Scope
An item's value is not a property of the item. It is a property of the item and the character considering it. The same rare is an upgrade for one character, filler for another, and inert for a third.
This lesson is a procedure for resolving that quickly, and an account of why the intuitive method — comparing the biggest numbers — reliably misleads.
1 · The anatomy
| Part | What it is |
|---|---|
| Base type | The item itself, and what it inherently provides — weapon damage, armour values, attribute requirements |
| Implicit | A modifier present on every copy of that base |
| Explicits | The rolled modifiers. Three prefixes and three suffixes at most on a rare |
| Item level | Determines which modifiers and which tiers can appear |
| Sockets and enchantments | Runes and similar, separate from the explicit modifiers |
Item level is the part most often ignored and the part that decides an item's ceiling. It is not a requirement to equip — it governs what the item was allowed to roll, and therefore what it could still become if crafted. A low item level base caps out low no matter what you spend on it.
Item level comes from where the item dropped. Rare monsters give +1, unique monsters +2, and certain map conditions add further levels (PoE2 Wiki).
2 · Tiers, and what a tier is worth
Each modifier exists at several tiers. A higher tier rolls a larger range of values and requires a higher item level to appear.
The trap is treating tier as a measure of value. It is not.
A top-tier roll of a modifier your character does not use is worth exactly zero, and it occupies a slot that could have carried something that does.
Tier only becomes a comparison tool after relevance has been established. The order of questions is therefore fixed: does this modifier reach my damage, then does it feed a starved total, and only then is it a good roll of its kind.
3 · The procedure
Applied in order, and most items are rejected in the first two steps.
- Does it reach me at all? Lesson 3's question. Minion modifiers on a non-minion character, local weapon damage on a spell character, physical scaling on a fully-converting skill — all zero, regardless of the numbers.
- Which totals do the relevant modifiers feed, and are those totals starved? Lesson 2's division. A modifier feeding a total already at 500% contributes little; the same modifier feeding a nearly empty total contributes a great deal.
- What does it displace? An item is never evaluated alone. It replaces something, and the comparison is the difference between the two, not the item's own contents.
- Does it cost me a cap? Losing resistance cap is expensive, per lesson 1 — the last points are the most valuable ones, so a nominal upgrade that uncaps a resistance is frequently a downgrade.
- What is its ceiling? Open affixes and item level determine whether this is a finished item or a starting point.
Which question should be answered first about an unfamiliar rare?
- Whether its modifiers reach this character at all
- Whether the modifiers rolled at high tiers
- Whether the item level is high enough to craft
- Whether it carries a full six modifiers
4 · Why three relevant modifiers beat one spectacular one
The arithmetic is the additive-total rule applied across several statistics at once.
Consider a character with 500% increased damage, 20% increased attack speed, and one resistance sitting 15 points below cap.
| Item | Contribution |
|---|---|
A — one modifier: +120% increased damage |
120 ÷ 600 = 20% damage |
B — three modifiers: +40% increased damage, +15% attack speed, +15% resistance |
40 ÷ 600 ≈ 7% damage 15 ÷ 120 ≈ 13% speed plus the resistance cap |
Item A produces a single 20% gain. Item B produces roughly 7% and 13% in separate multiplicative terms — which compound rather than sum — and closes a resistance gap whose value we established in lesson 1 as disproportionate near the cap.
Spectacular single modifiers pour into whichever total is already largest, because that is usually why it is largest. Modest modifiers spread across starved totals routinely outperform them.
This is the same principle as "which total is smallest", applied to a whole item rather than a single choice.
5 · Uniques
A unique carries a fixed effect, often unavailable anywhere else, in exchange for giving up the six rolled modifiers a rare would have had.
The real cost of a unique is the rare it displaces — six affixes of resistance, life and relevant statistics — not whatever downside is printed on it.
So a unique earns its slot when its effect is something no combination of affixes could produce: a behavioural change, an unavailable mechanic, a threshold that unlocks something. A unique whose effect is merely large numbers is competing directly against six affixes and usually loses.
5.1 When a printed downside is genuinely disqualifying
Downsides are not equal. The ones that actually disqualify:
- It breaks a gate. Anything that uncaps a resistance, drops you below a survivable maximum hit, or removes a recovery mechanism. Gates are not tradeable against damage.
- It conflicts with your base source. A downside to a statistic your build does not use costs nothing; one to the statistic your base depends on costs everything.
- It forecloses a supply line. Losing a slot that was your only home for a needed statistic — lesson 10's problem — is more expensive than the printed numbers suggest.
And the ones that usually do not: penalties to statistics you never scale, and penalties that look severe but apply to a small share of your damage.
Worked diagnostic
An attack character with 480% increased physical damage, 22% increased attack speed, and all resistances exactly at cap. Currently wearing the amulet on the left.
- +52 to maximum Life
- +34% to Fire Resistance
- +29% to Chaos Resistance
- +18% increased Physical Damage
- +145% increased Physical Damage
- +9% increased Attack Speed
- +41 to maximum Life
The dropped amulet has a far larger damage roll. Is it an upgrade?
The damage difference is smaller than it looks. Swapping loses 18 and gains 145, a net of +127 into a total of 480. That is 127 ÷ 580 ≈ 22% more damage — real, but not the fourfold impression the printed numbers give.
The attack speed is worth comparably much from nothing. 9 ÷ 122 ≈ 7%, in a separate multiplicative term. Together with the damage, roughly 1.22 × 1.07 — call it 30% more damage overall.
But it costs two resistances and eleven life. Resistances were exactly at cap, so this uncaps two of them. By lesson 1's arithmetic, points removed from near the cap are the most expensive ones on the character — losing 34 points of fire resistance takes what lands from 25% to 59% of incoming fire damage, well over double.
So: it is an upgrade only if the resistances can be recovered elsewhere. That is a lesson 10 question — is there a free suffix on another slot? If yes, this is a large gain and worth the rebalancing. If no, it trades a gate for damage, and gates are not tradeable.
Worth noting separately: the dropped amulet has item level 81 and only three modifiers, so it has three open affixes and a high ceiling. Even if it is not an upgrade today, it is a materially better starting point than the worn one — which is the subject of the next lesson.
What to carry out of this
- Relevance first, magnitude second. A top-tier roll you cannot use is worth zero.
- An item is always evaluated as a difference against what it replaces, never alone.
- Three modest relevant modifiers usually beat one spectacular one, because they land in separate multiplicative terms while the spectacular one dilutes into the largest total.
- Breaking a resistance cap is expensive and rarely paid for by damage.
- A unique's true cost is the six affixes it displaces. It earns its slot only for effects affixes cannot produce.
- Item level and open affixes describe the ceiling, which can matter more than the current contents.