The anatomy of a multi-condition alert

Most stock alerts have one condition: price crosses a level. The interesting alerts have four or five. Here's how Tickerbot turns a one-line idea into a query that watches the whole market — and what's actually happening under the hood when it fires.

Start with the idea

Imagine a swing trader who's been watching the small-cap technology space. They've noticed that the best entries lately have happened when several things line up at once: a stock breaks out of a consolidation, volume confirms, an analyst upgrade hits within a few days of the breakout, and EPS estimates have been revised up.

That's a real, specific, multi-factor setup. Here's what to do with it.

Step 1: Write it as a sentence

You don't open a query builder. You don't pick a category. You just write what you mean.

Step 2: Tickerbot translates the sentence into a query

Behind the scenes, an AI model parses the sentence into structured logic. The parsing isn't trying to be clever — it's just mapping each phrase to a known data field:

• "small-cap" → market_cap_class = 'small_cap' (precomputed flag)
• "tech stock" → sector = 'Information Technology'
• "breaking the 20-day high" → breakout = true (precomputed flag, edge-triggered)
• "on 2× volume" → relative_volume > 2
• "analyst upgrade in the last 7 days" → recent_upgrade_7d = true
• "EPS estimates revised up in the last 30 days" → eps_revisions_up_30d > 0

The query gets compiled and validated against the schema. If anything is ambiguous, the AI asks a clarifying question. If everything resolves, the alert is saved.

Step 3: The query runs every five minutes

During US market hours (9:30 AM to 4 PM ET, plus pre-market and post-market), an evaluator process runs every five minutes. It executes the saved query against the live market database — which has fresh prices, indicators, ratings, and fundamentals for all 12,000+ tickers.

The query returns a set of tickers that match all conditions. Most ticks return an empty set. Occasionally — when the universe of stocks shifts in a way that satisfies all the conditions — the set has one or more matches.

Step 4: State-change dedup decides whether to fire

The current result set is compared to the previous result set. If a ticker is in both, no notification — you already heard about it. If a ticker is in the new set but not the old set, it's a fresh match, and you get a push notification.

This is the difference between a useful alert system and a noisy one. You get one ping per event, not one ping per evaluation.

Step 5: You get the notification

Push notification on your phone. Includes the matching ticker, the price at match, and the conditions that fired. One tap opens the stock's detail screen in the Tickerbot app, where you can see why it matched and decide what to do.

What you didn't have to do

You didn't write SQL. You didn't pick filter values from dropdowns. You didn't schedule a cron job. You didn't debug a Pine Script. You wrote a sentence and clicked save. That's the entire lifecycle.

The hard part of alerting was never the alerting. It was the translation layer between "what I mean" and "what the database understands."