Argus Edge — Prediction Market Betting Engine

Argus-style prediction market edge detection and betting strategy. Computes expected value from TA-implied probability vs market odds, sizes bets via Kelly c...

Security Scan

This is an instruction-only betting strategy document that matches its description, requires no credentials or installs, and is internally coherent — but its performance claims are unverified and it leaves data‑gathering/implementation details to the agent.

Purpose & Capability

The name/description (Argus-style prediction-market betting strategy) matches the SKILL.md content (edge calculation, Kelly sizing, freshness/consensus rules). No unrelated credentials, binaries, or installs are requested.

Instruction Scope

The SKILL.md contains formulas, rules, and usage examples but no concrete implementation steps for obtaining market data or placing bets (no API endpoints, no commands). That makes the instructions intentionally high-level: coherent for a strategy doc, but it grants broad discretion to the agent to fetch market data or interact with exchanges if the agent is given such capabilities elsewhere.

Install Mechanism

Instruction-only skill with no install spec and no code files; nothing will be written to disk or installed by this skill itself.

Credentials

No environment variables, credentials, or config paths are requested. The skill does not ask for unrelated secrets or system access.

Persistence & Privilege

always:false (default) and normal autonomous invocation permitted. The skill does not request elevated persistence or modify other skills or agent-wide settings.

Assessment

This skill is essentially a strategy document — it doesn't install code or request credentials, so it's coherent with its stated purpose. Before using it: (1) recognize the win-rate/validation claims are unverifiable from the SKILL.md alone; treat them skeptically. (2) The skill does not provide market-data sources or execution instructions, so an agent using it will need to fetch market quotes and/or place bets — ensure the agent does not have API keys, exchange credentials, or automated trading permissions you don't intend it to use. (3) If you plan to implement these rules programmatically, validate with small, non-production stakes first and add safeguards (rate limits, bankroll caps, explicit confirmation before placing real bets). (4) If you want a fully-automated tool, insist on concrete data-source and execution steps (APIs, authentication model) before granting any credentials.

The complete Argus strategy for finding and exploiting edge in crypto prediction markets.

Core Formula

Edge = our_P(win) - market_implied_P(win)
Bet when: edge ≥ 10% AND fresh market (<30 min old) AND TA score ≥ ±2
Kelly stake = (edge × bankroll) / odds

Strategy Rules

Freshness Guard

• Skip markets >92% consensus (dead signal, L020)
• Prioritize markets <30 min old (primary window)
• Primary bets WR: 77.8% vs overall 56.6%

Counter-Consensus Rule (L023)

TA score ≥+1 + market DOWN >80% + ≥20 min remaining → bet UP (counter-consensus has positive EV, validated at 5x+ payout)

Asset Calibration

Asset	TA Reliability	Bias	Min Score
BTC	0.75	Neutral	±3
ETH	0.80	UP+0.05	±2
SOL	0.90	UP+0.05	±1
XRP	0.70	UP+0.08	±2

Usage

Use argus-edge to find the best Polymarket bet right now

Use argus-edge to calculate edge on BTC 11am ET market

Use argus-edge for full market scan with Kelly sizing

Battle-Tested

Derived from 100+ live Polymarket bets. 25 documented lessons (L001–L025).