Gold Trend MT4 Indicator: How to Spot a Repainting One

You dragged a gold trend MT4 indicator onto your XAUUSD chart and it looked perfect. A neat trail of buy arrows along every rally. Sell arrows sitting right on the tops. Six months of gold history, and the tool called almost all of it.

Then you turned it loose on a live account, and it stopped looking perfect.

The arrows moved. A buy signal that existed yesterday was gone today, replaced by an arrow three bars later. A sell signal that sat at the exact high in your screenshot now sits twelve bars lower. The history kept healing itself, and the live signals kept arriving late.

That is repainting. It is the single most important thing to understand about any gold trend indicator you are thinking of trusting with real money.

This page is not a round-up of indicators. It is about one question you can answer yourself, tonight, on a demo account: does this tool change its mind about the past? If it does, nothing else about it matters.

Why is a repainting gold trend indicator worse than no indicator at all?

Because a tool that does not repaint at least tells you the truth, and you can learn to work around it, while a repainting tool lies in a way that is built to look like success. The damage is not that it loses trades. The damage is that it destroys your feedback loop, so you never find out what actually went wrong.

Define the term once, because it gets used loosely.

A repainting indicator is one whose historical output changes after the fact. The value it shows on a past bar is not the value it showed while that bar was live. It recalculates, redraws, or deletes signals as new price data arrives, so when you scroll back through the chart, the indicator appears to have been right about moves that had not happened yet when the signal was given.

There are three common types, and they get worse in this order.

Recalculation. The indicator recomputes its buffer on every new bar. A signal that appeared at the close of a candle may move or vanish when the next candle forms. This is usually a coding choice rather than trickery. The developer read Close[0] — the price of the bar still forming — instead of Close[1], the last closed bar. The number is real, but it was never final.

Redraw. The tool keeps only the most recent leg of a move and erases the previous one when direction changes. ZigZag-style tools do this by design. On a gold chart, a redrawing trend line always connects the last significant swing to the current price, so it looks clairvoyant while it is really just describing where price already went.

Future repaint. The worst kind. The indicator writes a signal using information from bars that arrive after that signal’s timestamp. In a backtest it produces the smooth, near-perfect curve you saw on the sales page. Live, the same signal arrives late or not at all, because the future data it leaned on does not exist yet. This is look-ahead bias, and when it is deliberate it is a way of flattering every backtest built on the indicator.

Why does this bite harder on gold than on EURUSD? Because XAUUSD moves in long, fast legs and then reverses violently. A tool that draws its trend after the fact will look superb on that kind of chart, and a trader who trusts it enters exactly as the leg ends. If you want the long version of this problem, our walkthrough of the accurate signals no repaint indicator covers how vendors dress up the same trick, and the best non-repainting arrow indicator comparison shows what a fixed, honest signal looks like on a chart.

Here is the arithmetic that should end the argument. A trend filter that repaints shows a near-perfect hit rate on your history and a coin-flip hit rate live. Those two numbers cannot both be true, so one of them is fake. You will not discover which until real money is on the line — unless you run the tests further down this page first.

Blue is not a signal

One more trap that confuses this whole topic. On most gold trend indicators, the colour is not the signal. The blue stretch is simply the last computed value of a moving average or a momentum reading. A blue bar does not mean “buy”. It means the line is currently above its own threshold. Traders who read colour as a command end up chasing entries two bars after the move began, and then blame the indicator for being late. Read the rule, not the paint.

Worse than nothing, in one sentence

No indicator means you decide from price structure and you own the outcome. A repainting indicator means you make the same decisions but attribute them to a system that was never real, so you repeat the mistake forever.

What did a repainting gold indicator actually cost a real trader?

It cost Maya about $1,900 and four months, and the money was the smaller part. Her story is worth reading because nothing dramatic happened, which is how most of these losses really occur.

Maya had traded gold for a year. Not badly. She had a simple method: trade with the daily direction, enter on a pullback, risk 1% per trade. She was roughly break-even and she knew her real problem was second-guessing entries.

A friend sent her a gold trend indicator — a compiled .ex4 file, no source, sold under a name she had never heard. She put it on a demo account first, which is exactly what a careful person does. Over three weeks it printed 41 signals and 34 were winners. The drawdown was almost invisible. So she copied the settings onto a live account.

