Designing a game economy that has to last five years

An economy designed for launch and an economy designed to run for years are different artefacts. The first has to work on day one. The second has to still be adjustable after four years of content, events, and drift.

Most games get the first and discover the difference somewhere around year two, when every change has become risky because everything is entangled.

The property that matters is adjustability

The thing that separates a long-lived economy from a fragile one is not sophistication. It is whether you can still change a value without unpredictable consequences.

Economies become unadjustable when everything depends on everything. A price is set relative to a reward rate, which was set relative to a session length, which was tuned against a progression curve that assumed a particular price. Change any one and the others are wrong, and nobody can predict by how much.

This entanglement accumulates naturally. Each individual dependency was reasonable when created. The sum of them is a system where the safe number of changes per quarter approaches zero.

The design property that prevents it is having a small number of things that everything else references, rather than a web of mutual references. If prices are all expressed relative to one anchor value, changing that anchor moves everything coherently. If each price was set independently against whatever was relevant at the time, there is no coherent move.

Leave room at the top

The second property is headroom. A five-year economy will need content that is more expensive than anything at launch, because players will progress past everything you shipped with.

Games that launch with prices spanning their full intended range have nowhere to go. Everything added later either fits into the existing range — and therefore feels like more of the same — or breaks it, which makes the earlier content look trivial in retrospect.

Leaving the top of the range empty at launch costs nothing and preserves the ability to add genuinely aspirational content later. The same applies to progression: a curve that reaches its maximum at launch has no room for the players who are still there in year three.

The related decision is whether the numbers can grow. Economies where costs and rewards multiply over time — each tier several times the last — run out of meaningful precision and produce absurd numbers within a few years. Economies designed around additive growth stay legible longer but hit ceilings sooner. Which is right depends on the game, and picking accidentally is the common case.

Build the drain before you need it

The third property is somewhere for surplus currency to go that does not depend on progression.

Every long-running economy eventually has players who have progressed past everything and continue to earn. Without a permanent, non-progression use for currency, those players accumulate indefinitely and their purchases stop.

Adding a sink at that point is much less effective than having had one, because it arrives to a population with large balances and gets consumed immediately. A sink that has been running all along absorbs surplus continuously and never faces the backlog.

The sinks that work long-term are the ones that are consumed rather than owned, and the ones that are horizontal rather than vertical — cosmetics, personalisation, anything a player can want repeatedly.

See how we look at long-run economy structure →

The decisions that are expensive to reverse

A few choices are effectively permanent once players have state, and they are worth deliberating at the start even though they seem minor.

How many currencies there are, and what each is for. Adding one later is easy; removing one is a wind-down operation with a trust cost.

Whether balances are capped. Uncapped balances mean a small number of players will accumulate amounts that break assumptions later. Adding a cap afterward takes something from players who have it.

Whether anything expires. Expiry is a powerful tool for controlling supply and introducing it later is read as a takeaway. Deciding at the start that a particular currency is seasonal is fine; deciding it in year three is not.

Whether the economy is per-account or per-device. This one is close to irreversible and it constrains everything about how the game can expand across platforms.

None of these are hard to decide. They are hard to change, which puts them in the category worth spending an afternoon on before launch rather than a quarter on afterward.

Worth noting that none of this requires predicting what the game becomes. It requires leaving room for the parts you cannot predict — headroom at the top of the range, a sink that keeps working after the content runs out, an anchor that lets prices move together. Those are structural provisions rather than forecasts. A team that tries to design the year-four economy in advance will get it wrong; a team that designs a year-one economy which can still be steered in year four will be fine, and the second is a much smaller ask than the first.

Talk to us about designing for the long run →