How We Test AI Stock Tools

1. Real accounts, dated prices

We sign up for every tool ourselves. Prices are copied from the vendor's official pricing page and stamped with the date we checked them. When a price changes, we update it and keep the new date.

2. Weekly pick logging

Each week we record every top pick a tool publishes, with the closing price on the day of the pick. Nothing is removed afterwards, including losing picks, so the log is free of survivorship bias. Picks are logged as they are published rather than reconstructed from history, so a backtest that may suffer from overfitting never counts as a track record.

3. 12-week comparison with the S&P 500

After 12 weeks we compare an equal-weighted basket of the picks with SPY, which tracks the S&P 500 benchmark, over the same dates. The return above or below SPY is a simple measure of the tool's alpha. A tool gets a score only after this window. Progress is public on the pick tracker.

4. What we will not do

• Invent reviews, testimonials or reviewer personas.
• Let affiliate commissions decide a verdict or a ranking.
• Claim or imply guaranteed returns.

5. AI assistance

We use AI to help research and draft. A human editor checks every figure against its primary source before publishing.

Data downloads

Every price, plan and check date on this site is published as open data: tools.json and tools.csv. Free to reuse with attribution (CC BY 4.0): bestaistocktools.com.