Volume is the one input that tells you how many people cared. Price can drift on almost nothing; a move on heavy volume means participants actually transacted. That makes an unusual volume alert genuinely useful — and also one of the easiest alerts to set up badly, because the obvious way to define "unusual" is the wrong one.
Unusual means unusual for that stock
The mistake is setting a number of shares. Ten million shares is a sleepy session for a mega-cap and an extraordinary event for a mid-cap. A threshold in absolute shares will produce an alert list containing the same handful of enormous tickers every single day.
The fix is to measure each stock against its own recent average — volume today as a percentage of the average of the last 20 bars. Now the comparison is internal to each ticker, a 200% reading means the same thing everywhere, and the rule generalises across a whole index.
Choosing the threshold
| Threshold | Reads as | Across a large-cap index |
|---|---|---|
| 125% | Slightly busier than normal | Very frequent — close to noise |
| 150% | Clearly more active | A sensible default. Still fires often. |
| 200% | Double normal participation | Noticeably rarer, usually something behind it |
| 300%+ | Exceptional | A handful a week. Often news-driven. |
There is no correct threshold, only a correct way to choose one: pick the level that produces a number of alerts you will genuinely read. An alert stream you swipe away is worth exactly nothing, and it is worse than nothing because it trains you to ignore the channel.
Volume is a confirmation, not a signal
This is the part worth being blunt about. A volume spike is directionless. Heavy volume accompanies a stock gapping up on strong results and a stock collapsing on a profit warning. If you alert on volume alone, you are subscribing to both, and you will spend your time opening charts to find out which one you got.
Volume earns its place attached to something directional:
- Breakout + volume — price closes above its 20-day high and volume is 150% of average. The classic pairing, and the one we set up step by step here.
- Support bounce + volume — price at a recent low with participation coming in, rather than drifting down on nobody trading.
- Moving-average cross + volume — a trend change that people actually acted on.
We have not published a backtest of volume on its own, and it would be misleading to imply otherwise — it is a filter rather than a strategy, so there is nothing to enter or exit on. What we have published is what happens when you attach it to a directional trigger and test the result: our breakout test with volume confirmation lost roughly half the account over 81 trades. The confirmation reduced the number of trades; it did not rescue the exits.
Setting the alert up
- Describe it. In the builder, type something like "alert me when an S&P 500 stock breaks its 20-day high on volume 150% above average". You get scanner blocks back that you can read and edit.
- Check the volume block's settings. Confirm the averaging window and the multiple are what you intended. This kind of parameter fails quietly — a wrong window does not error, it just never fires, or fires constantly.
- Pick the timeframe. Daily bars compare today's volume to recent days. Hourly compares this hour to recent hours, which is a different and much twitchier question.
- Backtest it. Run it across the whole index before activating and look at the trade ledger, not just the headline.
- Connect Telegram or email and turn it on.
One thing that will surprise you
Once this is running you will notice that most unusual-volume days are not tradeable. The stock gaps, the volume arrives, and by the time you read the alert the move has already happened. That is not a flaw in the alert — it is what news looks like from the outside.
The setups worth acting on tend to be the ones where volume builds alongside a technical condition you were already waiting for, rather than the ones where volume is the whole story. Which is the same conclusion as above, arrived at the expensive way.
Related reading
- How to get a Telegram alert when a stock breaks out — volume used as the confirmation, end to end.
- How to scan the whole S&P 500 for one setup — running one rule across every ticker rather than a watchlist.
- How to backtest a trading strategy without coding