Complete your first Strategy run

What this guide is for

This is the one end-to-end workflow in Getting Started. It takes you from one Dataset and one existing Strategy to a completed fixed Study and a retained Run Output. The purpose is to learn how Arizmic hands exact objects between pages, not to find an optimized or deployment-ready Strategy on the first attempt.

Detailed import fields belong in Data Manager, Strategy authoring belongs in Strategies, Study methods belong in Studies, and interpretation belongs in Results. This guide links those pages instead of duplicating their complete instructions.

Before you begin

Confirm that:

• the launcher and workstation are ready;
• System & Jobs -> Jobs and Activity Log are available;
• you have a permitted market-data source or an existing Dataset;
• the Catalog shows the Dataset's Instrument, coverage, schema, and state; and
• the Strategies library contains a saved Strategy compatible with that data.

Start with a small, representative date range. A bounded first run is easier to diagnose and confirms the object handoffs before you spend time or provider charges on larger work.

1. Select or import a Dataset

Open Data Manager -> Catalog. If a suitable Dataset already exists, select it and verify:

• provider or local source;
• Instrument and venue meaning;
• timezone and session handling;
• data level and schema;
• first and last timestamp;
• quality and coverage state; and
• exact Dataset identity.

If you need to import data, use Local Import or Import API and follow the corresponding Data Manager guide. Review mappings and the candidate before activation. Do not continue merely because an import Job completed; the Catalog must show an available Dataset with the intended meaning.

2. Choose one Strategy version

Open Strategies -> Library and choose an existing Strategy that supports your Dataset and intended execution level. For the first run, prefer a shipped example with a small parameter set and clearly stated inputs.

Confirm:

• Strategy name and version;
• required Signals and upstream dependencies;
• required data level and preparation;
• parameter defaults and allowed ranges;
• supported execution fidelity; and
• any Instrument, session, or warm-up requirements.

If readiness reports a missing dependency, resolve it in Signals, Strategies, or Data Manager before opening Studies. Do not select a different data level solely to silence a compatibility warning.

3. Open a fixed Strategy Study

Use the Strategy's Open in Studies action, or open Studies, choose the Strategy subject, and bind the exact Strategy and Dataset versions.

Choose the fixed or Solo Study path for this first run. It asks one concrete question: what does this exact Strategy configuration produce on this exact Dataset under these execution assumptions?

Set:

• one exact parameter value for every required Strategy parameter;
• a bounded date or data window inside Dataset coverage;
• a supported execution fidelity;
• commission, fee, slippage, and other available cost assumptions;
• initial capital and supported sizing inputs; and
• retention settings needed for the Result views you plan to inspect.

Avoid a parameter search, walk-forward design, Portfolio subject, or Live deployment for the first run. Those are separate workflows with additional decisions.

4. Review resolved preparation

The Study validator determines which Prepared Data and Signal outputs the request needs. It may reuse a compatible retained build or propose a required preparation Job.

Review the resolved chain:

• selected Dataset;
• required Prepared Data kind and settings;
• Signal versions and dependencies;
• Strategy version and parameters; and
• requested execution fidelity.

A compatible build is not simply one with a similar label. Its Dataset, settings, data level, and declared requirements must cover the request.

5. Validate and estimate

Run the Study validation before queueing. Clear blockers at the owning stage:

• Dataset or coverage problems in Data Manager;
• missing or incompatible Signals in Signals;
• Strategy definition or parameter problems in Strategies; and
• Study window, fidelity, resource, or assumption problems in Studies.

Review warnings rather than automatically accepting them. A warning can permit execution while still materially limiting how you should interpret the Result.

Where available, review the resource estimate. Reduce the date range, retention, or workload if it exceeds the practical limits of the current computer.

6. Queue the Study

Submit the validated Study once. Record or follow the created Job in System & Jobs -> Jobs.

The Job can pass through preparation and execution phases. Use the active-job status in the application header for a quick view and the Jobs page for full state, progress, cancellation, and recovery controls.

You can navigate elsewhere while it runs. Do not submit a duplicate simply because the originating Studies page is no longer open.

7. Handle failure at the owning stage

Fix the first meaningful blocker. Repeatedly changing several inputs at once makes it difficult to know which correction mattered.

8. Open the retained Result

When the Job completes, use its result link or open Studies -> Results. First confirm that the Study identity, subject, selected versions, method, and status match the request you submitted.

Then open the produced Run Output in Results. Review in a stable order:

1. Overview for identity, completion, and headline values;
2. Chart for the performance and drawdown path;
3. Performance for available returns, risk, and trade summaries;
4. Parameters for the exact Strategy and execution inputs;
5. Positions & Trades for individual behavior;
6. Diagnostics for warnings and runtime findings; and
7. Outputs for retained artifacts and completeness.

Other tabs appear when the Study, subject, fidelity, and retention settings support them. An unavailable tab or field is not a zero value.

9. State what the run established

One fixed run tells you how one exact configuration behaved on one exact input under one set of modeled conditions. It does not establish robustness, out-of-sample performance, capacity, live fill quality, or future returns.

Record:

• Dataset and Prepared Data identities;
• Signal and Strategy versions;
• parameter values;
• Study method and window;
• execution and cost assumptions;
• Result and Run Output identities;
• warnings or missing outputs; and
• the next question you intend to test.

If you proceed, choose the next owning guide instead of repeating this entire workflow:

• Strategy Study Design for searches and validation designs;
• Portfolio Study Design for shared Portfolio evaluation;
• Python Notebooks for a custom bounded investigation;
• AI Companion for brainstorming or reviewed proposals; and
• Live for the separately gated supervised deployment workflow.

First-run checklist

The workflow is complete when:

• one exact Dataset and Strategy version were selected;
• Study validation resolved compatible inputs;
• one bounded fixed Study completed through a durable Job;
• the Study Result and Run Output are retained and open correctly;
• warnings and missing outputs were recorded; and
• the conclusion is limited to what this exact run can show.