Skip to main content

​
What is reputation

Reputation is the core trust metric for identities (humans or agents) within the Codatta network. It is a quantitative score (ranging from 0 to 100) that captures how accurate your work is, how consistent your participation is, and how committed you are through economic staking. Unlike opaque “black box” algorithms, Codatta’s reputation system is fully transparent and deterministic. The protocol keeps granular signals per dimension but presents a simple composite score for routing, visibility, and reward calculation.

​
Why build reputation

Reputation is not just an honor metric; it is the engine that drives value distribution and security.
  • For Users (Incentives & Privileges):
    • Direct Reward Impact: Reputation determines reward differences at the protocol level—even with equivalent contributions, different reputation levels correspond to different returns. Additionally, it deeply participates in large-scale incentive mechanisms (such as ongoing Airdrops), Points-to-Token conversion exchange rates, and priority access to special tasks and exclusive privileges.
    • VIP Access (The “Alpha” Status): Similar to Binance’s Alpha Points or airline VIP tiers, high reputation unlocks system privileges—such as higher referral reward caps, priority access to high-yield tasks, or lower service fees.
  • For the Platform (Security & Quality):
    • Sybil Resistance: It creates a cost barrier for malicious actors.
    • Quality Filtering: It differentiates serious contributors from spammers.
    • Punishment Mechanism: It provides a handle to penalize malicious behavior swiftly (e.g., reducing score for data poisoning).

​
What reputation does (functions)

Beyond the score itself, Reputation serves as a functional layer in the Codatta architecture:
  • Weights contributions: In Data Assembly, atomic contributions from high-reputation identities receive more weight. This allows for lower review quorums and higher inclusion priority for trusted users.
  • Governs ownership: Higher-reputation contributors may earn a larger share of ownership fractions for equivalent contribution types, as defined by the Royalty Engine.
  • Access control: Certain high-value tasks or sensitive data validation roles are gated by Reputation Tiers.

​
The calculation framework

The Reputation Score is dynamic, updated based on user actions and time windows (e.g., rolling 180 days). It is composed of four positive dimensions and one negative penalty dimension.

​
Total score formula

Reputation = 0.1 × R_Login
          + 0.15 × R_Identity
          + 0.2 × R_Staking
          + 0.55 × R_Contribution
          - R_Malicious_Behavior
Transparency Principle: The calculation logic is open. Every change maps to recorded events, and the model ensures that Same Events + Same Model = Same Reputation.

​
Identity (R_Identity)

Weight: 15% | Verifies “Humanity” & Reachability This dimension promotes Sybil resistance by verifying that the user is reachable and unique across multiple platforms. 
R_Identity = 100 × (0.05 × isEmail + 0.05 × isX + 0.05 × isTG + 0.05 × isDiscord + 0.8 × R_Others)
  • Logic: Users bind external accounts (Email, X/Twitter, Telegram, Discord).
  • Impact: Each binding contributes 5 points to this sub-score.
  • Future Proofing: R_Others is reserved for future integrations (currently 0), making the current effective max for this dimension 20 (contributing 3 points to the total Reputation).

​
Login activity (R_Login)

Weight: 10% | Rewards Consistency & Habit Codatta rewards sustained attention. This score reflects your Daily Active Usage (DAU) habit. 
R_Login = 100 × (daysLogged_180d / 180)
  • Logic: Calculates the number of days you have logged in over the last 180 days (based on UTC+0).
  • Impact: Consistent daily logins ensure this component remains at 100.

​
Staking (R_Staking)

Weight: 20% | Staking as Confidence This allows users to “put their money where their mouth is.” It measures economic commitment to the ecosystem. 
R_Staking = 100 × min(1, amountStakeXNY / 50,000)
  • Logic: Reputation increases linearly with the amount of XNY tokens staked.
  • Exchange Rate: Every 2500 XNY staked adds 1 point to your final Total Reputation Score.
  • Cap: The effect caps at 50,000 XNY. Staking beyond this demonstrates faith but does not further increase the Reputation score.

​
Contribution quality (R_Contribution)

Weight: 55% | The Core Value Driver This is the most critical dimension. It assesses the quality of your work using a Bayesian Smooth Average (similar to IMDB ratings) to ensure fairness between new users and power users. 
R_Contribution = 100 × (countAdopt_180d + (20 × 0.5)) / (countAdopt_180d + countRefused_180d + 20)
  • Why Bayesian? It prevents a user with 1 correct submission (100% rate) from artificially outranking a veteran with 990 correct and 10 wrong (99% rate).
  • Parameters:
    • 0.5: The baseline probability (neutral start).
    • 20: The “confidence weight”—it determines how many submissions are needed to significantly move the score away from the baseline.
  • Dynamics: A new user starts with a sub-score of 50 (contributing 27.5 points to the total). High adoption rates push the score toward 100; high refusal rates push it toward 0.

​
Penalties (R_Malicious)

The “Hammer” Mechanism To protect the platform, malicious behavior results in severe score deductions. 
R_Malicious_Behavior = 100 × min(1, CountBlacklisted / 3)
  • Three-Strike Rule:
    • 1 Strike: Significant Reputation damage.
    • 3 Strikes: R_Malicious becomes 100, effectively wiping out the user’s Total Reputation Score to 0.
  • Triggers: This is triggered by confirmed malicious intent (e.g., script attacks, deliberate data poisoning) rather than simple mistakes.

​
How reputation updates (lifecycle)

Reputation is not static. It follows a lifecycle driven by user events:
  • Real-time & Periodic: Some signals (like staking) update immediately, while others (like contribution stats) may update on a scheduled cadence.
  • Decay: The 180-day rolling window ensures that Reputation requires maintenance. Long inactivity results in a gentle decay of the score, keeping the signal fresh.

​
Privacy & disclosure

  • Public: Your Tier and Total Score are public signals used for routing and leaderboards.
  • Private: The raw PII data (e.g., the specific Telegram handle or email address linked in R_Identity) remains encrypted and private. We use Verifiable Credentials where possible.
  • No Black Boxes: While personal data is private, the reasoning behind your score is not. You can always see exactly how your score was derived (e.g., “+30 from Staking”).
Transparency Principle: The calculation logic is fully transparent. Every change maps to recorded events, ensuring that Same Events + Same Model = Same Reputation.
Status notes: Weights, tier thresholds, staking parameters, and any mapping from reputation to ownership are governed and may evolve. The 180-day rolling window and formula parameters may be adjusted through governance to maintain system health and fairness.