Price Alerts That Fire at the Right Time: Not Every Time

Three years ago I had 47 active price alerts on my watchlist and got an average of 14 triggers per day. I silenced the channel in week two. The alerts weren't wrong, AAPL really did touch 200 that day, but they fired on a single 3-second tick that immediately reverted, then fired again on the next re-touch 40 minutes later, then again at 4:02am on a European open ADR cross. "Price touched 200" is not information. What I actually wanted to know was "price has meaningfully committed above 200," which is a completely different question. The fix wasn't to set fewer alerts; it was to set smarter ones. I rebuilt the alert stack using three rules: daily-close crosses instead of intraday ticks, ATR-normalized distance instead of round-number levels, and relative-strength alerts for rotation work. Daily trigger count dropped from 14 to about 2. The remaining 2 were almost always worth looking at. Alert-spam fatigue was a design problem, not a volume problem.

This post is about the Price Alerts card, why three refinements (close-based, ATR-normalized, and relative-strength) separate useful alerts from spam, and how to design alerts that inform rather than interrupt.

TL;DR

• Intraday tick alerts fire on noise that reverts immediately. Close-based alerts fire on commitment.
• ATR-normalized distance filters tape chop. A 2-ATR move is meaningful; a 0.3-ATR move is noise.
• Relative-strength alerts catch rotations early. Individual-price alerts miss the market-leadership signal.
• Always set both sides of a key level. The failed setup is as informative as the successful one.
• Fewer well-designed alerts beat many sloppy ones. Target 2-3 alerts per day max per active watchlist.

Why most price alerts don't help

The generic retail price alert, "tell me when AAPL hits 200", has three fundamental problems:

1. Triggers on ticks that revert. A single 3-second tick through 200 followed by an immediate retrace means nothing, but the alert fires. On a volatile day this can happen 20+ times.
2. No volatility context. A 1% move in a high-vol regime is noise; the same move in a low-vol regime is a breakout. Round-number price levels ignore this.
3. Fires on 24-hour tape prints. ADR prices print through European and Asian sessions; alerts fire at inconvenient times when the US session hasn't even started.

The fix is to replace "price touches level" with three more-selective conditions:

• Close-based: fires only on the daily or hourly close, not intraday ticks
• ATR-normalized: measures move size in ATR units (2+ ATR = meaningful) rather than absolute dollars
• Relative-strength: fires when ticker outperforms or underperforms a benchmark by a threshold over N bars

These three reduce false positives by roughly an order of magnitude. The cost is that you occasionally miss a move that never translates to a close, but most of those were false starts anyway.

What the Price Alerts card shows

The Price Alerts card lets you configure alerts by type and delivery:

• Level-cross: fires when price closes above or below a level (daily or intraday bar close)
• Percentage-move: fires when return over N bars exceeds a threshold
• ATR-normalized move: fires when the move exceeds 2× average true range (filters chop)
• Relative-strength: fires when ticker outperforms or underperforms a benchmark by X% over N days
• Delivery channels: email, browser push, mobile push, webhook (for integration with trading systems)
• Active hours: limit delivery to US market hours to avoid overnight ADR prints
• Rule history: every firing with timestamp, price, and context
• Cooldown: prevents the same alert from re-firing within N hours even if the condition recurs

Three design rules

Prefer close-based to intraday-tick alerts. The intraday-tick alert fires every time the price crosses a threshold, including on immediately-reverting wicks. The close-based alert fires when the price commits, finishing the bar beyond the level. A stock that touches 200 intraday but closes at 198.40 hasn't broken out; the tick alert fires anyway, the close alert correctly stays silent. For swing traders, daily closes are the right granularity. For day traders, 15-minute or hourly closes. Pure tick alerts are useful only for explicit stop-out execution, not for signal generation.

Use ATR-normalized distance for chop filtering. A 1% move means different things in different volatility regimes. In a low-vol name with 0.8% ATR, a 1% daily move is a meaningful breakout. In a high-vol name with 3% ATR, a 1% move is routine noise. Setting alerts in ATR units (2+ ATR = fire) is regime-adaptive: the alert gets more selective when vol expands and more triggerable when vol compresses, which is the right direction of adjustment. Round-number level alerts (200, 50, 100) have the opposite problem, they don't adapt at all.

Use relative-strength alerts for rotation work. Individual-price alerts miss the market-leadership signal entirely. A ticker can hold its absolute price while underperforming the sector badly, or rally 5% while the sector rallies 8%, absolute-price alerts don't distinguish. A relative-strength alert ("fire when AAPL underperforms QQQ by 3% over 10 days") catches rotations as they happen, which is the tradeable signal for long-short or sector-rotation work.

The fourth rule, not a design primitive but a practice: always set both sides of a key level. If you alert on a break above 200, also alert on a failure below 198. The failed breakout ("price touched 200 then closed below 198") is as actionable as the successful one ("price closed above 200 with follow-through"). Setting only the success-side alert means you miss half the signal.

Example: rebuilt alert stack

My rebuilt alert configuration on a watchlist of 15 names:

Total fires per day across 15 names: about 1.6 on average. Before the rebuild: 14. The 1.6 that do fire are almost always worth a second look; the 14 were mostly spam.

What price alerts can miss

• Gap moves. A stock that gaps through a level opens past the trigger; the close-based alert fires at the new close, not at the level.
• Overnight news. Alerts fire on US-session closes; European and Asian session moves on ADRs aren't captured unless specifically enabled.
• Holidays and half-days. Volume and volatility are distorted on holiday sessions; ATR-based alerts may misfire.
• Ex-dividend and split dates. Mechanical price changes can trigger level alerts without any real information; the card flags these dates.
• Very illiquid names. Low-volume names have erratic ATR; alerts based on them may fire on single-print moves.

Common mistakes

• Too many tick-based alerts. Fast path to inbox spam. Convert to close-based.
• Single-side alerts. Missing the failure side means missing half the information.
• No cooldown. Without cooldown, one choppy session can fire the same alert 20 times.
• Alert on every watchlist name. Scope alerts to positions you'd actually act on; alerts on "watching only" names become noise.
• Ignoring the rule history. Review which alerts fired and whether the firing was useful; dead-wood alerts should be deleted, not left to spam.

Where it fits

Pair with Trade Alerts for entry/exit signal alerts (not just price), Trade Signals for the underlying technical triggers that generate the signal, and Watchlist to scope alerts to names you actually track. For portfolio-level alerts (drawdown, correlation breakdown), see AI Insights.

FAQ

How fast is delivery?

Email and push within 30 seconds of close. Webhook within 1-2 seconds.

Can I alert on options prices?

Yes, options-specific alerts support strike-level price, IV, and delta triggers. See the Trade Alerts card for those.

What's a reasonable alert count?

Target 2-3 per day across your full watchlist. More than 5/day is usually a design problem.

Does it support alert suppression during earnings?

Yes, the calendar-aware toggle silences price alerts in the 24 hours around earnings prints to avoid noise.

Can webhooks integrate with my broker?

Yes, if the broker supports webhook-driven order entry. Most do via third-party middleware.

Related reading

• Dashboard Recent Alerts
• How to Set Trade Alerts
• Dashboard Watchlist Overview

Open the Price Alerts card → /app/alerts