Crafting and trade

Craft or buy? How to compare the real cost of an upgrade

Compare the cost of obtaining a usable result, not the cost of pressing the craft button once. Include missing inputs, repeat attempts, fees and the time you would rather spend doing something else. Then compare an equivalent item you can actually buy.

By GG Atlas · Reviewed Oct 26, 2026 · About 6 min read

Define the item that would solve your problem

Start with the requirement: the properties an item needs to be useful to your character or project. Comparing a premium crafted result with the cheapest vaguely similar listing will not tell you much. Equally, you may be paying for a perfect property that makes no difference to your immediate goal.

Write a minimum acceptable result and keep it unchanged while comparing options. If you lower the target to make crafting look cheaper, recheck the market for that lower target too. Check progression restrictions and whether trading is available in the mode you play.

For a fixed recipe, count what you must give up

Expand the recipe, subtract the materials you own and identify the missing inputs. Then distinguish the cash or currency you need now from the total resources the item consumes. Materials in storage may require no new purchase, but using them means they cannot be used for another project.

For a decision about time, compare what else you would realistically do in that session. You do not have to assign a precise price to every minute. It is enough to notice that gathering a missing ingredient takes the whole evening when a modest purchase would let you play the activity you wanted.

Illustrative fixed recipe	Cost in game currency
Four materials at 22 each	88
Crafting fee	12
Total crafting inputs	100
Equivalent item available to buy	95

What that small price difference does—and does not—mean

In the made-up example above, buying is cheaper by 5 currency if the offers are available, the items are equivalent and no other costs apply. That is a narrow conclusion, not a rule that buying is always better. Crafting might also provide progression, unlock something or simply be the activity you enjoy.

If you already own all four materials, the immediate currency payment for crafting is only the 12 fee. That does not make the materials worthless, but it can matter when you have little spendable currency or cannot trade them. Keep immediate affordability separate from total input value.

For uncertain crafting, separate expected cost from a spending limit

A simulator needs the actual sequence and success conditions. A single attractive roll chance is not enough if failed attempts consume a base, require repairs or change the odds of the next attempt. Check what is lost, what can be reused and whether the model includes those costs.

For a simplified example, suppose each attempt costs 35 currency and independently has the same 20% chance of producing your acceptable result. With no other costs or salvage, the expected number of attempts is 1 / 0.20 = 5, and the expected spend to first success is 175. That average is not a promise that your fifth attempt succeeds.

After five attempts, the chance of at least one success is 1 − 0.8⁵, or about 67.2%. There is still roughly a 32.8% chance of no success. If an equivalent item is actually available for 140, buying costs less than this model's expected crafting spend and avoids this particular crafting uncertainty. Recalculate if any of the assumptions change.

Put it into practice

Path of Exile 2: check the item before modelling the craft

You want a usable equipment upgrade with a short list of required modifiers. Work out whether your target and method belong together before pricing repeated attempts.

Path of Exile 2 · PoE2DB
Look up the relevant base and modifier information. Check requirements and available modifier tiers for the item you intend to use; do not start a simulation from an impossible target.
Path of Exile 2 · Craft of Exile
Select the PoE 2 version, match the base and item level, and model only methods supported by the tool's current data. Inspect its update notes before relying on a newer crafting mechanic. Record the expected inputs and what each failed attempt loses, then compare an equivalent live trade offer separately.

Do not carry the illustrative 20% chance above into a real craft. Use the actual method's model, and postpone the craft if its mechanics are not represented. A simulated cost is not a guaranteed budget or a live market quote.

Decide the stopping rule before you start

Choose a resource limit you can tolerate losing without getting the result. An expected cost is not a suitable spending cap by itself. A small chance of an early success can make crafting tempting even when the unsuccessful outcome would derail your next project.

Keep failed attempts from rewriting the original comparison. Spending resources already does not improve the next attempt's chance in the independent example above. If the actual game has changing odds or guarantees, use that real rule rather than assuming either independence or an eventual rescue.

Check the market once more before committing

Confirm that the offer is still available, the properties match your minimum target and you are looking at the correct market, platform or season. A historical estimate, advertised listing and completed purchase are different things. Include relevant fees or required intermediate trades where they exist.

The final choice can be to craft, buy, gather first or postpone the upgrade. A useful calculation narrows the decision and exposes the uncertainty. It does not need to turn every play session into an exercise in maximising currency.

I am comparing equivalent usable results.
I counted missing inputs, fees and what a failure consumes.
I understand the model's odds and assumptions.
I set a spending limit independently of the expected cost.
I checked an offer I can actually use, not just a historic price.