01 / THE PRACTICAL QUESTION
A decision, before a model.
An option seller can trade to offset changing exposure, but every rebalance costs money.
Determine whether a learned trading rule improves tail loss under the same costs.
02 / DATA & COMPARISON
The idea in plain language.
A delta hedge offsets an option's local sensitivity to its underlying asset. A neural hedge learns a trading rule from simulated paths and a chosen training objective.
- Delta
- How much an option's price changes for a small move in the underlying asset.
Data. Seeded option price paths; models train separately from evaluation.
Baseline. Black–Scholes delta hedge and unhedged exposure.
03 / THE EXPERIMENT
What the saved experiment shows.
Across all three saved held-out seeds, the neural hedge has higher 95% ES than delta hedging. Both reduce tail loss substantially relative to no hedge. The learned model has not earned its complexity in this benchmark.
Inspect the supporting resultEvidence record: learning-benchmarks.json#projects/08
Explore the related lab04 / RESULTS & LIMITATIONS
Evidence with its boundaries attached.
The related lab is a cross-project demonstration. Read this investigation’s evidence and limits before transferring its conclusions.
Inspect numerical checks and validation records
Loading validation evidence…
Loading across-seed evidence…
Simulated dynamics, discrete trading, and model mismatch limit transfer to live markets.
05 / REPRODUCE
Reproduce and challenge the result.
Code, configuration, and reproduction
The project contains its implementation, configuration, tests, and walkthrough. Download the lab configuration to record the exact parameters used in an interactive run.
Project code and walkthrough