That is when the two lives of the same indicator split apart.

On live gold the tool produced 38 signals in three weeks and 19 winners. Signals arrived a bar or two after the moment they appeared when she refreshed the history. A buy arrow present at 10:04 was sometimes gone by 10:06. She began to doubt her own reading of the chart, then began overriding the tool, then stopped following any plan at all. By the end she had also paid for a “pro” upgrade, which is where most of the $1,900 went.

Two costs. The first is money. The second is that she nearly concluded technical tools do not work on gold, which is wrong and would have sent her back to trading on feeling. The tool was the problem. It was reading the future in the backtest and the present in the live account, so no amount of discipline could have saved it.

The fix was cheap. An afternoon, a notepad, and the four tests below.

How do you test a gold trend indicator for repainting yourself?

You test it by recording the same signal twice — once while it is live, once after the bar has closed — and checking whether it stayed put. If the live signal and the historical signal differ, the indicator repaints, and the size of the difference tells you how badly.

Everything below runs on a demo account this evening. The first three tests need no programming. The fourth settles the argument for good.

Test one: the bar-by-bar replay

Open the MetaTrader 4 Strategy Tester, load any Expert Advisor against your gold symbol, and switch visual mode on so the chart plays in its own window. You are not testing the EA. You are using the tester as a replay machine, which is the closest thing MT4 has to stepping through history one bar at a time.

Do this:

• Attach the trend indicator to the visual-mode chart.
• Set the replay to the slowest speed, or pause it and step forward bar by bar.
• As each new bar forms, write down the signal — buy, sell, or nothing — and the bar’s time.
• Let it run for at least 200 bars, and screenshot the chart at the start of that window.
• Scroll back to your first logged bar and compare the arrows shown now against what you wrote down.

If the arrows do not match your log, the indicator repaints. A reliable tool passes this unchanged, bar for bar, because it only ever reads values from closed bars.

Test two: the refresh-and-wait test

This is the fastest check and it catches careless repainting. Put the indicator on a live XAUUSD chart. Wait for a signal to appear on a bar that is still forming, and note the time. Do not act on it. Now do nothing to the chart for five minutes, then look again.

A repainting signal will often move or vanish before the bar even closes. Then right-click the chart and choose Refresh, and switch to another timeframe and back. If the historical arrows rearrange themselves, you have your answer in under ten minutes.

Test three: the chart-reload test

The last of the no-code tests is the bluntest. Take a screenshot of the chart with a date visible in the corner. Close the chart entirely. Reopen the same symbol and timeframe. Compare the two images.

An honest indicator draws the identical history. A repainting one draws a tidier version of the past than the one you photographed, which is precisely the lie that makes its backtest unusable. Keep the screenshots; dated evidence beats memory when you are deciding whether to trust a tool with money.

Test four: the source-code audit

This is why an indicator that ships as readable .mq4 is worth more than a compiled .ex4, and it is the reason for the download on this page. Open the file in MetaEditor or any text editor and search for these patterns:

• Reads of the current bar inside the signal logic. Close[0], High[0], Low[0]. If the signal depends on the price of the bar that is still forming, it can and will change before the bar closes.
• ObjectDelete next to a draw call. The tool is erasing something it drew, usually the previous signal. That is redraw behaviour by definition.
• Buffer values written for old bars. Code that sets a buffer value at a shift greater than one is rewriting history on purpose.
• SetIndexDrawBegin and IndicatorBuffers set up oddly, so the visible line is only ever the newest few values.
• Negative shifts. Any index like Close[-1] or a loop that reaches into the future is look-ahead bias, and it will never reproduce live.

A trend filter that only reads shift >= 1 and never redraws past output will not repaint. You now know that without trusting anyone’s marketing.

Test five: the EA-versus-arrows test

The final check is the one that cannot be argued with, if you can write a little MQL. Build a tiny EA that calls the indicator through iCustom and reads its buffer at shift 1 — the last closed bar — and logs the value. Run it in the Strategy Tester on real tick data over the same period you watched visually.

