Lesson 12 · Durable Core

Crafting Fundamentals

What each category of currency does, how to think about attempts rather than outcomes, and why in a self-found context unspent currency is worth nothing at all.

~15 minutes Conceptual, not recipes Builds on lessons 10 and 11

Scope

There are a great many currency items and they change between patches. Memorising them is not the skill.

The skill is recognising that they fall into a small number of operations, knowing which operation a given situation calls for, and being able to decide in advance how many attempts a plan is worth. That structure survives patches; the item names do not.

1 · The operations

Nearly every crafting currency performs one of these.

OperationEffectExample
Upgrade rarity Change an item's rarity, adding modifiers as it goes Normal → magic; magic → rare; normal → rare directly
Augment Add one random modifier, up to the rarity's limit Adding a modifier to a magic or rare item
Remove Delete a random existing modifier Annulment
Remove and add Both at once — a targeted reroll of one slot Chaos
Reroll values Re-randomise the numbers within their existing ranges, keeping the modifiers Divine
Quality, sockets, enchantment Alter properties outside the explicit modifiers Whetstones, socket currency, runes
Corrupt An irreversible gamble that also locks the item against further crafting Vaal

The limits from lesson 10 govern all of it: a magic item holds two modifiers, one of each half; a rare holds six, three of each. Adding is impossible once the relevant half is full, which is why removal is a real operation rather than a mistake.

2 · Determinism is a separate layer

The operations above are random. A second category of item constrains that randomness — omens, which modify how the next use of a specific currency behaves.

The naming is systematic and worth learning once:

TermMeaning
SinistralActs on prefixes only
DextralActs on suffixes only
GreaterActs twice instead of once
WhittlingActs on the lowest-level modifier rather than a random one

Combined with a base operation, this produces targeting: removing only a suffix, adding only a prefix, removing specifically the weakest modifier. That converts a coin flip into something closer to a decision.

Random currency is cheap and plentiful; targeted currency is scarce. Spend randomness where the item is worthless and determinism where it is nearly finished.

Two more scaling dimensions exist: currency comes in greater and perfect variants that guarantee a minimum modifier level, and item level caps which tiers can appear at all. A high-tier outcome requires both a high item level base and, often, the better currency variant.

3 · Attempts, not outcomes

The productive question is never "will this work". It is "how many attempts am I willing to spend, and what happens when they run out".

Each attempt is independent — nothing accumulates toward a guaranteed result. So the plan must be stated as a budget and a stopping rule, decided before the first attempt rather than during.

A limit worth being honest about

Computing a genuine expected number of attempts requires the modifier pool and its weights for that base at that item level. That data is not reliably published for PoE2. Anyone quoting precise attempt counts is either using datamined tables or guessing.

What follows is therefore relative reasoning — which approach is better than which — rather than absolute numbers. That is enough for every decision this course needs.

3.1 The relative rules

4 · Two economics that are specific to self-found play

4.1 Crafting on a base you would otherwise vendor is free

If an item's alternative use is being discarded, any outcome is an improvement and the only real cost is the currency.

This inverts the usual anxiety. The expensive mistake is not a failed craft on a worthless base — that has no downside at all. It is a failed craft on an item that was already good, where you destroyed something usable.

Practise on trash. The skill of crafting is built on bases where failure costs nothing, and the currency spent there is not wasted — it is what makes the later attempts competent.

4.2 Unspent currency is worth nothing

In a trade context currency has a market value, so hoarding it is a real position. Self-found, currency cannot be exchanged for anything. Its only possible use is attempts.

A stash of unused currency is therefore not savings. It is attempts you declined to take, on a character that would have benefited from them.

The correct posture is to spend up to the point where you would rather have the currency for a specific known future use. Absent such a use, holding it is a strictly worse option than trying something.

Check

Self-found, which situation carries the largest real risk?

5 · When to stop

Improvement on a single slot has the same shape as everything else in this course: each addition lands on a growing total, so the returns flatten.

Stop when the next improvement to this slot is worth less than the first improvement to a different one.

That comparison is the whole rule, and it is the audit loop applied to crafting. A slot with four good modifiers is competing against a slot with none, and the empty slot almost always wins — for exactly the reason an empty additive total does.

Three concrete stopping signals:

Worked diagnostic

Diagnostic

Three items in the stash and a modest amount of currency. Resistances are capped; the weapon is the character's weakest slot by a wide margin.

A — Boots
Rare · item level 68 · worn
  • 25% increased Movement Speed
  • +70 to maximum Life
  • +40% to Fire Resistance
  • +38% to Cold Resistance
  • two open prefixes
B — Weapon base
Normal · item level 82 · unused
  • no modifiers — a white base

Where should the currency go, and what is the reasoning for each option?

The boots (A) are the tempting option and the weaker one. They are already good, they have open prefixes, and adding to them feels productive. But every addition lands on a character whose boots are not the problem — and any operation that can remove risks the resistances holding the cap. Real downside, small upside.

The white weapon base (B) is the answer, for three separate reasons.

  • The weapon is the starved slot. For an attack character the weapon is layer one — the base every multiplier acts on — so improving it multiplies everything. This is the audit loop's answer before crafting is even considered.
  • Item level 82 is a high ceiling. It can roll tiers the level 68 boots never could. Currency spent on a low-item-level base buys a low maximum.
  • It is free to fail on. A white base has no value to destroy. Every outcome is an improvement over an unused item, so the only cost is the currency — which, self-found, has no other use.

The general shape: spend randomness on high-ceiling bases that are currently worthless, in the slot the audit loop already identified. Save targeted currency for the item that is nearly finished. The boots are nearly finished and should be left alone until something targeted is available.

What to carry out of this