🎁 Get the FREE AI Skills Starter Guide β€” Subscribe β†’
BytesAgainBytesAgain
πŸ¦€ ClawHub

Kalshi Econ Revision Drift Trader

by @diagnostikon

Trades CPI bin markets on Kalshi accounting for systematic upward revision bias (~0.03 pp) in initial CPI releases. Markets pricing off initial releases unde...

TERMINAL
clawhub install kalshi-econ-revision-drift-trader

πŸ“– About This Skill


name: kalshi-econ-revision-drift-trader description: Trades CPI bin markets on Kalshi accounting for systematic upward revision bias (~0.03 pp) in initial CPI releases. Markets pricing off initial releases underprice higher CPI bins. Requires SIMMER_API_KEY and simmer-sdk. metadata: author: Diagnostikon owner: Diagnostikon version: "1.0.0" displayName: Kalshi Econ Revision Drift Trader difficulty: advanced homepage: "https://simmer.markets/skills" repository: "https://github.com/SpartanLabsXyz/simmer-sdk" requires_env: "SIMMER_API_KEY" requires_pip: "simmer-sdk" default_mode: "paper" live_flag: "--live"

Kalshi Econ Revision Drift Trader

> This is a template. > The default signal uses the well-documented CPI revision bias (+0.03 pp) to shift the fair probability distribution -- remix it with seasonal adjustment patterns, BLS methodology changes, or real-time nowcast data. > The skill handles all the plumbing (market discovery, trade execution, safeguards). Your agent provides the alpha.

Strategy Overview

BLS CPI initial releases are systematically revised upward by approximately 0.03 percentage points. Markets that resolve on the initial release effectively ignore this revision bias, creating a persistent edge. This skill shifts the CPI probability distribution by the revision bias and trades bins where the shift creates meaningful mispricing.

Key advantages:

  • Documented statistical bias -- BLS revision history from 2000-2024 confirms the upward drift
  • Structural edge -- as long as markets price off initial releases, the bias persists
  • Directional prediction -- higher CPI bins are systematically underpriced
  • No timing dependency -- the bias exists regardless of the current CPI level
  • Signal Logic

    Revision Drift Model

    1. Parse CPI bin markets and extract probability-weighted expected CPI mean 2. Compute "initial release" bin probabilities using normal distribution 3. Compute "revision-adjusted" bin probabilities by shifting mean +0.03 pp 4. Edge = revised_prob - market_price for each bin 5. Trade when |edge| >= entry_edge

    Revision Statistics

    | Metric | Value | Source | |--------|-------|--------| | Mean revision | +0.03 pp | BLS CPI revision history 2000-2024 | | Revision std | 0.05 pp | Same dataset | | Direction | Upward | Consistent across decades | | Impact | Higher bins underpriced | Shifts CDF right |

    Example

    If CPI market-implied mean is 2.80%:

    | Bin | Initial Prob | Revised Prob | Revision Edge | |-----|-------------|-------------|---------------| | < 2.0% | 5.5% | 4.9% | -0.6% | | 2.0-2.5% | 22.7% | 21.5% | -1.2% | | 2.5-3.0% | 38.3% | 38.2% | -0.1% | | 3.0-3.5% | 24.6% | 25.8% | +1.2% | | > 3.5% | 8.9% | 9.6% | +0.7% |

    Conviction-Based Sizing

  • conviction = min(|edge| / entry_edge, 2.0) / 2.0
  • size = max($1.00, conviction * MAX_POSITION_USD)
  • Larger edge = larger position, capped at MAX_POSITION_USD
  • Remix Ideas

  • Seasonal adjustment: CPI revisions vary by month (e.g., January effect)
  • BLS methodology tracking: New CPI basket weights change revision patterns
  • Nowcast overlay: Combine revision bias with Cleveland Fed CPI nowcast
  • Multi-release: Track preliminary, final, and revised releases separately
  • Risk Parameters

    | Parameter | Default | Notes | |-----------|---------|-------| | Entry edge | 8% | Min revised-vs-market divergence to trade | | Exit threshold | 45% | Sell when position price reaches this | | Max position size | $5.00 USDC | Per market | | Max trades per run | 3 | Rate limiting | | Max slippage | 15% | Skip if slippage exceeds | | Min liquidity | $0 | Disabled by default |

    Installation & Setup

    clawhub install kalshi-econ-revision-drift-trader
    

    Requires: SIMMER_API_KEY and SOLANA_PRIVATE_KEY environment variables.

    Cron Schedule

    Cron is set to null -- the skill does not run on a schedule until you configure it in the Simmer UI.

    Safety & Execution Mode

    The skill defaults to dry-run mode. Real trades only execute when --live is passed explicitly.

    | Scenario | Mode | Financial risk | |----------|------|----------------| | python trader.py | Dry run | None | | Cron / automaton | Dry run | None | | python trader.py --live | Live (Kalshi via DFlow) | Real USDC |

    The automaton cron is set to null -- it does not run on a schedule until you configure it in the Simmer UI. autostart: false means it won't start automatically on install.

    Required Credentials

    | Variable | Required | Notes | |----------|----------|-------| | SIMMER_API_KEY | Yes | Trading authority. Treat as a high-value credential. | | SOLANA_PRIVATE_KEY | Yes | Base58-encoded Solana private key for live trading. |

    Tunables (Risk Parameters)

    All risk parameters are declared in clawhub.json as tunables and adjustable from the Simmer UI without code changes.

    | Variable | Default | Purpose | |----------|---------|---------| | SIMMER_ECON_REV_ENTRY_EDGE | 0.08 | Min divergence between revised fair and market to trigger trade | | SIMMER_ECON_REV_EXIT_THRESHOLD | 0.45 | Sell position when price reaches this level | | SIMMER_ECON_REV_MAX_POSITION_USD | 5.00 | Max USDC per trade | | SIMMER_ECON_REV_MAX_TRADES_PER_RUN | 3 | Max trades per execution cycle | | SIMMER_ECON_REV_SLIPPAGE_MAX | 0.15 | Max slippage before skipping (0.15 = 15%) | | SIMMER_ECON_REV_MIN_LIQUIDITY | 0 | Min market liquidity USD (0 = disabled) |

    Dependency

    simmer-sdk is published on PyPI by Simmer Markets.

  • PyPI: https://pypi.org/project/simmer-sdk/
  • GitHub: https://github.com/SpartanLabsXyz/simmer-sdk
  • Publisher: hello@simmer.markets
  • Review the source before providing live credentials if you require full auditability.

    πŸ’‘ Examples

    If CPI market-implied mean is 2.80%:

    | Bin | Initial Prob | Revised Prob | Revision Edge | |-----|-------------|-------------|---------------| | < 2.0% | 5.5% | 4.9% | -0.6% | | 2.0-2.5% | 22.7% | 21.5% | -1.2% | | 2.5-3.0% | 38.3% | 38.2% | -0.1% | | 3.0-3.5% | 24.6% | 25.8% | +1.2% | | > 3.5% | 8.9% | 9.6% | +0.7% |

    Conviction-Based Sizing

  • conviction = min(|edge| / entry_edge, 2.0) / 2.0
  • size = max($1.00, conviction * MAX_POSITION_USD)
  • Larger edge = larger position, capped at MAX_POSITION_USD
  • Remix Ideas

  • Seasonal adjustment: CPI revisions vary by month (e.g., January effect)
  • BLS methodology tracking: New CPI basket weights change revision patterns
  • Nowcast overlay: Combine revision bias with Cleveland Fed CPI nowcast
  • Multi-release: Track preliminary, final, and revised releases separately