If the signals the EA records match the arrows on the chart, the indicator is fixed. If the visual arrows are more accurate than the EA’s logged signals, the visual arrows were drawn using data the EA could not see at the time. That gap is repainting, measured in numbers rather than impressions.

The takeaway from all five tests is the same. Before you believe any gold trend tool, you must be able to show that the signal you would have traded live is the signal the chart now displays. There is a related problem with tools that redraw market structure rather than arrows, which the supply and demand no-repaint guide explains in the same spirit.

What is a trend filter actually good for on XAUUSD?

A trend filter is a permission gate, not a prediction engine. It answers one question — “am I allowed to trade in this direction right now?” — and it is genuinely useful for that and for very little else.

That framing changes how you use it. You are not asking the indicator where price is going. You are asking it whether the current direction is one you are permitted to trade, and letting your own plan handle everything after that answer.

Where a trend filter earns its place

• Filtering counter-trend entries. Most losing retail trades on gold are trades against the higher-timeframe direction, taken because a small move looked exciting. A trend filter removes that category of trade before you see it.
• Reducing the number of decisions. A filter that says “long only today” cuts your option set in half. Fewer decisions means fewer impulsive ones.
• Keeping you consistent. A rule you follow every day beats a judgement call you make when you are tired. The filter does not need to be clever. It needs to be the same rule each session.
• Working with structure, not against it. On gold, price respects obvious swing highs and lows. A trend filter aligns your entries with that structure instead of fighting it, which is easier to see on a higher timeframe than on the M5 chart you are staring at.

Where a trend filter cannot help you

• It cannot predict a reversal. Every trend filter is a lagging tool. It confirms a direction that already exists. Expecting it to call tops on gold is how traders end up selling every rally during a bull run.
• It cannot time an entry. “Trend is up” is not an entry. You still need a pullback, a level, or a trigger of your own choosing.
• It cannot size a position. Risk per trade is your decision, and it is the decision that keeps you in the market. A filter has no idea how big your account is.
• It cannot place or manage your stop. The tool draws on the chart. It does not know your drawdown limit, your broker’s spread, or your maximum daily loss.
• It cannot protect you from news. Gold can move 300 points in minutes around a US inflation release. A trend filter will happily hold a direction straight through it.

If you want a worked example of combining a trend tool with reversal logic instead of treating either as an oracle, our breakdown of the top trend reversal indicators and the Pro Trend Line Indicator both stay on the honest side of that line. The pattern to copy is simple: let the filter set direction, let your own rules set entry and size, and keep the two jobs separate.

How do you read a trend indicator on gold without fooling yourself?

Read it only on a closed bar, confirm it on a higher timeframe, and treat it as one input rather than as the decision. That is the whole discipline, and it is boring on purpose.

• Only closed bars count. A signal on the bar still forming is a suggestion, not a signal. If you cannot wait for the close, you are trading the noise the indicator was supposed to filter out.
• Let the higher timeframe set the bias. Read the trend on H4 or Daily, and take entries on H1 or M15 in that direction. One reading per day, not one per five minutes.
• Confirm once, with one instrument. A moving average, a structure break, or a momentum reading is enough. Stacking five indicators until they all agree is not confirmation, it is inaction with extra steps.
• Log the signal before you trade it. Write the direction, the bar time, and the price. This log is the only evidence that will tell you later whether the indicator was fixed, and it takes ten seconds.
• Never let it override your risk limit. The filter may keep you out of trades. It must never let you into a bigger one.

Put the same discipline to a printed checklist you can follow at the screen:

Question	Honest answer or stop
Did the signal print on a closed bar?	If no, wait
Does the higher timeframe agree?	If no, no trade
Is there an entry trigger, not just a trend?	If no, wait
Is the position sized to your risk rule?	If no, size it first
Is a stop defined before entry?	If no, do not enter

Answer “no” to any row and the correct action is to do nothing. A trend filter’s most valuable output is often the trade it stops you from taking.

How much does lag cost you if all you want is an honest gold filter?

A trend filter is always late, and that is not a flaw — lag is the price you pay for removing noise. An average of past prices cannot turn before price turns. The only question worth asking is whether the lag you accept buys you more than the whipsaws it removes.

