Astro Trader Insights

Test a historical rule against a benchmark

Compare cycle-based strategies with buy-and-hold using the backtesting tools.

Start researching for free

What you can explore

Inspect assumptions

Review the asset, period, rule and available history behind a result.

Keep simulation in perspective

Backtests depend on data and assumptions. Costs, overfitting and changing regimes can make live results different.

Fix the rule before judging the result

Record the asset, period, entry and exit rules, and any cost assumptions used by the simulation. Compare the result with buy-and-hold over the same interval. Repeatedly changing parameters to improve a past result can fit noise; reserve an independent period for evaluation and keep assumptions visible when discussing performance.

Questions about this tool

Can a profitable backtest predict live returns?
No. Historical prices, execution assumptions, costs and changing market conditions limit what a simulation establishes. A profitable run is a result to investigate, not a guarantee.

How to get started

  1. Sign in with Google to save your web research.
  2. Open the research tool you need and check its inputs, dates and coverage before interpreting the result.
  3. Optionally unlock AI in Settings for €9.90 once and connect your own provider key. Model usage is charged by your provider.

Astro Trader is a research tool, not financial advice. Data may be delayed, incomplete or incorrect. AI explanations require verification. Portfolio trades are simulated.

Methodology, sources and authorship

Astro Trader is developed by Víctor Balcells. The stock model weights valuation (40%), quality and trend (30%), and timing (30%). Risk filters and data coverage accompany the result. Crypto and ETFs have their own models; exploratory cycles do not enter the fundamental stock score.

Market data comes from external providers, including Yahoo Finance; insider evidence links to SEC filings. Consult each report's dates, original sources and coverage: a missing observation is not a confirmed zero. AI text is generated commentary, not a statement written or verified by the author.