Why some corrupted gems cost 100× the uncorrupted version

If you spend long enough looking at the Path of Exile economy data, you'll run into a strange pattern. A common skill gem at level 20, quality 20%, uncorrupted, lists for around 28 chaos. The same gem, same level, same quality, corrupted, lists for 2,700 chaos — sometimes more. Same effective stats. Same functional power in your socket. Almost a hundred times the price.

This pattern is real, it appears across many gems on every league, and it isn't because of any in-game mechanic that adds value to the corrupted version. There's no hidden modifier, no implicit mod, nothing that makes the corrupted gem objectively better. So what's going on?

It isn't corruption implicits

The first hypothesis to rule out: maybe corrupted gems sometimes roll a special implicit mod, like "+1 to Level of Socketed Support Gems," that boosts their value. This is true of equipment — Vaal Orbs on armor or weapons can roll powerful implicit modifiers. But it isn't true of skill gems. A Vaal Orb on a gem produces exactly one of four outcomes: no change, level ±1, quality rerolled, or Vaal-version transformation. None of these adds a random implicit modifier to a regular gem.

So when you see a "Detonate Dead L20 Q20 corrupted" listing at 2,700c, that gem has no special properties. It's just a corrupted version of the same gem you can buy uncorrupted for 28c.

The aspirational-listing hypothesis

The most plausible explanation is selection bias in what gets listed. Consider the seller's perspective. A Vaal Orb has a 25% chance of producing the "no change" outcome — a corrupted gem with stats identical to the uncorrupted one. From the seller's point of view, this is the least interesting result. The gem is functionally indistinguishable from any uncorrupted equivalent on the market, which trades at 28c. The seller could list it at 28c and compete with a deep, liquid market for uncorrupted versions.

Or they could list it at 2,700c — aspirationally, hoping an inexperienced buyer doesn't realize they should just buy the uncorrupted version for far less. Most sellers know perfectly well that 28c is the going rate. They list at the high price because the cost of listing is essentially zero, and they only need to hit on a few impatient or uninformed buyers to make the strategy pay.

The cheap, fair-priced corrupted gems don't really show up on the market — sellers just consume them as their own leveling fodder, or don't bother listing because they have to undercut both the existing corrupted listings and the uncorrupted market. What's left in the listings is the aspirational top, and the price aggregation averages over those listings rather than over actual sales.

What real clearing prices look like

If mass-Vaal-orbing 28c gems actually produced 25% × 2,700c gems in sale-price terms, the strategy would be a free-money exploit. A 25% chance of returning 2,700c on a 28c+1c input is a ridiculous expected return. The fact that nobody is doing this, that the prices don't equilibrate, that the listed corrupted gems aren't getting cleared out by arbitrageurs — all of this strongly suggests that the 2,700c listings aren't actually selling at 2,700c.

What's the real clearing price for a no-change corrupted gem? Probably close to the uncorrupted price. The corrupted version can't be modified further (no more Gemcutter's Prisms can be applied; no more leveling beyond max), which makes it slightly less useful to most buyers, not more. The fair value is plausibly 90–100% of the uncorrupted price, not 100× it.

How this site corrects for the bias

The expected-value calculations on the corruption table would be unreliable if they used the listed corrupted prices at face value. To handle this, the math caps the price of every "non-upgrade" corruption outcome at the uncorrupted input price. Non-upgrade means:

• The "no change" outcome (same level, same quality)
• The "level −1" outcome (lower level than input)
• Quality-reroll outcomes that land at a quality value at or below the input

Outcomes that are genuinely distinct upgrades stay uncapped, because their listings reflect real market activity rather than aspirational pricing:

• The "level +1" outcome (Vaal-corruption-specific, real upgrade)
• Quality reroll to a value above input (Q23 from Q20 input is a real, distinct, valuable outcome)
• Vaal-version transformation (entirely different gem with its own market)

The practical effect: corruption EV on this site is conservative. We explicitly trade a little optimism for a lot more honesty. If you actually run the strategy, you may sometimes hit one of those aspirational sales for a genuinely outsized return — but it isn't something you can plan around or compound, so we don't price it in.

The lesson for traders

The deeper takeaway isn't specific to corrupted gems. It applies to any market on poe.ninja with a low listing count: thin markets reflect what sellers are asking, not what buyers are paying. Any time you see a price that seems mechanically impossible — a corrupted version worth 100× the uncorrupted, or a rare base item priced at multiples of its visibly identical counterpart — the explanation is almost always that very few of those listings actually sell.

The leveling table's High confidence only toggle enforces a minimum-listings threshold on the input side of every row precisely because of this. Below that threshold, the listed price is too noisy to trust. Above it, you can be reasonably confident the displayed number is at least close to what's actually clearing.