Finlance SignalDiscover financial knowledge shaped by community consensus.
QUANTITATIVE FRAMEWORK

Continuous Contributor Reputation & Bayesian Impact Model

Complete mathematical transparency regarding how contributor reputation is initialized, updated via peer evaluations, smoothed using empirical Bayesian inference, and recorded in the verifiable ledger.

1. Continuous Reputation & Genesis Baseline

Every registered user on Finlance possesses a single continuous reputation score that operates like a personal stock price. All contributors start at an identical genesis baseline:

Genesis Score100.00 ptsEqual baseline for all contributors
ScopeContinuousCarries across all authored insights
Evaluation Cost0.00 FPFree voting for all community members

2. Binary Community Evaluation Mechanism

Rather than confusing multi-token staking, liquidity curves, or pay-to-vote mechanics, community members evaluate insights with two intuitive actions:

USEFUL EVALUATION

Awarded when an insight provides rigorous analysis, verifiable mathematical evidence, or high-conviction research. Positively impacts the author's continuous reputation score.

NOT USEFUL EVALUATION

Awarded when an insight is low-effort, unsubstantiated, promotional, or demonstrably misleading. Negatively impacts the author's continuous reputation score.

Anti-Sybil & Self-Voting Protection: Authors are strictly prohibited from evaluating their own insights. Voters may change their evaluation or retract it at any time, which automatically computes the net delta in the author's continuous score.

3. Mathematical Formulation (Bayesian Impact Model)

To prevent single rogue votes from causing wild swings on newly published insights, Finlance applies empirical Bayesian smoothing with an uninformative neutral prior ($m = 10$, prior mean $0.5$):

// 1. Smoothed Usefulness Probability

P = (Useful + 10 × 0.5) / (TotalVotes + 10)

// 2. Evidence Sample Confidence Multiplier

C = TotalVotes / (TotalVotes + 10)

// 3. Cumulative Post Reputation Impact

Impact = 32.0 × (P - 0.5) × C

// 4. Marginal Score Adjustment on Vote Action

Delta = Impact_new - Impact_prev

AuthorReputation_new = max(0, AuthorReputation_prev + Delta)

Key property: A post with 1 Useful vote has a minor positive impact ($+0.29$ pts). As more community members independently verify the post as Useful, confidence $C \to 1.0$, producing an accumulated maximum impact of up to $+16.0$ pts per universally praised post. Conversely, a post with heavy Not Useful votes accumulates up to $-16.0$ pts.

4. Verifiable Reputation Ledger

Every delta ($\Delta$) applied to a contributor's reputation is recorded immutably in the reputationLedger collection. Every record contains the timestamp, previous score, new score, delta, related insight, and event reason. Anyone can audit the author's score history from the genesis 100.00 pts baseline to their current score.

Simulated Knowledge Economy: Community notes and valuations reflect user sentiment, not investment advice. Always verify independently.