Fully automated XAUUSD system / 1 EA + 4 indicators · MT4 & MT5
A 35.9% win rate, and in the MT5 backtest $10,000 compounded to $96,150. It doesn’t need to guess right. It takes the loss fast when it’s wrong and holds on when it’s right: the biggest winner was 7× the biggest loser.
If you keep shopping for indicators, buying the arrow, holding the loser and grabbing a few dollars on the winner, another indicator won’t fix it. What’s missing is a set of rules that settles the entry and the exit before you click. ChanSMC is that set of rules. The ChanSMC_Trader EA opens the trade, sends the stop, adds, trails and closes on its own, while four companion indicators draw centres, structure and key levels live on the chart. Nine rules, each tested on its own against 2017–2026 XAUUSD M5 data. Any rule the backtest didn’t back up was cut.
On sale now. The three core indicators can be downloaded here and used free for a month. The full-system trial and paid licences are handled on Telegram (@up9933) or WeChat (up98999). The third-party live record isn’t finished yet, so read the verification log before you pay for anything.
Walk through a real week · 0:58
Unedited terminal screenshot, 16–23 Oct 2026, SmartExit channel only: green = long, red = short. Click to enlarge.
MT5 Strategy Tester backtest · XAUUSD M5 · 2023-01-01 – 2026-10-21 · $10,000 deposit · commission and swap included · compounded = the same trade list re-sized at 0.02 lot per $10,000 of balance · backtest results, not live trading, not a guarantee of future returns · full report
Backtest report
Every figure below comes from the MetaTrader 5 Strategy Tester report for ChanSMC_Trader v4.76. The original report screenshots are at the end of this section and the test settings are listed below.
The same 2,869 trades, sized at 0.02 lot base per $10,000 of balance (rounded to 0.01 lot), profits rolled into the next trade: $10,000 → $96,149.65, 81.4% a year.
| Year-end balance | Compounded | Fixed 0.02 lot |
|---|---|---|
| 2023 | $11,672 | $11,881 |
| 2024 | $13,715 | $13,656 |
| 2025 | $28,486 | $21,677 |
| 2026-10-21 | $96,150 | $35,251 |
The largest single loss was −$286.04. The best trade was about 7× that.
More than six trades in ten lose, and the account still climbs: the average win is 2.84× the average loss.
Median hold is 2.6 hours on losers and 6.9 hours on winners. Out early when wrong, in for the ride when right: almost all the profit comes from the big moves it stays in.
Test settings
| Holding time | Trades | Win rate | Net |
|---|
Original MT5 report screenshots
Hypothetical performance notice: the results above are historical backtest (simulated) results, not a live trading record, and they neither represent nor guarantee future returns. Backtests are designed with hindsight and do not involve real fills, and may over- or under-state the effect of liquidity, spreads and slippage. Leveraged trading can lose all of your capital. Full disclosure at the bottom of the page.
The edge
Jesse Livermore said it never was his thinking that made the big money for him. It always was his sitting. This system puts the sitting into code. When it’s wrong, it takes the loss straight away. When it’s right, it adds once at 2R, moves the stop to entry and trails it until the structure says the move is over. Each rule below has a matching input in the source, and each was tested against 2017–2026 XAUUSD M5 data.
Inclusion → fractals → pens → centres → MACD divergence → class 1/2/3 buy/sell points, all by fixed rules. Three overlapping pens make a centre, with ZG and ZD on the chart. Confirmed signals never repaint, so what you see now is what the trade was based on. Only the newest pen can still move, and it’s marked.
ChanTheory_Gold v1.73A close through the prior swing (BOS / CHoCH) confirms the structural top or bottom and the channel changes colour: one bar in a trend, three in a range, to keep whipsaw down. Once the reason for the trade is gone, so is the trade. No hoping, no holding.
SmartExit v9.13The stop goes in with the order, at structure or the channel edge. Setups that need more than 5× ATR are skipped. Risk per trade is capped at InpLots × $15,000 ($300 at 0.02 lot); a wide stop means a smaller size, and if even the minimum lot breaks the cap, there’s no trade. Across 2,869 backtest trades the largest single loss was $286.04.