On a currency pair that trade-off can be close. On gold it usually is not, because the legs are large. XAUUSD can travel 400 to 800 points in a single directional push once a move is underway. If your filter is 150 points late at the start, you still capture most of the leg. The same 150 points of lag on a pair that only moves 200 points has eaten most of the trade before you enter.

That is why identical trend rules behave so differently across instruments, and why copying a bias rule straight from an EURUSD strategy onto a gold chart rarely survives contact with real spreads. The instrument changes the verdict, not the maths.

Measure it instead of guessing. Take any filter you are considering and count two things across 200 closed bars of gold: how often it flips direction while price is ranging, and how much of each genuine trend leg it misses at the start. A useful filter scores low on the first count and tolerable on the second.

Filter speed	What it catches	What it costs on gold
Fast (short moving average)	Turns early, reacts to every swing	Whipsaws hard in a range, flips on noise
Medium	Balances reaction against noise	Still late at turns, occasionally caught in chop
Slow (long moving average)	Holds a direction through pullbacks	Misses the start of each leg, gives some back at the top

The rule that follows is unglamorous. Choose the slowest filter you can tolerate, not the fastest one that looks responsive on a chart. If a filter flips three times in a week on gold, it is not a filter, it is a second price line, and you will trade it like one.

And never confuse lag with a reason to stack indicators. Adding a second filter to a slow one does not remove the lag. It adds a second opinion that will disagree with the first at the worst moment, which is how a clean rule becomes no rule at all.

One limit worth stating plainly: correct filtering does not change the fact that most retail accounts lose money on leveraged gold. A trend filter improves consistency. It does not remove risk. Risk per trade and a defined stop still do the heavy lifting.

What are the red flags of a marketed gold trend indicator?

Honest tools are specific and boring while marketed ones are vague and superlative, and the split shows up in the same places every time. Learn the pattern and you can judge a new gold trend product in about two minutes.

Check any tool against this list before you install it:

Red flag	What it usually means
No source code, compiled file only	You cannot check for repainting or look-ahead, and the seller does not want you to
A near-perfect backtest with almost no drawdown	The signals were drawn after the fact, or the test looked into the future
“Never repaints” with no way to verify it	A claim, not proof. Real proof is source code plus a bar-replay test you run yourself
Arrows pictured on completed moves	The screenshot was taken after price finished, so the arrows prove nothing
Balance screenshots instead of a track-record link	Evidence is a third-party link read from the broker’s server, not an image
Language about “recovering” losing trades	Usually a martingale or grid underneath, which hides risk instead of bounding it
Pressure to buy immediately	Honest tools are versioned and documented. Urgency is a sales device

Now the other side. The signs that a gold trend tool deserves your attention:

• The source is available and the licence is named. MIT, GPL or Apache, linked from the repository.
• A named author with a version history. A real person, a repository, commits you can read.
• One job, done plainly. It draws a trend line or prints a value. It does not pretend to be a whole strategy.
• You can reproduce the signal. Run the bar-replay test and the arrows stay where they were.
• It tells you what it does not do. Every honest tool states its limits before you ask.

None of that produces a profitable result on its own, because nothing can. What it produces is a tool that is what it says it is, which leaves only one variable for you to work on: the trader.

What is in this free download?

The download is the Channel Pattern Detector, an open-source indicator for MetaTrader 4 and MetaTrader 5 published under the Apache-2.0 licence by EarnForex, with the complete MQL4 and MQL5 source in the repository. It finds ascending, descending and horizontal channel patterns and marks them with lines directly on the chart, and it can alert you when a new channel is detected. It is the most literal version of the tool this page has been describing: the slope of the channel tells you which direction you are permitted to trade and the two edges tell you where the boundaries sit — and because the source is text, you can run every repainting test above against a tool that shows you exactly how it decides where those lines belong.

