Balance guide

Balance guide

Generator economy revolves around five knobs per tier:

Knob	What it controls	Bundled curve
tick-interval (ms)	How often the generator fires.	Decreases ~9% per tier.
base-capacity	Loosely interpreted "yield" multiplier.	Increases ~1.55× per tier.
base-durability	Hit budget before corruption (in DURABILITY / HYBRID mode).	Increases ~1.4× per tier.
shop-price	Cost to buy a placeable.	Increases ~2.2× per tier.
repair-cost + repair-cost-per-level	Cost to repair.	Tied loosely to shop-price.

These five values together define how rewarding a tier feels. Tuning any one in isolation usually breaks the curve. This page is a checklist for keeping the relationship sane.

The implicit objective

The shop price for tier n+1 should pay back in a reasonable number of ticks at tier n — empirically around 2,000–5,000 ticks of the previous tier feels right. Faster than 2k feels trivial; slower than 5k feels grindy.

If you halve the bundled tick interval at tier n but don't touch shop-price at tier n+1, you've roughly halved the payback time without intending to. Always think in pairs:

• Faster tier n → bump shop-price of n+1.
• Bigger drop weights at tier n → bump shop-price of n+1.
• Cheaper levels at tier n → bump shop-price of n+1 and level cost of n+1.

Suggested progression curve

If you're authoring a new chain from scratch, this geometric progression is a sensible starting point:

tick_interval(n)  = round(2400 * 0.91^n)        # ~9% faster per tier
base_capacity(n)  = round(500  * 1.55^n)
base_durability(n)= round(75   * 1.40^n)
shop_price(n)     = round(2000 * 2.20^n)
repair_cost(n)    = round(1000 * 2.20^n)

(n is the 0-based tier ordinal — wood = 0, stone = 1, …)

The bundled chains follow this curve; the off-tree generator script tools/gen_chains.py in the plugin source repo writes the files with the same formulas. Adjust the multipliers (0.91, 1.55, 1.40, 2.20) to taste.

Upgrade-cost formula

config.yml keeps the upgrade pricing in two formulas:

upgrade:
  level:
    cost-formula: "500 * level^1.8"
  promotion:
    cost-formula: "10000 * tier"

level is the source level when quoting; tier is the 1-based ordinal of the destination tier in load order. The defaults are deliberately on the steep side — level 10 within a tier costs 500 × 10^1.8 ≈ 31,500, which is enough to make the player choose between levelling-up vs. promoting.

If you raise max-level, the level-cost formula should grow faster (raise the exponent) so the late levels still bite.

Repair-cost formula

config.yml:

corruption:
  repair:
    cost-formula: "12.5 * tier"

Cheap by design — repairs should be an annoyance, not an economy-driver. A diamond tier (tier 8) costs ~100 to repair under the default. Bump the constant if you want repair to feel like a real cost.

Drop weights

The default drops.yml puts the high-weight item (e.g. iron_ingot: weight 80) much higher than the side drops (coal: weight 20). Keep at least one "anchor" drop in every tier's loot table — that's the primary income flow — and use side drops for flavour.

A common mistake is making every drop equally weighted. The loot panel ends up looking impressive (eight items @ 12.5% each) but the average income per tick is much smoother than the player expects, which makes the tier feel like noise instead of progress.

Event scaling

Events stack with the per-tier balance:

• SELL_MULTIPLIER multiplies the final sale price.
• DROP_AMOUNT multiplies the stack size produced per tick.
• SPEED_BOOST multiplies the number of generateLoot() calls per tick (hard-capped at 5).
• DROP_TIER temporarily promotes loot to the next tier's table.

If you author an event with multiplier: 5.0, every income line on every generator gets a 5× window. Long durations (≥10 minutes) plus big multipliers can rewrite your weekly economy — test on a staging server first.

Sell-all unlock cost

config.yml:

sell-all:
  unlock-cost:
    type: MONEY
    amount: 5000

Tune this against your earliest sellable tier. If wood can be purchased for 2,000 and yields ~5/tick at 2.4s, unlocking sell-all at 5,000 takes ~40 minutes of wood-only play. That's about right for the QoL-vs-cost trade-off.

Quick sanity check

Once you've tuned your chain, eyeball these ratios:

Ratio	Healthy range
tier(n+1).shop-price / sum(tier(n) drop revenue per tick × 2000)	0.5 – 1.5
tier.upgrade.level (max-level cost) / tier.shop-price	5 – 20
tier.repair-cost / tier.shop-price	0.4 – 0.8

Numbers outside those bands usually feel "wrong" to players — too trivial, too grindy, or off-pattern.

---

Tip on Ko-fi keeps maintenance going.