Alerts feel delayed¶
When Trade Ideas alerts “feel delayed,” it’s usually not a data lag problem. It’s a definition + workflow problem:
- your lane is too broad, so the alert fires after the move is already obvious
- your setup definition is vague (“movement happened”) instead of specific (“setup is forming”)
- your alert volume is too high, so you only notice the alert after it’s buried
This page is a fast diagnostic flow to get alerts back to reviewable + actionable.
Quick reality check (30 seconds)¶
1) Is the market open / is your universe correct?¶
- If you’re testing during premarket or after-hours, check session settings.
- Make sure you aren’t scanning a dead symbol list.
2) Are you expecting Trade Ideas to be a “signal”?¶
Trade Ideas is best when you treat alerts like review prompts, not perfect entries. Your edge is usually:
lane → rank → alert → review → execute
If you expect “alert = buy now,” most alerts will feel late.
The 80/20 causes (in order)¶
Cause #1 — Your alert describes movement, not a setup¶
If your alert is basically: - “% move is happening” - “new high” - “volume spike”
…it will fire after the thing is already visible.
Fix: Add at least one setup gate so it triggers earlier in the lifecycle: - position in range (near HOD/LOD, near break level) - time-of-day window - relative volume threshold (in-play filter) - liquidity floor (so you’re not reacting to junk)
Cause #2 — Your lane is too broad (noise hides the good alerts)¶
When your lane is broad, you get: - too many alerts - too many “meh” tickers - delayed attention (not delayed data)
Fix: Tighten the lane first: - add a dollar-volume floor - add average volume floor - use relative volume as a gate - constrain price band to what you actually trade
Related: - Too many alerts - Filters that matter
Cause #3 — Time-of-day mismatch¶
Some setups behave in specific windows. If your alert runs all day, you’ll get: - early noise you ignore - late signals you notice
Fix: Add time windows. Related: - Time of day filter
Cause #4 — You’re validating in the wrong window¶
A common pattern is: - alert fires - you look at a chart/layout that isn’t built for fast validation - by the time you decide, it’s “late”
Fix: Your review station should be ready: - Top List for triage - Chart for context - Alert Window for setup history
Related: - Alert Window tips
Practical troubleshooting flow (do this exactly)¶
Step 1 — Reduce to one alert window¶
Disable everything else. You want to debug one thing at a time.
Step 2 — Confirm it fires at all¶
If it doesn’t fire, you filtered to zero or the condition is too rare. Related: - No alerts
Step 3 — Remove one constraint at a time¶
If it fires late: - remove one condition - observe how the timing changes - add it back only if it improves signal quality
Step 4 — Add one “earlier” condition¶
Pick one: - pre-break trigger (e.g., approaching level) - in-play gate (RVOL) - time-of-day
Step 5 — Add one “quality” condition¶
Pick one: - liquidity floor - universe/symbol list quality
Goal: fewer, earlier, higher-quality alerts.
Latency vs slippage vs routing: quick differential diagnosis¶
| Symptom | Likely issue | First fix |
|---|---|---|
| Alerts are timely but fills are worse than expected | Slippage / spread regime | Tighten lane + order controls |
| Alerts seem “late” only in fast names | Setup triggers too downstream | Move trigger earlier in setup lifecycle |
| Inconsistent behavior after reconnects | Routing/API state | Validate connection stability + dedupe rules |
Related: Broker API risk controls
If you still suspect true latency¶
Actual latency is rare, but if you want to rule it out:
- compare the alert timestamp vs the chart timestamp for the same event
- test on a highly liquid name (SPY / AAPL) where prints are constant
- verify your connection stability
If it’s truly a platform latency issue, it will show up even on ultra-liquid names.
Related pages¶
- Too many alerts
- No alerts
- Top List Window (build a daily gap list)
- Alert Window (example workflow)
- Alert Window tips (reduce noise)