> ## Documentation Index
> Fetch the complete documentation index at: https://docs.notareum.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Slashing

# Slashing

Slashing is the economic deterrent that keeps validators honest. When a validator is found guilty of misconduct through the dispute process, a tier-dependent percentage of their stake is burned and a reporter reward is paid to the whistleblower. This page covers the slash rates, burn mechanics, reporter rewards, and the design rationale. For the full dispute lifecycle, see [Dispute Resolution](dispute-resolution.md).

## The economic logic

Let `P_detect` be the probability of detection for a malicious validator with stake `S` and tier slash rate `gamma`. Expected gain from a single false verification is bounded by the share of the verification fee. Expected loss is `P_detect * gamma * S`. For the system to be incentive compatible:

```
P_detect * gamma * S >> fee_share
```

At `INSTITUTIONAL` tier (gamma = 75%), a 1M NOTA stake, and even 10% detection probability, expected loss is `0.10 * 0.75 * 1,000,000 = 75,000 NOTA`, which vastly exceeds any single verification fee (max 2,000 NOTA). The economics make honest attestation the dominant strategy.

## Slash rates by tier

| Tier          | Slash Rate     |
| ------------- | -------------- |
| BASIC         | 25% (2500 bps) |
| PROFESSIONAL  | 35% (3500 bps) |
| ENTERPRISE    | 50% (5000 bps) |
| INSTITUTIONAL | 75% (7500 bps) |

Higher tiers face larger slashing, matching the larger verification authority and throughput they receive. An `INSTITUTIONAL` validator with 1M NOTA staked stands to lose 750,000 NOTA on a guilty verdict.

## Slash flow

```mermaid theme={"system"}
flowchart LR
    A[Reporter files dispute] --> B[Posts NOTA bond]
    B --> C[Arbitration committee<br/>reviews evidence]
    C --> D{Verdict}
    D -->|Guilty| E[Compute slash amount]
    E --> F[ValidatorStaking.slash]
    F --> G[Tokens to SlashingManager]
    G --> H[Forward to burnAddress]
    G --> I[Reporter reward: 10% of slashed]
    G --> J[Return bond to reporter]
    D -->|Innocent| K[Bond burned]
    D -->|Dismissed| L[Bond returned, no slash]
```

## Slash computation

```
slashedAmount = (stakedAmount * slashBps) / 10000
```

The slashed amount flows from `NotareumValidatorStaking` to `NotareumSlashingManager` and is subsequently forwarded to `burnAddress` (a dead address or token contract with burn capability). Supply is permanently reduced.

## Reporter reward

```
reporterReward = (slashedAmount * REPORTER_REWARD_BPS) / 10000
```

Where `REPORTER_REWARD_BPS = 1000` (10%). The reporter receives `bondAmount + reporterReward` on a guilty verdict. This rewards whistleblowing with real economic upside while anti-spamming the system via bond burning on innocent verdicts.

## Burn mechanics

Slashed tokens MUST be burned. In v1.0, burning is implemented by transferring tokens to a `burnAddress`. The effect is equivalent to permanent supply reduction. The whitepaper specifies three burn scenarios:

* **False verification:** 25 to 75% of stake depending on tier.
* **Repeated offenses:** governance MAY increase effective slash rate up to 100%.
* **Collusion:** all involved validators subject to full slash via separate governance action.

## Tier tracking after slashing

A slashed validator's stake drops. Tier is automatically recalculated. If the drop crosses a tier threshold, the validator is downgraded: lower daily limit, lower reward multiplier, lower future slash rate. This matches the reduced authority to the reduced commitment.

## Interaction with unbonding

In v1.0, slashing reduces `stakedAmount` but does not reach tokens already in an unbonding queue. The 14-day unbonding period still prevents fast exit for most realistic dispute windows, since disputes typically resolve within days of a misbehavior. Future protocol versions may extend slashing to cover unbonding amounts.

## Example: the economics of attacking BASIC quorum

Suppose an attacker wants to falsely verify a resource at `BASIC` level. They need three approving validators at minimum 10,000 NOTA each, totaling 30,000 NOTA of committed stake. At the `BASIC` slash rate of 25%, a single detected false verification costs them `3 * 10,000 * 0.25 = 7,500 NOTA`. At a verification fee of 100 NOTA per request, they would need to successfully pull off 75 false verifications per detection to break even, assuming their share of each is the whole fee. With any non-trivial detection probability, the math turns against them immediately.

## Arbitrator trust

Dispute resolution is gated by `ROLE_SLASHING_ARBITRATOR`. In v1.0 this is centralized to a multisig-held role as a bootstrapping compromise. The arbitration role SHOULD be held by a multisig with a minimum 3-of-5 threshold to prevent unilateral abuse. Future versions will transition to on-chain arbitration via governance voting.

## Related pages

* [Dispute Resolution](dispute-resolution.md) for the full lifecycle
* [Staking and Tiers](staking-and-tiers.md) for stake mechanics
* [Validator Network](validator-network.md) for operator responsibilities
* [Security](security.md) for the broader threat model
* [SlashingManager contract](../smart-contracts/slashing-manager.md)
