- Home
- Skills
- Trading & Investing
- Honest Policy Simulation
Honest Policy Simulation
Bootstrap Monte Carlo for strategy evaluation — distributions and stated artifacts, not one lucky equity curve.
$24
Works with the AI Agents you already use
Honest Policy Simulation
Example session with this skill installed
Run an honest policy simulation on the provided ETH-USD 5m bars. Calibrate slippage from my journal.csv and test the 'Volatility-Adjusted Exit' policy. Focus on ruin probability.
- Read your context and instructions
- Compiled the honest policy simulation
- Generated the data export
I have completed the bootstrap Monte Carlo simulation for the 'Volatility-Adjusted Exit' policy. The engine performed 1,000 resampled iterations using 20-bar block bootstrapping to preserve local autocorrelation.
Policy Evaluation Report: ETH-USD Volatility-Adjusted Exit
| Metric | Historical Replay | Stochastic (Mean) | 5th Percentile (Left Tail) |
|---|---|---|---|
| Expectancy (R) | 1.42 | 1.18 | -0.45 |
| Max Drawdown | 12.4% | 18.2% | 31.9% |
| Ruin Probability | N/A | 2.4% | 8.7% |
| Win Rate | 58.2% | 54.1% | 46.8% |
| Profit Factor | 1.85 | 1.52 | 0.92 |
Artifact & Regime Analysis
Volatility Clustering: The block bootstrap successfully captured the ETH spike on 2024-05-21; 82% of paths containing this block showed elevated drawdown.
Slippage Calibration: Using journal.csv, a median slippage of 2.1 bps was applied. The policy expectancy dropped by 0.15R compared to zero-slippage assumptions.
Regime Performance: The policy maintains a positive mean in trending regimes but shows a negative expectancy (-0.22R) in mean-reverting ranges.
Reproducibility Gate: PASSED. The deterministic replay matched the stochastic mean on the original path within 0.01%, confirming engine integrity.
Next steps
- Review the Regime Gating logic; the current exit underperforms in sideways markets.
- Tighten the Sizing Policy to account for the 8.7% ruin probability seen in the 5th percentile tail.
- Run a Walk-Forward validation on the 2025-Q1 held-out data to confirm the edge persists.
journal.csv
CSV · data export
Example file from a real run - the skill writes it into your workspace.
Connects securely to your tools. The creator never sees your data.
About this skill
One historical replay is one realization — this skill teaches you to price it as one. Block bootstrap (not GBM) preserves volatility clustering and fat tails; bootstrapping the instrument's own paths preserves nonlinear payoff geometry. The same policy interface runs deterministic replay and stochastic sim — if they disagree on the same bars, the engine is wrong, and that disagreement is a free test. Includes the evaluator output spec, walk-forward tune/validate discipline, artifact modeling for degenerate prints and stale marks, and the reading rules that matter: the left tail decides, not the mean; exits can't create expectancy that entries don't bring; when exit debates hit diminishing returns the next lever is regime gating and sizing; and negative results ("no causal gate in the data") are shippable deliverables when the reproduce gate passes. What's inside: SKILL.md (the method), reference/eval-report-template.md (report structure, reproduce gates, verdict taxonomy), examples/worked-example.md (a filled verdict section comparing two exit rules).
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Trading & Investing.
Requires local historical price data and trade journals. Does not model real-time liquidity or order book depth. Simulation results are distributions, not guarantees of future performance.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 5 days ago