InpMaxStopATR · risk capM5 decides when, H1 and H4 decide how much. Both higher timeframes agree: the main position doubles. Only one agrees: it’s halved. Big size goes where the wind is at your back.
InpUseMTFSizeAt +2R one add-on goes in with the trend and the main stop moves to entry, so if price turns, that trade closes about flat (gaps and slippage aside). From +3R the stop follows the channel edge bar by bar and never moves back. In the backtest, the 194 trades held 12 hours or more won 85.1% of the time and netted $24,560, close to the entire profit.
InpUsePyramid · InpTrailOnBandAn entry on the wrong side of yesterday's pivot P is half size. After a main trade wins 2R or more, the next one is half size too, because the first flip after a big move is usually a fake.
InpPSideMult · InpPostWinRAfter a leg of 20× ATR or more, the next three flips are skipped. No entry with spread above 30 points. No new trades 30 minutes either side of NFP, CPI, Fed decisions and press conferences (v4.77). Flat before the Friday close, nothing held over the weekend.
cool-down · spread · news filterA daily loss breaker, “stop after a good day”, cutting at −0.7R, calling ranges in advance, standalone mean reversion. They all sound safe and they were all tested. After costs each was worse or unstable, so each was cut, with the reason noted in the code.
2017–2026 · M5 · costs includedThe status panel shows channel side, the flip trigger, H1/H4 colours, the stop it would use right now, spread and pivot position. The EA also logs every trade, every flip decision and an hourly account snapshot: which trades it took, which it skipped, and why.
MT4 & MT5 · ledgers · panelVideos
Start with the narrated film, then a walk across a real week of the live chart. The three short explainers break the method into single ideas you can check against your own chart.
Free download
Put the SmartExit long/short channel, the Chan centre engine and the daily pivots on your own MT4 / MT5 and test them on live prices. Compiled builds only (.ex4 / .ex5), no source code; valid for a month from first activation, then they switch off automatically.
Green = long, red = short. BOS / CHoCH confirmed on the close (one bar in a trend, three in a range), plus a four-line panel showing the flip trigger and the stop level.
Inclusion, fractals, pens, centres, MACD divergence and class 1/2/3 buy/sell points, computed automatically; confirmed signals don't repaint. Popup, push and e-mail alerts.
Pivot P and R1–R3 / S1–S3 from yesterday's high, low and close; switches to weekly / monthly pivots automatically with the chart timeframe.
Trial rules: 30 days from first activation, tied to your trading account and this computer; re-activating does not extend it, and the indicators switch off when it ends. To keep using them, buy the full system or get install help: Telegram @up9933 · WeChat up98999.
Live charts
These are unedited screenshots from the developer’s own MetaTrader terminal. Centres, divergence labels, buy/sell markers, pivot levels and the status panels are all drawn by the software itself. They show what the tool displays. They are not a performance record.
Want to see today’s chart?
Message us on Telegram for a screenshot of the long/short display on the current market, or take the free 1-month trial and watch it on your own terminal.
The problem
Dozens of indicators, a dozen parameter sets, and the account still slides back to where it started? It usually comes down to three things.
Every entry has a reason. Every exit has a mood. Up a few dollars, you grab it before it turns. Down, you give it “one more bar” and a small loss becomes a big one. By the end of the year the winners are small and the losers are big.
The same chart gets drawn one way today and another way tomorrow. Support and resistance always look right afterwards and never when you’re clicking. Anything you can’t write down as a rule, you can’t repeat.
Most indicators tell you when to get in. Almost none tell you when to get out. Yet what a trade makes or loses is decided the moment it closes. Without exit rules, even good signals can’t hold on to the profit.
What it is
The system has three parts: the SmartExit exit engine, the ChanSMC structure engine and a risk layer. Below is what each one does. Nothing here is about what it might earn.
The developer has traded gold live for 11 years.
