For experts · Economics
AdvancedSlashing & accountability
Slashing is the mechanism that gives expert reviews their economic weight. This page explains when it triggers, how the 25% penalty is calculated, and how to appeal.
Last updated
What slashing is
Slashing is the forfeiture of part of your staked VETD tokens as a penalty for reviewing badly. It only applies to votes where you explicitly staked; unstaked votes can still cost reputation but cannot be slashed.
The purpose: make bad-faith reviewing expensive enough that doing it at scale is economically unviable. Any review system without slashing has a trivial attack surface — stuff the panel with compromised reviewers. Staked voting raises the cost of that attack to whatever you're willing to lose.
What triggers slashing
Two conditions can trigger a slash:
- Misalignment from consensus. Your revealed score lands more than 1 × IQR from the median. This is the main trigger and accounts for the vast majority of real-world slashes.
- Enforcement action following a governance vote. If the guild votes to uphold a conflict-of-interest or code-of-conduct complaint against you, the penalty can include slashing in addition to reputation loss.
The math
Slashing only operates on the stake attached to the specific application being reviewed. It does not touch other stakes you have elsewhere in the guild or other guilds.
The formula is:
slashed_amount = stake_on_application × slash_percent
Where slash_percent is determined by your alignment tier(see below). The slashed VETD is transferred to the guild treasury, where it's used to fund future review reward pools — so slashing is net neutral for the guild as a whole.
Alignment classification
| Tier | Distance from median | Reputation | Stake slashed |
|---|---|---|---|
| Aligned Your 78, consensus 80 → no penalty, full reward. | ≤ 1.0 × IQR from median | +10 | 0% |
| Misaligned Your 45, consensus 80 → max penalty, reputation loss. | > 1.0 × IQR from median | −20 | 25% |
The asymmetry between reward and penalty is deliberate. Aligned votes earn +10 reputation; severe deviations lose −20 reputation and 25% of stake. That asymmetry exists because a single bad actor can do more damage to the system than a single good actor can repair.
A worked example
Imagine you staked 20 VETD on a candidate in the Engineering guild. Reveal phase closes and the scores look like this:
- Expert A: 78
- Expert B: 82
- Expert C: 76
- Expert D: 84
- You: 45
The median of the five scores is 78 and the IQR is 6.5. Your 45 is about 5×IQR below the median — well beyond the 1×IQR threshold — so you're classified as misaligned.
Result:
- Reputation: −20 in the Engineering guild.
- Stake slashed: 25% of 20 VETD = 5 VETD transferred to the guild treasury.
- Remaining stake: 15 VETD returned to your spendable balance, subject to the normal unstake cooldown if you want to withdraw it.
- You do not lose the whole 20 VETD — slashing is a percentage, not a full forfeiture.
Appealing a slashing
Slashing is deterministic, but the inputs aren't always clean. A guild admin could have misconfigured the rubric; a technical bug could have corrupted a revealed score; you might have genuinely had a conflict of interest that nobody flagged. For these cases, there's an appeal process.
- Open the finalized application in Vetting → Applications. A slashed review shows a Dispute button.
- Submit an appeal, describing the basis for the dispute. Keep it factual — appeals that read as emotional are routinely dismissed.
- The appeal goes to a guild governance vote. Other experts in the guild review your argument and vote to uphold or reverse the slash.
- If the appeal passes, you get the slashed VETD back and your reputation delta is reversed. If it fails, the slash stands and appealing the same event again is disallowed.
How to protect yourself
- Only stake what you're willing to lose on a single review. If losing 25% would meaningfully hurt you, the stake is too large.
- Stick to your highest-confidence reviews for staked votes. Unstaked votes still earn reputation on alignment and are free to be wrong on.
- Calibrate on unstaked votes first. New experts should avoid staking during their first 10–20 reviews while they get a feel for their guild's rubric and what the consensus actually looks like.
- Take conflicts seriously. Abstain rather than review when you know the candidate. Slashing triggered by a post-hoc conflict-of-interest finding is harder to appeal because the facts are rarely in dispute.