Spec	Detail
Files	Channel Pattern Detector source for MQL4 and for MQL5 (source, not compiled)
Platform	MetaTrader 4 and MetaTrader 5
What it draws	Ascending, descending and horizontal channel patterns, marked with lines on the chart
Signal output	Drawn channel lines, plus optional alerts when a new channel is detected; no trade is placed
Repainting status	Checkable — read the source, then confirm with the bar-replay test
Tuning	Inputs that relax or restrict the channel rules, so you can see how sensitive the detection is to your settings
Licence	Apache-2.0 — free to use, modify and share with the notice kept and changes noted
Author	EarnForex
Cost	Free
What it does not do	Not a trading system, no entry or exit rule, no gold preset, no verified live track record

Licence: Apache-2.0, verified against the repository on 29 September 2026. Apache-2.0 is a permissive open-source licence that also carries an explicit patent grant. You can use, modify and redistribute the code, including commercially, provided the copyright notice, the licence text and a note of your changes travel with it. You are not required to publish your changes.

What it does not do

This is the section most download pages skip, and it is the one worth reading.

• It does not place trades. This is an indicator. It draws lines on the chart. Every entry, size and stop remains your decision.
• It does not repaint-proof itself for you. You still run the tests. Source code lets you check honestly; it does not hand you the answer.
• It does not hand you a signal you can trade blindly. A channel is drawn structure, not an entry rule. You decide what to do at the edge, and the alert on a new channel is a heads-up rather than an instruction.
• It does not ship as a compiled binary. You compile the MQL4 or MQL5 source yourself in MetaEditor. That takes seconds and means you know exactly what is running.
• It does not come with a verified live track record. No third party has published results for it on a tracked account.
• It does not include a gold preset. The channel rules are general. You choose the symbol, timeframe and inputs, and you test them on your own broker’s data.
• It does not fix bad risk management. No channel does. Position sizing and stops are still the decisions that decide whether you survive.

If you would rather start from a curated set of licence-verified tools, the free download library holds the rest of the collection, and the /category/free-forex-indicator/ pages group them by tool type. The top ranking tables are a useful second reference when you compare indicators side by side.

Why does readable source code matter for a trend indicator?

Because a compiled .ex4 is a black box you are handing your chart and your decisions to, and you cannot test it for the one flaw that ruins trend tools. When the source is text, you can confirm the tool does not read the forming bar and does not redraw its own history.

Think about what you were asked to accept before this page. Someone you have never met sent you a file with no readable contents, claimed it identified gold trends, and left you to trust the compiled result. A repainting bug and a deliberate look-ahead trick look identical from outside the binary. They are not identical from inside the source.

What you are checking	Source-available MQL4/MQL5 (this download)	Compiled .ex4 from a vendor
Does it read the forming bar?	Readable in seconds	Undetectable
Does it redraw past signals?	Findable by searching the file	Undetectable
Look-ahead in the signal logic	Auditable bar by bar	Undetectable
Can you change the inputs	Yes, then recompile	Only the ones exposed
Who fixes a bug	You can	The vendor, if they still care
Cost	Free (Apache-2.0)	Often paid

None of that makes an open-source trend filter better at predicting gold. It makes it checkable, and for a tool whose entire value depends on whether its history is real, checkable is the only thing that matters.

What should you do in the next ten minutes?

Download the source for your terminal, compile the Channel Pattern Detector in MetaEditor, and put it on a XAUUSD chart on a demo account. Then run test one: replay the chart bar by bar in the Strategy Tester, log where the channel lines sit, and compare them with the lines the chart shows after the fact.

That single afternoon answers the only question that matters about a gold trend MT4 indicator. If the signals hold still, you have a trend filter you can reason about, and a clear picture of what it can and cannot do: set a direction, filter counter-trend entries, and stay out of the way of everything else. If they move, you have saved yourself the version of this lesson that Maya paid $1,900 to learn.

Test on a demo account first, every time. Compile the source yourself so you know what is running. Keep your risk per trade small and your stop defined before you enter. Then, and only then, decide whether a gold trend filter has earned a place in your process.

For more free, licence-verified tools to run alongside it, browse the blog and the best MT4 EA guides, or compare licensed options on the shop pages.

Trading forex and CFDs on gold carries a high risk of loss and is not suitable for everyone. Leverage magnifies both profits and losses, and most retail accounts lose money. No indicator, expert advisor or strategy on this page removes that risk, and nothing here is financial advice. Past performance, including every backtest mentioned, does not indicate future results. Only risk money you can afford to lose, and always test on a demo account first.