Atlas ai backtesting agent

Saved Strategies

For the complete documentation index, see llms.txt. Markdown versions of documentation pages are available by appending .md to page URLs; this page is available as Markdown.

Atlas lets you save any backtested strategy returned in chat to a dedicated dashboard, then re-checks each saved strategy weekly against the latest strategy database. Strategies that still pass the quality filters are tagged Current; those that drop out are tagged Deprecated.

How to save a strategy

  1. In an Atlas chat, ask for a strategy (see Your First Strategy Search and Prompting Tips).
  2. On any returned strategy card, click the star icon in the top corner.
  3. The strategy is added to your Saved Strategies dashboard with a snapshot of its metrics at save time.

A strategy is identified by its asset, timeframe, direction, and entry/exit conditions — not by how the prompt was worded. Re-saving the same setup (even if you asked for it differently) updates the existing entry instead of creating a duplicate.

The Saved Strategies dashboard

Open the dashboard from your account menu to see every strategy you've starred, most recently updated first. Each row shows:

  • Asset, timeframe, and direction
  • The entry/exit conditions
  • Snapshotted metrics: Net Profit, Win Rate, Sharpe, Profit Factor, Max Drawdown, Total Trades, MCPT p-value, and Quality Tier
  • The equity curve from the saved backtest
  • A status badge: Current or Deprecated

Current vs Deprecated

StatusWhat it meansMetrics behavior
CurrentThe exact set of trade conditions still appears in the latest weekly strategy database.Metrics refresh to the latest values on every dashboard load.
DeprecatedThe conditions are no longer present in the refreshed database.Last-known snapshot preserved; metrics frozen at the point the strategy dropped out.

A Deprecated tag means one of three things happened in the most recent weekly rebuild:

  • The strategy's out-of-sample performance dropped below the quality thresholds.
  • Its Monte Carlo permutation p-value no longer cleared the significance gate.
  • The instrument/timeframe set was rebuilt and this specific condition combination didn't re-emerge in the top-ranked set.

A Deprecated flag is a signal to re-investigate, not a verdict that the idea is dead. The frozen snapshot is preserved as an audit log of why the strategy was once worth trading.

How re-validation works

The Atlas strategy database is rebuilt weekly (see Backtesting Configurations for the full pipeline). When you open the Saved Strategies dashboard, every saved strategy is re-checked against that latest build — no manual refresh required.

Re-validation is exact, not fuzzy: the same asset, timeframe, direction, and condition set must survive the same quality gates applied to every new candidate. Those gates include:

  • The standard quality-score thresholds used for all strategies in the database.
  • A Monte Carlo permutation test on log returns — if shuffling the return sequence destroys the edge, the strategy fails the gate.

If those gates fail in a given week, the strategy drops out of the rebuilt database and your saved entry flips to Deprecated.

What to do with a Deprecated strategy

  • Ask Atlas for a variant. Loosening or tightening one condition often reproduces the edge.
  • Check related instruments. Sometimes a single instrument dropped while the same condition set still works on related markets.
  • Compare snapshot vs. recent performance. If Sharpe collapsed but drawdown didn't, regime shift is more likely than overfitting.
  • Don't delete the entry. The frozen snapshot is its own audit log of why you stopped trading the setup.

What gets saved

FieldSaved at first saveRefreshed when CurrentFrozen when Deprecated
Asset, timeframe, directionYes
Entry/exit conditionsYes
Net ProfitYesYesYes (snapshot)
Win RateYesYesYes (snapshot)
SharpeYesYesYes (snapshot)
Profit FactorYesYesYes (snapshot)
Max DrawdownYesYesYes (snapshot)
Total TradesYesYesYes (snapshot)
MCPT p-valueYesYesYes (snapshot)
Quality TierYesYesYes (snapshot)
Equity CurveYesYesYes (snapshot)

Agent Instructions

This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

Querying This Documentation

If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the ask query parameter:

GET https://algoalpha.gitbook.io/algoalpha-user-manual/atlas-ai-backtesting-agent/saved-strategies.md?ask=<question>

The question should be specific, self-contained, and written in natural language. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.

⌘K