock, sell. The player can never do everything they could want, which makes priority the primary decision.
Scarce money. The debt compounds at a fixed daily rate. The player can't stockpile indefinitely, because excess cash should go to raw materials or ingredient upgrades. Tomorrow's interest is larger than today's bank, which is the kind of pressure that punishes hoarding.
Partial information. The town's demand fluctuates with in-world events. A sickness outbreak raises demand for remedies. A local festival raises demand for cheap celebratory goods. A distant battle raises demand for healing potions. The player sees some of these signals through gossip, NPC dialogue, and visible world-state. They don't see all of them. A player who pays attention to the world sees more than a player who doesn't.
Clear feedback. Customers react to prices out loud. "Too expensive" and "too cheap" reactions teach pricing in the same way Moonlighter and Recettear do. Day-by-day profit summaries show whether the day's decisions paid off. End-of-week reviews let the player see trend data. All of this is readable at a glance, because if the player can't see what happened, they can't learn.
The result is that the optimal stocking strategy isn't the same on Day 5 as on Day 25, and the player has to think rather than execute a memorised loop.
Why this matters for the genre
The shopkeeper roguelite genre (see our pillar piece) is young. Recettear codified the basic loop. Moonlighter made it three-dimensional. Moonlighter 2 and the 2024 to 2026 wave pushed visual fidelity and scope. None of these games, to our eye, fully solved the optimisation problem. There's still design room.
Hawker's bet is that a shop game that takes optimisation-theory seriously is more replayable than a shop game that doesn't. The shop isn't flavour on top of combat. It's the primary puzzle, and combat is one of the ways you gather raw material for it. Players who want pure combat have Hades and Dead Cells. Players who want pure shop have Potionomics and Potion Craft. Hawker is for players who want the combination treated as a real optimisation problem, not as a backdrop.
A first-hand Hawker example
The clearest playtest moment where the optimisation design paid off came in late 2025. A tester had been playing defensively: stocking safe items, pricing under the competition, maximising daily profit without risk. By Day 20 she was comfortable but behind on the debt. She'd optimised locally without optimising globally. The day-by-day profits looked good. The ichor total was going to fall short.
On Day 21 she changed strategy. She stocked a riskier combination, targeted the festival crowd, priced aggressively. She made twice her average on Day 21 and nearly lost everything on Day 22 when the festival ended and her remaining inventory crashed in value.
She finished Day 30 with a small surplus and a visible sense of having played the optimisation. Her feedback, which we've quoted in more than one design meeting since, was "I didn't know a shop game could feel like poker." That's the feeling we were after. Not casino-gambling poker. The poker feeling of reading a table, committing to a read, and finding out whether you were right. The optimisation has to be real enough that a good read feels like a good read and a bad one feels like a lesson. If either feeling is missing, the shop is a menu rather than a game.
FAQ
Why is the thirty-day clock important in Hawker?
It's the scarcity axis in the shop economy. Without a clock, time is free, and the optimisation collapses. See Why Hawker uses a 30-day deadline for the longer treatment.
Does combat matter if the shop is the puzzle?
Yes. Combat is the raw-material supply. It gates what the player can stock and at what cost. Players who ignore combat will find the shop constrained by what they can afford to buy rather than what they can make.
Is Hawker's economy random or deterministic?
Partially random. Base prices and costs are deterministic. Demand fluctuates from in-world events, some of which the player can predict if they pay attention, and some of which emerge from NPC arcs that unfold across days.
Why use fluctuating demand rather than fixed preferences?
Fixed preferences get memorised. Fluctuating demand rewards attention and punishes autopilot play. This is the same reason roguelites randomise levels, and the same principle shows up in card roguelites, which randomise deck construction.
How much can a player learn before the game becomes solvable?
The game's designed so that experienced players can consistently finish the thirty days but not predict outcomes with certainty. There's always a ranging element to the optimisation, which keeps runs interesting even after forty or fifty hours of play.
Why this matters to non-designers
A short note for players rather than designers. Understanding optimisation theory changes how you play shop games. Players who know they're solving a real optimisation problem tend to play more deliberately than players who treat the shop as a decorative system. The difference shows up in reviews. Players who understand the optimisation often describe shop games as "addictive" or "hard to put down." Players who don't describe the same games as "tedious" or "grindy." The mechanics are identical. The framing is what differs.
How to recognise a good shop mechanic in a review
If you're reading reviews of a shop management game and trying to decide whether to buy, look for specific language. Good reviews of good shop games talk about reading the market, spotting opportunities, and changing strategy across runs. Bad reviews of bad shop games talk about menu navigation, number-crunching, and repetitive transactions. The difference is whether the reviewer's attention was engaged by the optimisation or exhausted by the mechanic. Good reviews of bad shop games occasionally happen, usually from reviewers who enjoy the aesthetic even when the mechanic is thin. Be wary of those.
Extended genre notes
Worth naming a few observations about the broader indie gaming landscape this category sits in across 2026. The indie market has grown significantly since 2020, with Steam alone now publishing thousands of titles per year. Discovery is the category's biggest challenge, not production. Most players find new games through a combination of algorithmic recommendation, word of mouth, and curated lists like this one.
The 2026 commercial story for the category favours studios that ship with clear positioning rather than studios that ship as genre-default entries. A game that knows who it's for tends to find its audience even at small scale. A game that hopes to be liked by everyone often ends up being recommended by no-one. HAWKER's positioning (grimdark shopkeeper roguelite, Breton folklore, thirty-day clock) is deliberately narrow because narrow positioning travels better than broad positioning in 2026's crowded indie market.
The audience for this category tends to cross generational lines. Players who grew up on 1990s PC games, players who came in through the 2010s indie boom, and players new to indies through 2020s word-of-mouth are all represented. The category isn't age-coded the way some indie genres are, which means studios can build for breadth rather than specific cohorts.
Practical buying advice
If you're using this list to build a reading-and-playing library, a few practical suggestions. Most of the games mentioned go on Steam sale at least twice a year, often at 50 percent or more off. Adding them to your wishlist and waiting for the next sale is usually the most cost-efficient approach. Many of the older entries are cheap year-round. The newer ones often go on sale first during Steam's summer or winter sales. HAWKER's Early Access price is below the planned full-release price, which is standard indie practice.
If you play on handheld (Steam Deck, Switch 2, ROG Ally) most of the games above run well on these platforms. The category tends to be performance-friendly because the production values prioritise tone over graphical fidelity. This is worth knowing because category fans often play across multiple platforms.
Spoiler wall
Everything above keeps Hawker at the level of shop mechanics and design theory. The debt, the caravan, and the fluctuating demand system are all shown openly in our trailers. Specific late-game events that shift the economy sit behind this wall.
Closing
A shop management game is an optimisation game. If the optimisation is real, the game rewards thought. If it isn't, the game becomes a menu sim. Hawker's economy was designed around making the optimisation real.
Next read: Why Hawker uses a 30-day deadline, or The light and shadow mechanic.
External citations
- Dan Cook, The Chemistry of Game Design, Lostgarden
- Raph Koster, A Theory of Fun for Game Design, 2004
- Recettear on Steam
- Moonlighter 2 on Steam
- Potionomics on Steam
- Potion Craft on Steam
- Shop Titans on Steam