Vision-driven iteration

Lemma: Balance and content decisions are driven by a stated design philosophy, not by player metrics. The studio's identity is anchored in what the game should be, even when the data says players prefer something else. The cost is short-term player pain; the benefit is long-term coherence.

What it solves

Long-running live-service games drift. Every patch optimizes for engagement; engagement-optimized patches incrementally erode whatever made the game distinctive in the first place. A game that's been running for ten years and chasing metrics is often unrecognizable from its launch.

Vision-driven iteration is the antidote. The studio has a stated thesis about what the game is — articulated in manifestos, GDC talks, dev posts — and balance decisions follow that thesis. Players who deviate from the thesis (find degenerate builds, exploit feel-bad combinations) get nerfed because they're outside the thesis, not because they're popular or unpopular.

This is a hard stance. Players hate having earned builds nerfed. The studio accepts that pain as the cost of identity preservation.

Variants — vision vs. metrics

The cleanest contrast is between two long-running studios that pick opposite stances.

Studio / Game	Stance	Decision-driver	Example
Grinding Gear Games — Path of Exile	Vision-driven	"What fits the Vision; what feels degenerate; what new content the next league will support"	Harvest crafting nerfed in 3.13/3.14 because it became "near-perfect items easily craftable," even though it was beloved
Mega Crit — Slay the Spire	Metrics-driven	"Win-rate and pick-rate telemetry across millions of runs"	Cards / relics patched based on whether they're systematically over- or under-picked

Both studios use telemetry. The difference is what the decision-driver is when telemetry conflicts with vision.

For Mega Crit: data wins. If a card has a 70% win rate, it's overpowered, full stop — adjust until it's in the target band.

For GGG: vision wins. A build with 70% win rate isn't necessarily nerfed because of that — it's nerfed if it's "degenerate," "feels bad," "trivializes content," or "doesn't fit the Vision." The data informs the decision; the philosophy makes it.

When to use this pattern

• Long-running games (5+ years) where identity drift is the existential threat.
• Sandbox / sprawling design space where "right" is a values question, not a number.
• Games with a clear stated thesis — GGG's "the Vision," Bungie's "10-year Destiny," Blizzard's "WoW vanilla feel" debates. The thesis must be articulable to be applied.
• When you have a die-hard audience willing to absorb short-term pain for long-term coherence.

Avoid when:

• The game is a closed-form puzzle with clear right answers (Spire-style — metrics work better).
• Your audience is general-mass and the F2P model depends on individual-build engagement (chase the metrics).
• The studio doesn't have a clearly stated thesis to apply (then you're just nerfing on vibes).

Pitfalls

• The thesis must be public and consistent. If the studio nerfs based on private vision the community can't see, the nerfs feel arbitrary. GGG publishes manifestos before major nerfs explicitly to show their work.
• Visions can be wrong. GGG's commitment to no-AH friction has costs that have become more visible over a decade. Vision-driven studios are not immune to bad theses; they're just more committed to whatever thesis they hold.
• Vision-driven studios alienate players who liked what got nerfed. This is the unavoidable cost. PoE's "Path of Nerfs" community joke is also a community wound — one that GGG accepts as the price of long-term identity.
• Slow community trust building. It takes years for a community to internalize "yes, this studio nerfs hard, but they ship coherent design." New players don't have this context and bounce off the first nerf.

Adjacent patterns

• metrics-driven-balance — the opposing stance. Both work for different game shapes.
• aspirational-crafting — a specific manifestation of vision-driven iteration in PoE: the recurring "ship deterministic crafting → nerf it" cycle is downstream of the thesis that crafting should be aspirational.

Why this is worth its own concept page

This is one of the most-debated tensions in live-service game design, and it surfaces clearly in two of the games in this knowledge base. Studios pick a stance early, often implicitly, and reverse rarely. Knowing which stance a game picks tells you a lot about how it'll evolve.

For my own games: pick a stance deliberately. Don't drift into one or the other by accident. If you ship metrics-driven, build the data pipeline first. If you ship vision-driven, write the manifestos first.