Real-time alert reliability is how often a signal reaches you or your bot, on time, once, and matching what the closed candle showed. For scalpers on 1m to 15m charts it breaks into four checks: delivery speed, delivery certainty, duplicates, and signal honesty (no repainting). A measured study of TradingView alerts found a median delay of about 4 seconds on 1-minute charts, with a long tail well beyond that. Plan around the tail, not the median.
TL;DR
- Alert delivery and order execution are two separate steps. A fast alert can still meet a slow fill.
- A 2026 study of 34,174 TradingView alerts measured a 4.03 second median delay and a 16.1 second 99th percentile on 1-minute candles.
- Webhooks are fire-and-forget: no retries, no acknowledgment. Add your own monitoring and reconciliation.
- Use bar-close alerts and non-repainting logic, then test both with timestamps instead of trusting the chart.
- No alert system predicts markets. U.S. regulators warn against any promise of guaranteed returns from signals or bots.
Table of Contents
- What real-time alert reliability means for a trader
- What the data says about TradingView alert delay
- Why alerts arrive late, fail or repeat
- How to check whether an indicator repaints
- How to measure your own alert latency
- Safeguards that reduce missed and false alerts
- Our perspective
- Build a reliable setup with Scalping-Algo
- What regulators say about limits
- FAQ
- Sources
What real-time alert reliability means for a trader
Most people say "my alert was late" and mean one of several different failures. Separate them, because each has a different fix.
| Failure | What you see | Usual cause |
|---|---|---|
| Late | Signal arrives seconds after the candle closed | Delay in the charting platform's dispatch, a slow network, or a slow receiver |
| Missing | Chart shows a signal, nothing arrives | Webhook never delivered, alert expired, endpoint down |
| Duplicate | Same signal twice | Wrong alert frequency, retried or double-configured alerts |
| Misleading | Signal vanishes or moves after you acted | Repainting on the unfinished bar |
The second distinction matters more: alert delivery is not order execution. The alert is a message that a condition was met. Execution is your broker or exchange accepting, routing and filling an order. Delivery latency is measured from candle close to message receipt. Execution latency starts when your order leaves your system. On a 1-minute scalp targeting a few ticks, you pay both, plus slippage.
For context on how fast the execution side can be in regulated equity markets, FINRA reported that 99.55% of held customer market orders in NMS stocks up to 20,000 shares were executed in under 500 milliseconds in September 2021. That says nothing about your alert pipeline. It shows that once an order reaches a venue, the biggest delays are often upstream of it.
What the data says about TradingView alert delay
The best public measurement we found is a 2026 analysis from the TV-Hub team. They logged bar-close time, the platform's alert-generation timestamp and the time their server received the webhook, using a server log and the {{timenow}} placeholder. Across 34,174 alerts, the median total latency from 1-minute candle close to server receipt was 4.03 seconds, the 90th percentile 7.62 seconds and the 99th percentile 16.1 seconds.
Three takeaways for scalpers:
- The median hides the risk. If your edge is gone after 3 seconds, a median of 4 seconds already breaks it. A 16-second tail breaks even slower systems.
- The delay sits mostly upstream of your server. The same study attributes it largely to the charting platform's dispatch pipeline rather than the receiving server. Buying a faster receiver helps less than you expect.
- Timeframe decides what is tolerable. The study reported that lower timeframes are more affected by drift, with 1-minute marginal and 15-second charts unusable.
This is one company's dataset, so treat the figures as a benchmark to compare against, not a promise. Your own numbers will differ by plan, region, chart load and time of day. That is why testing matters (see below).
Busy sessions run slower
The same analysis found that U.S. market sessions (13:00 to 19:59 UTC) ran 27% slower at the median than off-peak hours. Volatility spikes and session opens are exactly when scalpers want speed, and exactly when the pipeline is busiest. Size positions for the slow case, not the quiet-hour case.
Why alerts arrive late, fail or repeat
Platform dispatch
The alert has to be generated at bar close, queued, and sent. You do not control that stage. What you control is how much work you add to it: fewer alerts per chart, simpler conditions and a lean message payload.
