Validator: Miner Evaluation and Selection

Overview

In the SecureTrade Bittensor Subnet, the validator's most critical function is to act as a sophisticated auctioneer and risk manager. After receiving a Credit Line Issuance Request (CLR) from the SecureTrade Trader application, the validator initiates a competitive process among miners (Underwriters) to find the most optimal Credit Line Issuance Offer (CLO).

This document details the step-by-step process a validator follows to query miners, evaluate their offers, select a winner, and ultimately score their performance for the distribution of TAO rewards. This mechanism is the engine that drives competition, ensures transaction security, and maintains the health of the subnet.

The Miner Evaluation

The entire process, from receiving a request to scoring miners, can be visualized as a lifecycle. This cycle is triggered each time a validator receives a new CLR.

1: The Query - Distributing the Credit Line Issuance Request (CLR)

When a validator receives a CLR from the SecureTrade Trader application (when credit limit request is sent to marketplace), it forwards the request to all active miners registered on the subnet that are connected to its Axon.

The query contains all the information a miner needs to make a decision:

• Borrower Data: Anonymized but comprehensive data for risk assessment, such as transaction history, industry context, and other financial indicators.
• Lock-in Period: (in days) represents an estimate time for which miner's collateral is gonna be locked and can be used for risk assessment by the miner.
• Credit Amount: Can also be included in CLR if specified by trader while requesting so miner decides to offer for demanded credit line.

2: The Response - Receiving Credit Line Issuance Offers (CLOs)

The validator sets a timeout period to wait for responses. Miners analyze the CLR and respond with their Credit Line Issuance Offer (CLO), which includes:

• Quoted Premium: The price the miner is willing to accept to underwrite the risk.
• Credit Amount: Collateral for the transaction amount.
• Transaction Reference (tx_ref): A unique identifier for their offer, which is also registered on-chain.

3: The Evaluation and Selection Process

Once the timeout is reached, the validator evaluates all received CLOs through a multi-stage filtering and selection process. This logic is typically implemented within the validator's forward() method.

A. Initial Filtering & Sanity Checks Before comparing offers, the validator disqualifies any that do not meet the basic requirements:

• Full Collateral Coverage: The CLO must offer 100% collateral coverage. Any offer proposing less is immediately discarded as it fails to meet the subnet's core security principle.
• Rationality Check: The validator filters out irrational or outlier offers. This includes:
  • Offers with a premium higher than the underwriting budget.
  • Offers with a negative or zero premium.
  • Malformed or incomplete responses.

B. Selecting the Optimal Offer From the pool of valid CLOs, the validator selects the single best offer based on a clear hierarchy of criteria:

• Premium Competitiveness (Primary Factor): The most crucial factor is the quoted premium. The validator will always select the valid offer with the lowest premium. This creates a hyper-competitive environment that directly benefits the platform and its users by driving down the cost of credit line issuance.
• Historical Performance (Tie-Breaker): In the rare event that two or more miners submit identical, winning premiums, the validator may use historical performance data as a tie-breaker. This data can include:
  • Reliability: The miner's historical success rate in completing transactions.
  • Activity: The frequency and volume of valid offers submitted over time. A miner with a stronger track record may be favored to ensure maximum reliability.

4: On-Chain Execution

After selecting a winner, the validator sends a p2p message through dendrite to the miner for on chain tx to deposit TRADE-S to the borrower's public safe

The Scoring & Weight-Setting Cycle

The tempo (configured as 24 hours) is the heartbeat of the Bittensor incentive mechanism. At the end of each tempo, the validator aggregates all the performance data from the transactional cycles and translates it into scores that determine TAO rewards.

Throughout the 24-Hour Tempo:

• The validator is continuously running the Transactional Validation Cycle described above. It may issue dozens or hundreds of CLRs, evaluate thousands of CLOs, and execute multiple deals on-chain.

At the End of the Tempo (The "Bookkeeping" Phase):

• Aggregate Performance Data: The validator reviews all the activity from the past 24 hours: which miners responded, their offered premiums, who won, and who were uncompetitive.
• Calculate Miner Scores: Based on this data, the validator calculates a performance score for every miner it observed.
  • Miners who won bids receive the highest scores.
  • Miners who participated with competitive (but not winning) offers receive partial scores.
  • Miners who submitted irrational offers or failed to respond receive a score of zero.
• Normalize Scores into Weights: The scores of all miners are normalized to a sum of 1. A miner's final "weight" is their proportional share of the total score pool.
  • Example: If Miner A has a score of 50, Miner B has 30, and Miner C has 20, their total score is 100. Their weights would be A=0.5, B=0.3, and C=0.2.
• Set Weights on the Subnet: This is the most critical step for the incentive mechanism. The validator calls the set_weights function on the Bittensor blockchain, broadcasting its calculated weights for all miners. This public declaration is its "vote" on the quality and performance of each miner.
• Handle Low-Activity Periods: If, for any reason, the validator did not process any new CLRs during the tempo, it re-applies the weights from the previous tempo. This ensures network stability and prevents miners from being unfairly penalized (losing their rewards) due to validator downtime or a lack of simulated "deal flow."

These weights, submitted by all validators on the subnet, are aggregated by the Bittensor blockchain. The final, aggregated weights determine how the subnet's TAO emissions (in this case, 41% of the total) are distributed among the Underwriter miners. This creates a powerful feedback loop: high-quality performance leads to higher scores, which leads to higher weights, which results in greater TAO rewards, incentivizing every miner to be as competitive and reliable as possible.