Most traders end up doing this by hand. You have one setup you trust. You want to know every time it happens, anywhere in the index — not just on the six tickers you happen to have on a watchlist. So you run a screener in the morning, you flick through charts at lunch, and you accept that you are missing most of them.
The reason it is done by hand is that the two obvious tools each solve half the problem. This guide covers what the missing half is, and how to set up the whole thing once so it runs without you.
Why screeners and chart alerts both fall short
| Tool | What it does | What it misses |
|---|---|---|
| Stock screener | Which stocks match right now, when you press the button | Anything that triggered while you were not looking. It is a snapshot, not a watch. |
| Chart alert | Watches continuously — one symbol at a time | The other 499. You would need to create and maintain an alert per ticker. |
| Scanner | One rule set, every ticker, continuously | Nothing structurally — but it has to be set up correctly, which is the rest of this page. |
The distinction is worth being precise about because it is the one people get wrong: a screener is a query, an alert is a subscription. You want a subscription to a query — every stock, all the time, tell me when.
Step 1 — define the rule once, not per ticker
The important shift is that your rule must be written in terms relative to each stock, not in absolute prices. "Above $180" only means something for one ticker. "Above its own 20-day high" means something for all 500, and it recalculates per stock every day.
Rules that generalise across an index look like:
- price closes above its highest high of the last N bars
- RSI(14) below 30 while price sits at a 100-bar low
- volume more than 150% of its own recent average
- the 50-period moving average crossing above the 200-period
In the builder you can type this in plain English and it produces the blocks. Read the blocks it gives you before you go further — the rules are shown precisely so you can confirm it is watching what you meant. A lookback that is off by a factor of five looks exactly like a correct one from the outside, and it will fail silently rather than error.
Step 2 — pick the universe deliberately
You choose one per strategy: the S&P 500, the NASDAQ 100, a broader large-cap US list, or BTC/USD and ETH/USD. Crypto is watched around the clock, because it has no market sessions.
A wider universe is not automatically better. A rule that fires four times a week across the S&P 500 might fire thirty times across a broader list, and thirty alerts a week is a notification you learn to swipe away. Start narrow.
Step 3 — test the rule across the entire index before you switch it on
This is the step that genuinely is hard to do elsewhere, and it is the one worth doing. Rather than testing your idea on one chart and assuming it generalises, you replay it bar by bar against every member of the index and look at the whole trade ledger.
One ticker takes seconds. All 503 takes about a minute. The engine walks historical bars one at a time, opens at the next bar's open and exits at your target or stop — and it is the same evaluation code the live scanner runs, so a rule that tests one way does not quietly behave differently when you turn alerts on.
Every result we have published came from exactly this: the MACD crossover across 503 stocks (1,761 trades), the hammer candle, the support bounce, and a breakout test that lost 50%. Three of those four lost money. That is the normal outcome of testing an idea properly, and it is the reason to test before switching on rather than after.
The caveats apply to your runs as much as ours: index membership is as it stands today, so companies dropped during the period are absent and removals skew toward poor performers. Fills assume no slippage and no commissions. Both flatter the result.
Step 4 — let it run and deliver
Once the rule holds up, activate it and connect Telegram or email. From that point it evaluates every ticker in the universe against your conditions around the clock and messages you when one clears all of them.
The mental adjustment is that most of the work is rejection. On a normal day the scanner checks hundreds of names and tells you about none of them. That silence is the product working, not failing — the single alert is worth reading precisely because of the hundreds it suppressed.
What this is not good for
Being straight about the edges, because the wrong expectation wastes your time:
- Not sub-minute trading. This is built for daily and hourly setups you hold for hours or days. It is not an execution tool and it does not place trades.
- Not small caps or penny stocks. The universes are large-cap US equities.
- Not forex. Not supported.
- Not a source of setups. It automates a rule you already have. It will not tell you which rule is worth running — only the backtest will, and often the answer is no.
Related reading
- How to backtest a trading strategy without coding — what a backtest proves, and the mistakes that fake a good result.
- How to get a Telegram alert when a stock breaks out — a worked example of one rule, end to end.
- What win rate do you actually need? — the number to check your ledger against.
- Best free stock screeners in 2026 — if a snapshot is genuinely all you need.