Lessons for my own (Godot) games

Notes from 1000+ hours. Some patterns are wildly stealable; others only work because PoE has 12 years of context to support them.

Steal: currency-as-crafting

The single most-stealable PoE design move, and almost no one steals it. Most games have gold + materials, where gold has no intrinsic use (only as a trading abstraction). PoE collapses the two: every currency does something, and trade prices are denominated in those same currencies.

This produces:

• Built-in inflation control — trade flows are also sinks.
• Real opportunity cost on every drop.
• Denomination as game design.

If your game has loot and economy, try this. Even a small game can have "every currency is a verb." Roguelite deck-builders, looter-action games, factory games, even cozy games — all candidates. Don't give the player gold; give them materials that flow through the economy as currency and as inputs.

Steal: shared progression tree where class is a starting position

The 1500-node tree is too big to copy literally. The principle generalizes: instead of class-locked trees, give every player the same progression space, with class as a vector into it. Marauder's start position is in the south, Witch's in the north — but the geometry is shared.

This solves "every Sorcerer plays the same" without giving up class identity. You can scale it down dramatically — even a 100-node shared tree is more interesting than a 30-node-per-class tree. The size is for sustaining a 12-year game; smaller games don't need that.

Steal: bonus-with-drawback keystones

PoE's keystones are textbook: Chaos Inoculation, Resolute Technique, Mind Over Matter. Each one flips a fundamental rule with a real cost. Rules as commits, not as stat bumps.

This pattern works in any progression system. Have ~5 of them; make each transformative; make each bind decisively. They become the moments players talk about ("I'm running CI this league") for years.

Steal: trade friction as a real choice

This is the contested one, but worth thinking about. Easy trade is a design choice, not the default. If your game has any economy, ask: does trade help my game, or does it short-circuit it?

For roguelikes, RPGs, and some looter-shooters, trade friction or trade-as-opt-in preserves drop excitement. PoE's separate SSF / trade ladders is one model; "trade for premium currency only" (Warframe) is another. Don't add an auction house just because the genre conventionally has one.

Steal: delegate tooling to the community

GGG ships a filter language. NeverSink ships the actual filter. The community can iterate faster than the studio. This pattern works wherever "the right configuration" depends on live data the studio can't track in patches.

For my games: ship hooks (mod systems, save format, stat APIs), not defaults. If a wiki or forum will inevitably emerge, ship a foothold for it.

Steal: separate ladders, not difficulty options

PoE's HC and SSF as separate leagues rather than a difficulty checkbox. Players who want HC commit; players who don't are unaffected.

For multiplayer / leaderboard games, this is much cleaner than "Hard Mode" toggle. The cost is content duplication — economies, ladders, sometimes balance tuning per ladder. The benefit is that each ladder has its own meta and its own community.

Steal: meta-progression as a build, not a stat sheet

The Atlas Passive Tree is brilliant. The endgame gets its own build. Most games have meta-progression as a stat sheet (perks, levels, points to spend on bonuses); the Atlas tree is a full progression axis that's orthogonal to the character.

For roguelites: instead of "+5% damage forever" as your meta-progression reward, give players a meta-tree that affects which mechanics spawn. Same time investment, vastly more decision space.

Cautionary lesson: friction filters out players you may want

PoE's friction is load-bearing for its 1000-hour audience. It also turns away the 10-hour audience. You're picking which audience you want. PoE chose its long-tail audience; that's an explicit decision, not an oversight. If you copy the friction without copying the deep end that justifies it, you'll just have a frustrating game.

Cautionary lesson: aspirational crafting requires careful calibration

The "ship deterministic system → nerf it" cycle is expensive — every cycle the community burns. GGG accepts the cost because they think the alternative (build crafted items become baseline) is worse. You probably can't afford this cycle if your audience isn't already invested.

For smaller games, pick one stance: deterministic-end (you can build perfect items eventually) or RNG-end (perfect items are pure luck). The middle is hard.

Cautionary lesson: complexity has a tax

PoE's onboarding is famously brutal. Even after 1000 hours I had to look up what some new league mechanics did. The complexity is the depth, but it's also the wall — and PoE2's existence is partly an admission that PoE1's wall got too high. If you ship complexity, ship onboarding too — guided builds, in-game tooltips that aren't just stat names, glossaries, build templates.

Cautionary lesson: nerfs need a clear story

GGG nerfs because of "the Vision." That story is real and consistent, but the community still revolts every league. The story has to be told constantly. Without the manifestos and the explanation, nerfs feel arbitrary. With them, they feel like calibration. Be ready to write the manifesto.

Cautionary lesson: don't ship live-service complexity if you can't sustain it

GGG ships a 3-month league cadence. Each league = new content + balance pass + economy reset. This is a huge engineering and design commitment. A small team cannot sustain this without burning out. PoE2 EA crunch slowed PoE1 league cadence; the community noticed and complained loudly. Picking up the live-service treadmill is a one-way decision.

Compared to my hours in Warframe

The contrast is the most useful lesson. Both games are F2P, both 2013, both hugely deep. Warframe gives you power fantasy; PoE makes you earn power. Warframe's economy is "trade premium currency between players"; PoE's is "currency is also crafting." Warframe is "Mastery as breadth"; PoE is "Build identity through commitment." Both work. The question for my own games is which player I'm designing for — the one who wants to feel godlike (Warframe) or the one who wants to feel clever (PoE). They're different games for different moods.