Hold CEX model positions to their declared prediction horizon - #1183
Conversation
|
Important
This repository does not receive automatic reviews because it has fewer than 10 stars. ⚙️ Run configurationConfiguration used: defaults Review profile: CHILL Plan: Advanced Run ID: Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
Codex Review SummaryThis comment shows the latest Codex review activity on this pull request.
ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings. |
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 1cdf5afa99
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
1cdf5af to
5cf917d
Compare
Change
A model predicting an h-second return could previously close or resize its position every second. Bind a cost-aware holding policy to the prediction horizon: one episode per instrument, fixed entry quantity, ignore new signals while holding, and close before another entry. The signal clock determines t+h; both IOC legs retain the frozen arrival-latency rule. Actual fill-to-fill duration and incomplete exits are derived from the accounting trace and independently checked on readback.
The same policy now governs supervised validation, frozen-model evaluation, portable FormulaStrategy behavior and native event replay. The quantity ledger charges executable entry/exit quotes and fill-notional costs without double-counting spread. Incomplete tails and gaps fail explicitly. Entry thresholds and costs remain unchanged; this does not claim holding caused historical zero trades.
Preparation reuse #1180 is merged. Ridge-only three-arm execution and the full workflow integration follow under the active research task; no real H1 run has started.
Issue relationship
Refs #1182
Validation
Runtime impact and rollback
Merge may publish research and strategy artifacts through existing automation; this change does not activate live trading or bypass runtime gates. Roll back the source/image identity and retain existing immutable research evidence. New H1 experiments require merged prerequisites and genuine signed budget authority; the prior six-group budget remains exhausted.