Probity

Probity is a benchmark-driven forecasting subnet on Bittensor that creates a decentralized competitive marketplace for calibrated probabilistic intelligence.

Problem

Prediction markets are capital-weighted, not skill-weighted.

There is no decentralized system that:

• Identifies consistently well-calibrated forecasters

• Rewards probabilistic accuracy over time

• Produces a programmable probability intelligence layer

Market prices reflect capital allocation — not persistent forecasting skill.

Solution

Miners compete to submit calibrated probability forecasts for objectively verifiable binary events.

Validators deterministically score each forecast using:

• Strictly proper log-loss

• Relative performance vs market baseline

• Rolling skill aggregation

• Exponential weight routing

Emission flows exclusively to persistent outperformance.

The result is a decentralized, skill-weighted probability ensemble (SWPE) that can be exposed as a programmable oracle.

Scoring & Incentive Design

• Skill = Market LogLoss − Miner LogLoss

• RollingSkill smooths short-term variance

• Weight = exp(β × RollingSkill)

• Normalized weights determine TAO emission

Truthful reporting is the dominant strategy.
Mirroring the market yields zero expected skill.
Short-term luck cannot dominate long-term emission.

Key Features

• Permissionless miner participation

• Deterministic validator scoring

• Commit–reveal anti-copy mechanism

• Relative benchmarking vs prediction markets

• Continuous probability aggregation

• Fully auditable emission routing

Target Use Cases

• DeFi derivatives pricing

• Parametric insurance

• DAO governance decision support

• Risk analytics

• Probability API for prediction platforms

Why Bittensor

Probity is inherently suited to Bittensor because its core objective is continuous, benchmark-driven intelligence evaluation.

Bittensor provides:

• Native emission routing to measurable performance

• Decentralized validator consensus

• Permissionless miner competition

• A transparent, stake-weighted weighting mechanism

Probity leverages these primitives to transform calibrated forecasting into a measurable digital commodity.

Unlike centralized reputation systems, Bittensor allows forecasting skill to be economically enforced, continuously evaluated, and permissionlessly monetized.

In Probity, emission becomes the reward function for probabilistic truth.

Tech Stack (Planned)

• Python

• Bittensor SDK

• Market data ingestion layer

• REST Probability API

• EVM-compatible oracle integration

During Round I, we focused on formalizing Probity’s incentive mechanism and validating its economic properties.

Completed work includes:

• Designed full incentive and emission routing logic

• Defined strictly proper log-loss scoring framework

• Implemented relative benchmarking vs market baseline

• Specified rolling skill aggregation and exponential weight routing

• Formalized commit–reveal anti-copy mechanism

• Drafted miner and validator I/O specification

• Published public GitHub repository with scoring prototype

• Prepared validator scoring pseudocode and weight computation logic

• Structured roadmap toward testnet implementation

We also conducted internal edge-case analysis on:

• Market mirroring strategies

• Identity splitting attacks

• Short-term variance dominance

• Collusion scenarios

The current repository contains a functional scoring prototype and architecture specification.
Round II focus will be full testnet deployment with working miner–validator loops.

Probity is currently pre-seed and self-funded.

We have not raised external capital and are focused on mechanism design, subnet implementation, and testnet validation during the hackathon phase.

We are open to strategic ecosystem support and partnerships aligned with long-term subnet sustainability, particularly within the Bittensor ecosystem.