Webhooks have no delivery guarantee
Webhooks from TradingView are described as fire-and-forget requests with no delivery guarantees, retries or acknowledgments. If your endpoint is down for 20 seconds, that signal may be gone. Reliable automation therefore needs a middle layer that logs every inbound message, confirms what was acted on, and reconciles against your broker's open positions.
Your own network and hardware
The SEC has long warned that slow internet service providers, inadequate hardware or heavy network traffic can cause significant delays or failures when accessing automated trading systems. If you run a bot or terminal at home on Wi-Fi, this is your weakest link. A dedicated always-on machine helps, and a hosted server is a common answer. If you go that route, check the trading VPS requirements for CPU, memory, storage and location before you commit, since an undersized server adds its own delays.
Duplicates from alert settings
The alert frequency option decides how many times one condition can fire. "Once per bar" and "once per bar close" behave differently on an unfinished candle. Combined with a signal that flips intrabar, you can get repeats or ghosts. For scalping, set bar-close frequency unless you have a tested reason not to. Our guide to TradingView alert message syntax covers placeholders and formatting so each message carries a unique, parseable identifier you can de-duplicate on.
Repainting
A repainting indicator changes past or current-bar signals after the fact. The alert may fire on a state the closed candle never confirms. This is the most damaging failure because the chart looks perfect in hindsight while live results disappoint.
How to check whether an indicator repaints
You can test this in under an hour with no special tools.
- Watch it live. Put the indicator on a 1-minute chart and watch a signal form. If an arrow appears mid-candle and disappears before the close, it repaints on the open bar.
- Reload and compare. Screenshot the last 20 signals, refresh the chart, change timeframe and return. Any signal that moved, vanished or appeared is suspect.
- Use Bar Replay. Replay the same stretch forward bar by bar and compare to the static chart. Differences point to future-leaning logic.
- Read the code. Open-source Pine Script lets you check whether signals are gated by bar confirmation. Closed-source scripts leave you guessing.
- Check the alert, not the chart. Log the alert's timestamp against the candle close. A signal confirmed on close should arrive after the close, never before it.
A non-repainting signal is not a profitable signal. It only means the alert tells the truth about the past bar. That is the honest minimum.
How to measure your own alert latency
Copy the method from the study, scaled down.
- Create a test alert on the chart and timeframe you trade, with bar-close frequency.
- Put
{{timenow}}in the message so the platform's own timestamp travels with the alert. - Point the webhook at a simple endpoint that writes the receipt time to a log.
- Subtract the candle's close time from the receipt time for each alert.
- Collect at least a few hundred samples across quiet hours and the U.S. session.
- Report the median, 90th and 99th percentiles. Design for the 99th.
Step-by-step versions with working code live in our posts on cutting TradingView alert latency and measuring with {{timenow}} and copy-paste TradingView webhook examples. For the bigger picture of how alerts feed a bot, see webhook trading alerts explained.
Test the mobile and Discord path separately
Push notifications and chat messages travel on different routes from webhooks. Time them the same way: note the candle close, note when the message lands on your phone. If you act manually, this is the number that matters, not the server log.
Safeguards that reduce missed and false alerts
- Bar-close alerts only for entries on 1m to 15m charts.
- One alert, one purpose. Fewer alerts per chart means less to dispatch and less to debug.
- Unique ID in every message so your receiver drops duplicates.
- Heartbeat alert. Fire a known alert every few minutes. If your log stops showing it, your pipeline is down and you should stop trading automatically.
- Reconcile positions. Compare what your system thinks is open to what your broker says, on a schedule.
- Staleness filter. Reject any alert older than a set number of seconds at receipt, using the timestamp in the message.
- Redundant channel. Send critical signals to a second route, such as push plus a chat channel, so one failure does not blind you.
- Size for the tail. Assume a late fill on your worst latency percentile and check that the trade still makes sense.
A broader pre-flight list is in our automated trading checklist for retail traders, and build or buy a unified alerts dashboard compares the options for centralizing alerts from several charts.