SmartExit (v9.13) reads SMC structure and an adaptive moving-average channel. A close through the prior swing (BOS / ChoCH) locks the structural top or bottom, and the channel changes colour: after one closing bar in a trend, after three in a range, to cut down whipsaw. When the structure the trade was based on breaks, the position is closed.
From Chan theory: inclusion handling → fractals → pens → centres (zhongshu) → MACD divergence → the three classes of buy/sell point. A centre is built by fixed rules (three overlapping pens, with ZG and ZD), not drawn by eye. Points confirmed by a later fractal never change. Only the newest pen, still unfinished, can move, and the indicator marks it.
Plenty of sellers put “safe” or “prop-firm ready” on a banner and publish nothing behind it. The EA’s actual sizing and stop rules are listed below, including the ones that add size, so you can read them before you decide.
Opens on channel flips, attaches the structural stop, manages pyramiding, trailing and the Friday close.
ExecutionBOS / ChoCH structure, adaptive channel, four-line status panel with live stop level.
Exit enginePens, centres, divergence and class 1/2/3 buy-sell points; alerts by popup, push or e-mail.
StructureDaily / weekly / monthly Fibonacci pivots that switch automatically with the chart timeframe.
LevelsBox + EMA34 view for quiet sessions: buy the lower edge, sell the upper edge, KD-confirmed.
CompanionThese are the EA’s design rules, written out so you can check them against how it behaves and against its inputs. They describe how risk is controlled. They are not a promise about results and don’t guarantee passing any firm’s evaluation.
The protective stop is sent with the order, placed at structure / the channel edge. If that stop would be wider than 5× ATR, the trade is not taken.
Money at risk to the stop is capped at InpLots × $15,000 (0.02 → $300). A wide stop shrinks the size; if even the minimum lot breaches the cap, the trade is skipped.
The EA never adds to a losing position and never layers orders at intervals. One main position at a time.
H1 + H4 both aligned: ×2. Only one aligned: ×0.5. Entry on the wrong side of the daily pivot P: ×0.5. After a win of 2R or more: next trade ×0.5.
After a losing trade the next main position is one step larger (max 2× base), and it resets on the next win. It is on by default and switches off with InpUseLossAdd = false. There is no uncapped martingale.
At +2R one add-on position is opened and the main stop moves to break-even; from +3R the stop trails the channel edge.
Spread above 30 points: no entry. After a move of 20× ATR or more, the next three flips are skipped (cool-down). Positions are flattened before the Friday close.
A daily loss limit was tested and not adopted because it cut the following winners. If you need one, set it at account or broker level; the EA does not enforce it.
Trade anatomy
These are explainer charts. The price path is made up, but everything is drawn by the documented rules, and each picture shows one mechanism. For the real market, see the live charts above.
Explainer charts and the three animated explainers use an illustrative price path, not a recorded market. Real charts are in the Live charts section.
How it works
The loop is simple, and no step needs a human decision at the moment of execution.
The EA runs on one XAUUSD M5 chart in MetaTrader 4 or 5. That’s the only configuration it has been validated on.
Centres, structure breaks and channel state come from closed bars, so everyone running it gets the same result.
Entries and trade management trigger on channel flips and R-multiples, not on judgement calls or alerts you chase by hand.
From the moment a position opens, its structural stop, its trailing rule and its flip exit are already attached.
Side by side
This compares common patterns in retail gold EAs. It isn’t about any named product.
| Topic | Typical gold EA | ChanSMC 金枢 |
|---|---|---|
| Grid / averaging down | Common, often unadvertised | None, and it can’t be switched on |
| Size after a loss | Uncapped martingale doubling | One optional step, max 2× base, resets on a win. Disclosed, and can be switched off |
| Stop-loss | Hidden, virtual or none | Real structural stop sent with every order |
| Structure | Repaints, redrawn zones | Rule-built centres; closed-bar signals don’t change |
| Parameters | Locked away | Published below, with defaults |
| Weak conditions | Never mentioned | Named: ranging markets, with examples |
| Proof | Cherry-picked profit screenshots | Full backtest report + original screenshots + settings; live record pending third-party verification |
Verification log
Backtest research, a system running every day, and a public live record anyone can check are three different things, so they’re listed separately here. Backtest figures are published with their settings. No live performance is claimed until the third-party record exists.
