Strategy Library

Use Strategy Library

Strategy Library is where you find, inspect, organize, and maintain the versioned trading rules available in Arizmic. A Strategy combines declared inputs, parameters, entry and exit rules, position handling, risk rules, and execution requirements. It does not choose a Dataset or run research by itself.

Open Strategies -> Strategy Library before using a Strategy in a Study, after syncing a local Python plugin, or when you need to inspect the exact rules behind a saved version.

Find a Strategy

Use Search, Group, and pagination to find a catalog entry, then select a row to open its Inspector.

Search and groups help you locate a Strategy; they do not prove that its requirements are met for the Dataset or time window you intend to use. That is resolved during later validation.

Read the Inspector

The Inspector is the quickest way to understand the selected Strategy.

• Metadata shows source, minimum data context, minimum execution, and the last update.
• Compile Check reports whether the available compiled forms passed their checks. It is not a performance result or a release certification.
• Parameters lists defaults and bounds. Adjust them deliberately; a change affects future use, not earlier results.
• Inputs and Signal Inputs show the values the Strategy reads. Inspect an exact Signal version before assuming a similarly named output is the same input.
• Entry rules, Exit rules, and Risk show the actual decision and position-handling logic.
• Advanced contains exact IDs, versions, plugin details, and compile hashes for support or reproducible work.

Read the requirements and rules together. A Strategy may be structurally valid while still needing a different Dataset, Signal output, or prepared context.

Upstream and downstream dependencies

A Strategy's upstream dependencies are its declared inputs and everything needed to produce them. Those inputs can be raw market fields or exact Signal outputs. A Study follows that chain during validation and preparation; the Strategy does not need to repeat every intermediate Signal as a separate input.

For example, Session VWAP True Z-Score Mean Reversion reads two outputs from one direct Signal: session_vwap_true_zscore:seed1:zscore and session_vwap_true_zscore:seed1:weighted_sd. That Signal calculates both from session_id, close, and volume, so its chain ends at the Dataset. By contrast, a Strategy using Session VWAP Distance has a transitive chain: Session VWAP → Session VWAP Distance → Strategy.

A Strategy's downstream dependencies are the Studies, results, and any later deployment configuration that use its exact saved version. Changing a Strategy's rules, inputs, or requirements creates a new version; it does not rewrite earlier work. Before sharing a plugin, make sure the recipient has each required Signal version or an approved package that supplies it.

Organize and sync

Use Manage Strategy Groups and Labels to add local groups and labels. They make the Library easier to browse but do not change the Strategy's rules, version, package, or behavior on another workstation.

Use Sync Plugins to rescan the configured local Strategy Library. Use Sync in the Inspector when you changed one known Python plugin and only want to import that Strategy. Sync updates the catalog; it does not rerun Studies or silently replace versions used by earlier results.

Only sync local code you wrote or reviewed. A Python Strategy plugin is trusted local code during discovery.

Choose an authoring path

All three paths need the same final steps: inspect the saved Strategy, validate it against the intended data, and create a Study when the requirements are clear.

Versions and deletion

Changing a Strategy's rules, inputs, parameter meaning, risk behavior, or execution requirements creates a new version. Keep the old version when prior research or comparisons need to remain understandable.

Delete removes the active catalog entry. It does not rewrite old results, and it does not remove a plugin's source files. Review downstream use before deleting a Strategy; an installed plugin can be discovered again by a later sync.

Before you create a Study

1. confirm the exact Strategy version and its inputs;
2. review parameters, entry/exit rules, position policy, and risk handling;
3. check the required data context and execution floor;
4. resolve any missing Signal inputs or dependencies; and
5. validate the Strategy with the intended Dataset and time window.