The appeal is obvious. You watch thirty instruments, you cannot look at all of them on four timeframes, and a scanner claims to point at the ones forming a setup. The mechanics behind that claim are worth understanding, because they explain both what the tool is good for and the specific way it will mislead you.
How detection actually works
Almost every pattern scanner starts with the same step: reduce the price series to a set of swing highs and swing lows. That reduction uses a fractal or zigzag rule, typically something like a bar whose high exceeds the highs of the N bars on each side. Change N and you get a completely different set of pivots, which means you get a completely different set of patterns. The pivot sensitivity setting buried in the tool's options is the single most important parameter, and most users never touch it.
From those pivots, the geometry is arithmetic. A symmetrical triangle is a sequence of lower highs and higher lows whose trendlines converge within a tolerance. A head and shoulders is five pivots with a specific height ordering and a neckline that the tool fits with a straight line. Harmonic patterns are the strictest case: they are defined by Fibonacci ratios between the legs, with a permitted deviation, so a scanner that finds a Gartley is really reporting that four ratios fell inside their windows. The underlying measurement is the same one described in the Fibonacci retracement guide.
The repainting problem
Here is the mechanism that trips people up. A swing pivot is only confirmed after N bars have printed to its right. That means the most recent pivot in any live chart is provisional, and if price extends, the tool deletes the pattern it drew ten minutes ago and draws a new one somewhere else. The historical chart then looks flawless, because every pattern on it was finalised with full hindsight.
This is the same failure family as indicator repainting, and it is why a screenshot of a scanner's historical hit rate tells you nothing. If you want to evaluate one honestly, record its live alerts as they arrive, timestamped, and score those. Anything else is measuring the past against itself, which is the trap in backtest overfitting.
Ask one question of any pattern tool: does the detection depend on bars that had not printed when the alert fired? If yes, its historical statistics are hindsight, no matter how the results are presented.
What the tools are genuinely good at
Scanning breadth. A human cannot watch forty symbols across the M15, H1, H4 and daily. A scanner can, and its output as a shortlist of charts worth opening is a fair use of the technology. Treat every hit as a prompt to look, never as a signal to click.
They are also good at the mechanical patterns where the definition leaves little room for argument. Inside bars, engulfing candles, opening range breaks and gap conditions are single bar or two bar rules with no pivot dependence, so they do not repaint at all. Those detections are trustworthy in a way that a five pivot harmonic never will be. The vocabulary is in candlestick chart basics.
Where they are weakest is context. A triangle sitting under a daily supply zone and a triangle in the middle of nowhere are the same shape to a geometry engine and completely different trades to anyone who has held both. Every serious user of these tools pairs them with a manual read of structure, along the lines of supply and demand zones and support and resistance.
Choosing one
Four things separate a usable scanner from a toy. The pivot sensitivity has to be exposed and adjustable, because a fixed setting is fitted to whatever the vendor tested on. It has to scan a defined universe with a stated refresh interval, so you know whether you are seeing live conditions or a five minute old snapshot. Alerts must be server side and deliverable to a channel you monitor, in the manner described in the alerts guide. And it should tell you which bar confirmed the pattern, so you can check the timing yourself.
On the commercial side, be sceptical of any scanner advertised with a hit rate. Nobody can publish a meaningful success percentage for pattern detection, because the exit rule, the stop distance and the instrument all change the result, and the vendor controls none of them. A tool that sells on published win rates is selling the wrong thing, which is a point we make more broadly in how to evaluate a signal service.
Using the output without getting hurt
The workflow that survives contact with a live account is narrow. Let the scanner produce a shortlist. Open each chart yourself and mark the higher timeframe structure. Discard anything without a level to trade against. Define the entry, the stop and the invalidation before the pattern completes, using the sizing arithmetic in risk management rules. Then let price come to you.
The scanner saves you the looking. It does not save you the deciding, and leveraged trading carries a high risk of loss, so the deciding is the part worth your attention.
"A scanner is a pair of extra eyes, not a second opinion. It tells you where to look. What is on the chart when you get there is still your call."
— Alex Onta, Executive Director, SINGUARD
Key Takeaways
- Pattern detection begins by reducing price to swing pivots, so the pivot sensitivity setting changes every pattern the tool finds.
- Pivot based patterns repaint because the newest pivot is provisional, which makes any historical hit rate a hindsight measurement.
- Single bar and two bar rules such as inside bars, engulfing candles and opening range breaks do not repaint and are far more reliable detections.
- Use the scanner as a shortlist generator, then mark structure and define the stop yourself before the pattern completes.
Frequently Asked Questions
Do chart pattern scanners repaint?
Any scanner built on swing pivots does, because the most recent pivot is not confirmed until several more bars print to its right. Patterns drawn in real time can be deleted and redrawn as price extends. Single bar and two bar detections such as engulfing candles or inside bars do not have this problem.
Are pattern scanners accurate enough to trade automatically?
Detection accuracy and trade outcome are different questions. A tool can identify a textbook triangle correctly and the breakout can still fail. Because the exit rule, stop distance and instrument all sit outside the scanner, no published success rate for pattern recognition is meaningful, and automating straight from detections is high risk.
What is the best way to test a pattern scanner?
Record its live alerts with timestamps as they arrive and score those, rather than reviewing its markings on a historical chart. Historical markings were finalised with information the tool did not have at the moment of the alert, so they flatter the result.
About the Author
Alex Onta is an Executive Director at SINGUARD. He built eTrader, the terminal, the mobile apps, eTrader Broker, Copytrading, Business and Community, along with the worldwide clustered-server infrastructure it all runs on, with his brother Roman Onta helping on the design, and he leads that division today. Together with Roman he builds the Prop Firm CRM, the Broker CRM, Scalegram and CopySignals, and the two of them carry worldwide compliance, payment processing and international business structuring side by side. He lives and works in Dubai for most of the year. Meet the executive duo leading Singuard's five divisions.