How TradeStupid works

Every stock gets a score built from six parts. It describes how well current conditions line up for buying — it is not a prediction of how much the stock will return, and the parts overlap (several read the same price movements).

The parts add up to 107 possible points and the score is capped at 100, so a stock doesn't need a perfect mark everywhere to reach the top. The score maps to a verdict: 72+ Buy, 55–71 Consider, 38–54 Watch, below 38 Wait.

A good-looking chart can't override real problems. The verdict is held down — and the page says why — when:

• Analysts' average price target is more than 20% below the current price (40%+ drops the rating to Wait).
• The company is burning cash while revenue shrinks.
• The stock is down more than 55% over the past year.
• Net debt is large relative to the company's size and it's burning cash.
• There's no P/E and no analyst coverage, so most of the score would be placeholders.
• The price is more than 18% above its 50-day average — stretched, so the verdict waits for a dip instead of saying Buy (held at Consider).
• The S&P 500 is below its 200-day average — every rating is held at Consider or below.

The ideal buy zone runs from a stock's 50-day average to 8% above it. Suggested stops are based on how much the stock normally moves in a day, widened for longer time horizons and lower risk tolerance so ordinary swings don't shake you out. A wider stop never means more money at risk: the position size shrinks to match, so the estimated loss if the stop is hit stays at the same small share of your portfolio (lower for a lower risk tolerance). A stop is not a guaranteed exit price: once triggered it sells at the next available price, and a stock that gaps down overnight or on earnings can fill well below it. Position sizes use the portfolio size in your profile — until you set one, they're based on an assumed $10,000.

Every morning an automated scan scores a fixed list of 47 large companies. The first day one of them is rated Buy, that's logged as a call (at most once per 20 days per stock). Twenty calendar days later it's graded against the price the scan recorded then, and compared with the S&P 500 (SPY) over the same window.

Limits worth knowing: the sample is small and only covers the time since tracking started; it ignores trading costs and taxes; 20 days is a short window for a strategy meant to be held longer; and calls logged before mid-October 2026 used a looser definition (score alone, ignoring the caps above), so they're reported separately until enough new calls resolve.

Prices, fundamentals, analyst estimates, and news come from Yahoo Finance. "LIVE" means the app re-checks the price every 15 seconds while the market is open — the quote itself can still lag the exchange. Fundamentals and analyst data update less often, and upcoming earnings dates are frequently estimates until the company confirms them. Index bars show the ETF that tracks each index (SPY, QQQ, IWM), so prices are the fund's, not the index level.

The written summaries are generated by open-weight AI models (OpenAI's gpt-oss family, run by Groq) from the same numbers shown on the page. The short "Why / Biggest risk / What would change it" box above each analysis is not AI — it's built directly from the score. It's given the final verdict and told to explain it, not change it, and responses that contradict the verdict are automatically rejected — but it can still get details wrong. The score breakdown is the source of truth.

• Your watchlist, positions, and profile are stored in your browser. They're only kept anywhere else if you sign in.
• If you sign in, that data is synced to a database tied to your email so your devices match. Signing out (and confirming) erases it from that device.
• Sign-in links and the daily briefing are sent through an email provider. AI features send the relevant numbers to an AI provider to write their text.
• No ads, no third-party analytics or trackers, and nothing is sold.

TradeStupid is for information and education only. It isn't a registered investment adviser or broker-dealer, and nothing here is a recommendation to buy or sell any security. Algorithms and AI can be wrong, data can be delayed or incorrect, and investing carries the risk of losing money. Do your own research and consider talking to a licensed professional before making investment decisions.