“Backtest research spanning roughly ten years (2017–2026, XAUUSD M5, including tick-level testing in MetaTrader) exists, and the system currently runs daily on both MT4 and MT5. Neither of those is a public live track record. A real-account public record, verified by a third party (MQL5 Signals / Myfxbook), will take 6–12 months to build and isn’t finished yet. Paid licences went on sale before it. No live results are claimed on this page. The figures shown come from an MT5 backtest, with the settings stated.”
| Item | What it is | Status |
|---|---|---|
| Backtest research | A full MT5 Strategy Tester run (2023-01 – 2026-10, default inputs) is published in the Backtest report with the original screenshots and settings; the earlier 2017–2022 work is a Python bar-by-bar replica, published as method and conclusions only. | Published |
| Daily operation | The system runs daily on MT4 and MT5 in its validated configuration (XAUUSD, M5). The live charts above come from that operation. Daily operation shows the software works. On its own it isn’t a verified performance record. | Running |
| Public live record | A real account published on a third-party platform (MQL5 Signals / Myfxbook) that computes the statistics itself. This is the only form of live performance evidence this project will ever cite. | In progress |
| Commercial launch | Paid licences are sold directly: Telegram consultation, payment by bank transfer, WeChat Pay, Alipay or crypto, and manual licence delivery bound to your MT4/MT5 account number. Platform storefronts (MQL5 Market, Whop) may follow. | Open |
Every backtest figure states the platform, period, deposit, lot size and cost assumptions, with the original report screenshots and the CFTC Rule 4.41 notice below. A backtest is not live trading; live performance will only be cited from a third-party platform.
The method depends on market conditions. In quieter, range-bound years (roughly 2017–2022), this kind of structure-following has done clearly worse than in strong trends. Expect drawdowns and flat stretches, and size positions with that in mind.
The method was developed and validated for XAUUSD only. It isn’t offered for other symbols or timeframes and hasn’t been tested on them, so how it behaves there is unknown.
Public live record (link to come)
When the third-party record is published, the platform link and how long it has been running will go here, and any performance figure this site quotes will come from that page. Until then, this space is left empty.
Parameters
The table uses the EA’s actual input names. The full manual comes with the product.
The EA plus its indicators, installed through the MetaTrader terminal in the
standard way. InpLots is the only sizing input you need to set: the
total-lots ceiling and the per-trade dollar cap are derived from it automatically.
Licence validation for copies sold outside the MQL5 Market is bound to your account
number.
This tool was developed, tested, and is being verified exclusively on XAUUSD (gold) on the 5-minute timeframe. Running it anywhere else means running an untested system with real money. Results also vary between brokers because spreads, execution and data feeds differ. On very small accounts the minimum lot can exceed the per-trade cap, and then the EA skips the trade.
| Parameter | What it controls | Default |
|---|---|---|
InpLots | Base lot size. Per-trade dollar cap = InpLots × $15,000; the total-lots ceiling is derived from it too. | 0.02 → cap $300 |
InpUseMTFSize | Doubles the main position when H1 and H4 channels both agree with the M5 flip (one agreeing: ×0.5). | On |
InpUseLossAdd | One-step size increase after a losing trade, max 2× base, reset on the next win. | On (set false to disable) |
InpUsePyramid | One add-on position at +2R; the main stop moves to break-even at the same time. | On |
InpTrailOnBand | From +3R, the stop trails the SmartExit channel edge. | On |
InpMaxStopATR | Skip the trade if the structural stop is wider than this many ATR. | 5.0 |
InpMaxSpreadPt | No entry when the spread is above this many points. | 30 |
InpCoolAfterBig | After a channel leg of ≥20 ATR, skip the next three flips (post-trend whipsaw filter). | On |
InpPSideMult | Size multiplier when the entry is on the wrong side of yesterday’s pivot P. | 0.5 |
InpPostWinR | After a main trade closes at ≥ this many R, the next main trade is half size. | 2.0 |
InpFridayClose | Flatten before the weekend; optionally restore the same direction on Monday. | On · 23:00 server |
Grid | Averaging into losers, layered orders at intervals. Not a parameter. This EA doesn’t do it. | None |
Pricing
There’s no checkout page and no automated store. Each licence is agreed in a Telegram chat, paid by bank transfer, WeChat Pay, Alipay or crypto, and set up by hand, bound to your MT4/MT5 account number. Start with the free indicators or the 1-month trial and see it on your own chart first. Read the verification log too: the public record isn’t finished, so what you’re buying today is a documented method and a working tool, not a completed track record.
