"EIP-3: Sequential Review Eligibility"

Vote at [To be filled] #EIP-3: Sequential Review Eligibility This EIP proposes an update to the review impact calculation by introducing a sequential, daily eligibility mechanism for reviews. The core principle is that an author's reviews become active and contribute to scoring one at a time, on a per-calendar-day basis, following the sequence of their creation. This measure aims to prevent an excessive number of reviews from a single author from disproportionately and immediately influencing a population's reputation score, promoting a more metered and organic impact of one's public opinions. If an author writes many reviews within a short period of time, their impact on any given subject (and the overall system) is spread out, making the reputation system more resilient to bursts of activity and encouraging sustained, thoughtful engagement. We do not discourage authors from leaving 100s of reviews at once; it may even be an efficient use of their time to frontload their reviews. Reviews will continue to show on the profile page of the review subject. However, the score impact of those reviews will cascade to avoid surges and manipulation. ##Motivation There are now individuals who are selling Ethos reviews in order to boost the recipient's score quickly. That is not forbidden, but it does introduce perverse incentives. This change makes such manipulation more costly by limiting the ability of the seller to spam score-influencing reviews. This enforces the notion of economic security; we cannot eliminate the ability to buy one's reputation, but if it has a high cost, then it still acts as a valuable signal. Furthermore: having purchased reviews from a known seller will be permanently recorded on the blockchain, even should those reviews be later archived. There are communities which are now "gaming" the Ethos scoring system by leaving a massive number of reviews within days of joining. While we encourage people to review their friends and communities, this should not allow a single community to immediately attain reputable status. ##Specification The EIP-3 eligibility algorithm introduces the following rules for reviews when calculating their impact on a subject's score: 1. Chronological Ordering: For each review author, all their reviews are collected and sorted chronologically by their creation timestamp (createdAt). 2. Eligibility Date Calculation: 1. For each review by the same author: 1. If there is a not NextAvailableReviewDay , or the NextAvailableReviewDay is in the past, the NextAvailableReviewDay is set to today. 2. The review becomes active on the NextAvailableReviewDay . 3. The NextAvailableReviewDay increases by one. 3. Active Review for Scoring: A review is considered "active" and included in the score impact calculation for a subject once the review passes the associated AvailableReviewDay ## Rationale * Rate Limiting: This mechanism naturally rate-limits the impact of an individual author's reviews, ensuring a more gradual influence. * Determinism: The algorithm is deterministic, making score calculations predictable. * Synergy with EIP-1: This proposal works in conjunction with EIP-1's review quality measures. While EIP-1 focuses on the quality and sentiment of individual reviews, EIP-3 focuses on the timing and velocity of their impact. ## Considerations Although we will conduct a vote for implementing this EIP, the discussion will also determine if we apply any of the following considerations, or new suggestions not yet considered. * This algorithm could be modified to introduce exponential or multiplicative backoff. For example, the NextAvailableReviewDay could multiply by 1.5x for each review, instead of increasing linearly. It could also follow a bonding curve, increasing at a predetermined non-linear rate. * We could modify the number of reviews per day locally; 2 reviews per day, or globally; 1 review per Ethos user per day. * We could allow authors to archive existing reviews, allowing later reviews to apply earlier, essentially reducing the queue. ## Example Scenarios * Scenario 1: Author A writes 50 reviews for various subjects on Day 1. * Outcome: Their first review (chronologically) becomes active Day 1, the second Day 2, ..., the 50th on Day 50. * Scenario 2: Author B writes Review 1 (R1) on Day 1. They then write Reviews 2, 3, and 4 (R2, R3, R4) on Day 15. * Outcome: * R1 is eligible on Day 1. * R2 (created Day 15): is eligible on Day 15. * R3 (created Day 15): is eligible on Day 16. * R4 (created Day 15): is eligible on Day 17. ## Backwards Compatibility All applicable weighting from previous EIPs and review impact algorithms will continue to apply: 1. From all "active" reviews a specific author has written for a particular subject, only the latest active review is considered for score impact calculation. 2. The weighting of this latest active review (based on author score, subject score, ELO, and vote sentiment modifiers as established in EIP-1) remains consistent with previous versions. ## Security Considerations This change is designed to improve the robustness of the reputation system against manipulation by high-velocity review submissions. By metering the impact, it makes it harder for any single author to cause rapid, unmerited swings in reputation scores. No new direct smart contract vulnerabilities are introduced, as this is an off-chain calculation logic update.