Our perspective
We build signal tools for 1m to 15m traders, so we see the same complaint every week: "the alert was late." Our view is that reliability is a chain, and most traders only look at one link.
First, a signal that repaints is unreliable before it is ever sent. That is why our indicators confirm on candle close, and why our scripts are open-source: you should be able to read the logic that decides when an alert fires, not trust a screenshot.
Second, we aim for fast, native Discord webhook delivery inside our Command Center, and we say plainly that we do not control the charting platform's dispatch stage. Nobody who builds on TradingView does. Our job is to keep everything we control lean: simple conditions, bar-close logic and clear alert messages.
Third, we tell members to measure. Run the {{timenow}} test on your own setup, then pick a timeframe your measured tail can survive. Slower charts forgive slow alerts. For help choosing, see our scalping timeframe selection guide.
Fourth, reliability does not equal profit. Even a perfect alert pipeline only delivers a hypothesis, and risk management decides whether you survive being wrong. Our trader risk management checklist is where we tell members to start.
— Tran
Build a reliable setup with Scalping-Algo
If you want signals that confirm on close, alerts that go straight to Discord, and code you can read, our TradingView suite was built for that workflow. It includes Smart Scalping Signals, Edge Finder and the Momentum Oscillator, plus the Command Center dashboard, education and community sessions.
Plans on our pricing page: $79 per month, $799 per year, or $1,999 one-off for lifetime access.
Start with one chart, one timeframe and the latency test above. Add automation only after your numbers hold up.
What regulators say about limits
Better alerts do not remove market risk. The CFTC warns that AI trading bots and signal strategies are often marketed with promises of unreasonably high or guaranteed returns, and that AI cannot predict sudden market changes. It also reports that roughly two out of three retail forex traders end each quarter in the red.
If a third party places trades in your brokerage account, the SEC calls auto-trading arrangements highly risky and says they need careful due diligence on registration and track records. Our tools send alerts and never trade your account for you. If you automate, you hold the keys and the responsibility.
On the broker side, SEC staff note that firms with market access must maintain risk controls for all orders, and that errors in automated processes can compound quickly across venues. Your own system needs the same mindset: a daily loss limit, a maximum position size and a kill switch that works without an alert arriving.
If you trade with real savings, consider speaking with a registered financial professional about whether short-term trading fits your situation. Check any signal seller's claims against the CFTC and SEC pages above, and be skeptical of anyone promising returns.
FAQ
How fast should a scalping alert arrive?
Faster than your edge decays. In the 34,174-alert TradingView study, the median was 4.03 seconds on 1-minute candles and the 99th percentile 16.1 seconds. If your setup needs a fill within a couple of seconds, test it yourself and consider a slower timeframe. Your measured 99th percentile is the number to plan around.
Are TradingView webhooks guaranteed to deliver?
No. They are fire-and-forget requests without retries or acknowledgments, so a missed message is not resent. Use a receiver that logs everything, add a heartbeat alert, and reconcile positions against your broker on a schedule.
Why did my alert fire but the chart signal disappear?
Most often the indicator repaints on the open bar, or the alert frequency let it fire before the candle closed. Switch to bar-close frequency and use an indicator whose signals confirm on close. Then compare alert timestamps with candle closes to verify.
Is alert latency the same as order execution speed?
No. Alert latency is the time from candle close to message receipt. Execution latency starts when your order leaves your system and ends at the fill. You pay both, plus slippage, so measure each one separately.
Can reliable alerts guarantee profits?
No. Alerts only report that a condition was met. The CFTC cautions that automated tools cannot predict the future, so pair any signal with position sizing, stop rules and a loss limit you set in advance.
Sources
- TradingView Webhook Delay: 34,174 Alerts on Real Latency (TV-Hub)
- Can TradingView Place Trades By Itself? (StaxInvesting Blog)
- FINRA Regulatory Notice 22-04
- SEC: Tips for Online Investing
- SEC: All About Auto-Trading
- SEC staff: Responses to Frequently Asked Questions
- CFTC Customer Advisory: AI Won't Turn Trading Bots into Money Machines
- CFTC: Forex Frauds
- Trading VPS requirements to verify