¥0 for 1 month
SmartExit long/short channel, ChanTheory centres and FibPivot daily pivots. Download them here and run them on your own chart.
¥0 for 1 month
The complete system (SmartExit exit engine, ChanSMC structure engine and risk layer) with every feature on, free for a month.
¥56 / month
The full system, month by month.
¥168 / 3 months
The full system, three months at a time.
¥449 / year
The full system for a year, paid once, ¥223 less than paying monthly (12 × ¥56 = ¥672).
¥1600 one-time
The full system, licensed for good. Only a limited number are available, so ask whether one is open.
How it works: Message us on Telegram → pay by bank transfer (UnionPay / WeChat Pay / Alipay) or crypto (USDT / BTC / ETH) → your licence is delivered on Telegram, bound to your MT4/MT5 account number.
No discounts, countdowns or coupons. The price here is the price today and next month.
Payment
Two ways to pay, both handled in the same Telegram (or WeChat) chat. Payment details aren’t published on this site; the developer sends them to you in the chat once you’ve picked a tier. Bank details and wallet addresses on a public page tend to get copied onto scam sites.
Pay in RMB through your usual channels.
Pay from any wallet or exchange. USDT on TRC20 is recommended; the network fee is about $1.
Step by step: Message us with the tier you want and your MT4/MT5 account number → we send the payment details (bank or crypto) → you pay and send the transfer receipt or transaction ID → your licence is delivered in the same conversation, bound to your account number.
Prices are in RMB. Bank charges and network fees are the sender’s. Please check your MT4/MT5 account number before paying. The licence is bound to it and activated by hand.
Read the Refund Policy before you pay. Paid licences aren\'t refundable once delivered. The full system has a free 1-month trial.
Contact
Everything on this page runs through one conversation.
Free indicators, trial activation, paid licences, screenshots of the current market and any questions all go through a direct Telegram chat with the developer. There’s no ticket system, sales team or checkout form. You send a message, pay by bank transfer or crypto if you pick a paid tier, and the licence arrives in the same chat.
Message @up9933 on Telegram
FAQ
These are the questions that decide whether it’s right for you.
No. Be wary of any tool that says it does. This system will have losing trades and losing periods, including long drawdowns and flat stretches in ranging markets. Past results and backtests don’t guarantee future results. Only trade money you can afford to lose.
No. ChanSMC is a software tool. It applies a fixed, documented method to chart data and follows predefined rules. It gives no personal recommendations, doesn’t look at your circumstances, sends no trade calls and manages no accounts. Which account you run it on, with what settings and risk, and whether you trade at all, are your decisions.
XAUUSD on M5, nothing else. That’s what it was built and tested for, and what the live verification covers. It hasn’t been validated on other instruments or timeframes, and we make no claims about how it behaves there.
A public record needs time on a real account, with the numbers worked out by a third party, and that can’t be rushed. Roughly ten years of backtest research (2017–2026) exists and the system runs daily on MT4 and MT5, but neither is a third-party-verified live record and this site won’t present them as one. The verified record (MQL5 Signals / Myfxbook) will take 6–12 months. Licences went on sale before it, which is why the gap is stated here rather than left for buyers to find out later.
