Lessons for my own (Godot) games

The Battle Network series ships some of the most-stealable mechanics in handheld history. A surprising number of indie games have ported pieces of it; almost no one has assembled the whole stack.

Steal: pause-to-draw inside real-time combat

The Custom Gauge is the single most-original mechanic the series produced. A real-time-replenishing card-draw step that's also a tactical resource you attack and defend.

This is portable to almost any real-time combat game where there's a "skill / spell / ability" deck to draw from:

• Action roguelite where you draw new abilities every N seconds.
• Survivors-likes where you draw new weapons mid-run.
• ARPGs where active skills are drawn from a small hand rather than a fixed bar.

Almost no one has tried this in modern indie design. Don't separate menu time from combat time — embed the menu as a tactical resource in the action.

Steal: NaviCust as a genre-portable grid-inventory pattern

NaviCust is the cleanest grid-inventory implementation in the knowledge base. Tetris-shaped buff blocks on a small grid, adjacency rules, opt-in chaos via Bugs.

Direct copies in modern indie are rare. It's been re-discovered in Sparklite (patch board), Moonlighter 2 (backpack puzzle), Risk of Rain 2 (item interactions, less spatial), and various others — but BN3 shipped this in 2002.

For my own games: a 4×4 grid is enough. The constraints carry the design. Don't think you need a sprawling Sparklite 5×5 board — start with 3×3 and let the placement rules do the work.

Steal: opt-in chaos via Bugs

NaviCust's killer feature isn't the grid — it's that violating the rules doesn't reject the placement. It applies the buff and a debuff. Players can opt into chaos to fit more.

This is universally applicable to any constraint-satisfaction system. Loot games where you can over-equip and pay a stat penalty. Spell-loadout games where you can cram in one more ability and accept randomness. Allow rule-bending with a cost; don't enforce the rules with rejection.

Steal: code-constrained deckbuilding

BN's letter-code rule is a combinatorial constraint on co-play. Two chips combine if they share a name or share a code. That single rule determines folder shape.

This is portable to any small-hand card game. Instead of mana costs (MTG) or always-stackable (Hearthstone), give cards tags and let cards that share tags combine. A simple combinatorial rule replaces mana curves entirely.

For my own games: code-as-constraint is a much smaller mechanical surface than mana curves. Easier to balance, easier to teach.

Steal: yearly iteration on a frozen core

Capcom shipped 6 BN games in 5 years on the same engine. The combat shell never changed; the meta-progression layer iterated every year.

Implication: decide what your "frozen core" is. What part of your design must not change between iterations? For BN it was 6×3 grid + Folder + Custom Gauge. For my games, what's the equivalent? Pick it deliberately. Don't iterate the core; iterate the layers above.

Steal: Beast Out as bonus-with-forced-fallback

Beast Out (BN6) is one of the cleanest bonus-with-drawback implementations I've seen. The trick is that the drawback is forced, not optional. After your super-form runs out, you're locked into a worse state for at least 1 turn.

This produces a different feel than "+stat / -stat" trades. You commit to the burst, you eat the consequence. The decision is "burst now or save it for later" — a real timing choice, not a math problem.

For my games: when designing super-forms or burst abilities, give them a temporal aftermath that makes them not freely repeatable. Spire boss relics do this with permanent costs; BN does it with timed punishments. Both work; pick the one that fits.

Steal: depth that doesn't punish casual play

BN's depth ceiling is genuinely high. Folder optimization, PA recipe memorization, NaviCust packing, transformation system mastery. None of it is required to finish the campaign. The depth is opt-in for postgame.

This is rare. Most ARPGs (PoE) gate their endgame on deep system mastery. BN doesn't. Casual players can complete the campaign with a haphazard Folder; deep players unlock postgame superbosses with optimized everything.

For my games: separate "complete the game" from "master the systems." Don't gate completion on deep play. Let the deep play be its own opt-in reward layer.

Cautionary lesson: yearly iteration produces messy middle entries

BN4 is universally regarded as the franchise low point. Forced multi-playthroughs, unbalanced Souls, Dark Chip morality system that didn't quite work. BN5 was structurally weird (Liberation Mission mode took half the runtime). BN6 had to clean up a lot.

If you commit to fast iteration, expect bad ideas to ship. The benefit of the cycle is fast feedback — but the cost is that some entries will be remembered as the bad one. Plan for this. Don't pour every BN4-equivalent's design budget into a "definitive" version that won't get superseded.

Cautionary lesson: hidden knowledge depth has a community gradient

PA recipes and Compression Codes are great for engaged players. They're terrible for new-player onboarding. The community knows things the game doesn't tell you, and players without access to that community are at a permanent disadvantage.

For my games: if I'm shipping hidden-knowledge depth, I need to think about how new players discover it. Maybe in-game NPCs hint at recipes. Maybe a dedicated codex unlocks postgame. PoE delegates to NeverSink; BN delegated to magazines and now wikis. Pick a discovery path or accept the new-player gradient.

Cautionary lesson: combining genres needs a dialectic, not a list

BN's "action + card" framing was a dialectic — both halves were load-bearing, and the Custom Gauge made them inseparable. That's why it works.

A genre fusion that's just "this game has cards AND it's an action game" wouldn't have the same structural cohesion. The fusion needs a single mechanism that requires both halves at once.

For my games: if I'm tempted to combine two genres, ask what the single mechanism is that makes the combination unavoidable. If I can't name it, the fusion will feel like a list of features, not a coherent design.