The trial runs for 30 days from first activation and is tied to your trading account and computer. When it ends, the indicators remove themselves from the chart and leave an "expired" notice; reinstalling, switching accounts or changing the PC clock does not restart it. To keep using them, ask about a licence on Telegram @up9933 or WeChat up98999.
Yes. Message @up9933 and ask for a real-time screenshot of the long/short display on the current XAUUSD M5 chart. Better still, take the free 1-month trial and watch it on your own terminal, where nobody can pick what you see.
No grid, no averaging down, no uncapped martingale. There is one rule that adds size after a loss: the next main position is one step larger (at most 2× the base lot) and it resets on the next win. It’s on by default because the backtest ran with it on. It’s listed in the risk specification and the parameter table, and it switches off with InpUseLossAdd = false. Every order still carries its stop, and the per-trade dollar cap applies after every multiplier.
Yes. Message us on Telegram (@up9933), say which tier you want, pay by bank transfer (UnionPay / WeChat Pay / Alipay) or crypto (USDT, BTC or ETH), and your licence is delivered in that same conversation, bound to your MT4/MT5 account number. The three indicators can be downloaded from this page for a free 1-month trial, and the full-system trial is set up on Telegram. Start with those. Telegram @up9933 and WeChat up98999 are the only places it’s sold. Anyone else offering it isn’t us.
No. Once a paid licence has been delivered it isn't refundable. The reasons, and the few cases we do cover (no delivery, double payment, a software defect we can't fix), are in the Refund Policy. Use the free 1-month trial first and decide on your own charts.
Who shouldn’t buy this
If any of these sound like you, this system isn’t for you. Please don’t buy it.
Next step
Download the three indicators for a free 1-month trial, ask for the full-system trial, or ask for a screenshot of today’s market, and check the rules against real candles. It’s all done on Telegram. Read the disclosure below before you decide anything.
What this product is. ChanSMC, including the ChanSMC Trader EA and the SmartExit indicator, is computer software licensed for use as charting and trade-execution utilities. The publisher is a software developer, not a Commodity Trading Advisor (CTA), not an investment adviser, not a broker, and not a financial institution. Nothing on this site is an offer, solicitation, or recommendation to buy or sell any instrument, and nothing here takes into account your objectives, financial situation, or needs. It is not a signal service, not a copy-trading service, and not account management. This site carries no broker referrals and no introducing-broker (IB) or affiliate links. If any other material conflicts with this section, this section prevails.
General risk disclosure. Trading foreign exchange, contracts for difference (CFDs), spot gold, and other leveraged products carries a high level of risk and may not be suitable for all investors. The high degree of leverage can work against you as well as for you. Before deciding to trade, carefully consider your investment objectives, level of experience, and risk appetite. You could lose some or all of your initial investment and, with some products and venues, more than that. Do not invest money you cannot afford to lose, and seek advice from an independent, appropriately licensed financial adviser if you have any doubts. Automated trading software does not remove these risks: it executes rules, and rules can be wrong, can be applied to unsuitable conditions, and can fail technically (connectivity, platform, or data-feed failures) at the worst possible time.
Hypothetical performance (CFTC Rule 4.41 notice). Hypothetical or simulated performance results, including any backtest of this system, past, present or future, have certain limitations. Unlike an actual performance record, simulated results do not represent actual trading. Also, since the trades have not been executed, the results may have under- or over-compensated for the impact, if any, of certain market factors, such as lack of liquidity. Simulated trading programs in general are also subject to the fact that they are designed with the benefit of hindsight. No representation is being made that any account will or is likely to achieve profits or losses similar to those shown by any simulated or historical test.
Jurisdictions and licence terms. Availability of leveraged trading products, and of this software, varies by jurisdiction and may be restricted where you live (for example, retail CFDs are not available to residents of the United States). You are solely responsible for ensuring that your use of any trading product or venue is lawful where you are. Use of the software is governed by the licence terms confirmed with you before payment, covering licence scope, activation, permitted and prohibited uses, the refund policy and limitation of liability. This page does not itself grant any licence.