Gold Trend Lock EA is a free MetaTrader 5 (MT5) expert advisor for gold (XAUUSD) that opens a trade at the start of an hour in the direction gold has moved over the previous half day, with a fixed stop loss on every trade. For this 2026 review we read its tester logs trade by trade, matched our real-tick backtest against a $10,000 live account and ran it over five years of gold prices at every risk setting it ships with. Free download at the bottom.
The part worth reading is the exit. A profit lock at +$16 closes almost every winning trade, while losing trades run to the full $40 stop, so the lot setting you pick decides how deep the account falls.
TL;DR – Free MT5 gold trend robot: hourly trades, a $40 stop on every trade, up to two positions, no grid. Wins lock at +$16, so one stop costs 2.5 wins. Every full month of 2026 was positive in our real-tick test, but the default lot risks 13% per stop.
What is the Gold Trend Lock EA?
Gold Trend Lock EA is an MT5 robot for XAUUSD. At the top of an hour it opens a buy or a sell in the direction gold moved over the last 12 hours, with a $40 stop loss and a $25 take profit. Once a trade is $16 in profit, the stop locks that gain. It holds up to two positions and closes everything late on Friday.
Core Strategy and Logic
Everything below comes from the logs of our own tests: all of 2026 on Tickmill real ticks, plus 2022 to 2025 on generated ticks.
- Entries on the hour – Every trade is a market order sent in the first second of an hour (194 of 195 in 2026 at exactly hh:00:00). Entries run from 15:00 to 11:00 server time the next day, so it never opens a trade between 12:00 and 14:59, and in 2026 never at 00:00 or 01:00. It reads its own price data, so the chart it sits on doesn’t matter: H1, M15, M5 and H4 gave the same trades to the cent.
- Direction from the last half day – In 9 of 10 trades it buys when gold is higher than 12 hours earlier and sells when it is lower, which makes it a trend-following robot. It trades in about 1 in 20 of the hours inside its window, and the rule that picks them is not visible in the log.
- A $40 stop and a $25 target on every trade – All 195 trades of 2026 carried a stop loss $40 from entry and a take profit $25 from entry.
- The +$16 lock is the real exit – When a trade is $16 in profit, the EA moves the stop to +$16 and trails it in 5-cent steps. The next tick back closes it: all 160 trades that reached the lock closed within a minute, most within a second, and only 2 of 195 reached the $25 target. A normal win is $16 and a normal loss is $40.
- Up to two positions – A second trade can open at a later hour while the first is still running (28 times in 2026). Each is sized from the balance and carries its own stop, so on a bad day both stop out together.
- Friday close – Open trades close at 23:27 server time on Friday. That closed only 11 trades in 2026 but about a third of all trades in 2022-2023, when gold moved fewer dollars a day.
- No grid, no averaging – Nothing is ever added to a losing trade. A “Recovery” option, off by default, keeps the lot at its pre-loss size during a drawdown instead of shrinking it with the balance. It never goes above the earlier size, so it is not a martingale, but each later stop costs a bigger share of a smaller account.
The original name, My AI Gold, points to artificial intelligence. Nothing in the log needs it: fixed dollar levels, an hourly timetable and a half-day trend filter account for every trade we saw. The file is encrypted, so we can see what it does but not what is inside it. That is the same pattern as the “neural network” in Mavrik Scalper EA, which traded a 15-minute timetable, and the opposite of Gold Cortex EA, whose file does carry a neural-network model. Point 6 of our EA scam checklist explains why an AI label is not evidence of anything.
Please test in a demo account first for at least a week. Also, please familiarize yourself and understand how this Gold Trend Lock EA works, then only use it in a real account.
Gold Trend Lock EA Settings
These are the inputs that change how it trades. The chart timeframe is not one of them.
| Input | Default | Notes |
|---|---|---|
| Lot | 0.0 | Above 0 sets a fixed lot. At 0 the lot follows the balance setting below |
| balance / percent | 300 / 0.01 | 0.01 lot for every $300 of balance, growing with the balance. This is a high-risk setting (table below) |
| riskk | 0 | Built-in levels 1-7: 0.01 per $150, $300, $500, $750, $1,000, $2,000 and $2,500 (prop). 0 uses the balance input |
| MaxCountOrders | 2 | Positions at once. 1 avoids the double stop-out |
| Recovery | false | Keeps the pre-loss lot through a drawdown |
| BeginHour / EndHour | 15 / 11 | Entry window in server time |
| TKProfit / RST | 250 / 400 | The $25 target and $40 stop |
| trail / TrailStart / TrailStep | true / 160 / 5 | The +$16 lock and its 5-cent trail |
| Friday / FrHour | true / 0 | Friday close (23:27 server time in our tests) |
| NewsFilter | false | Needs a live news calendar, so it can’t run in the Strategy Tester |
| DelayD / DelayDT | false / 10 | Waits a few seconds before each entry. Off, every copy of the EA enters in the same second |
| AutoGMT | true | Made no difference on our test broker (UTC+2 / UTC+3 server time) |
The download includes six preset files. They change only the lot setting, and the prop preset also allows one position and switches the news filter on. Because every stop is $40 per 0.01 lot, each setting fixes how much of the balance one stop costs:
| Setting | One stop costs | Both positions stopped |
|---|---|---|
| Aggressive preset (0.01 per $100) | 40% | 80% |
| High-risk preset (0.01 per $250) | 16% | 32% |
| Default (0.01 per $300) | 13% | 27% |
| Normal-risk preset (0.01 per $500) | 8% | 16% |
| Low-risk preset (0.01 per $750) | 5% | 11% |
| Conservative preset (0.01 per $1,000) | 4% | 8% |
| riskk 6 (0.01 per $2,000) | 2% | 4% |
| Prop preset (0.01 per $2,500, one position) | 1.6% | – |
Rounding to whole 0.01 lots moves these figures a little on small accounts.
Recommendations for the Gold Trend Lock EA
- Deposit – Size it from the drop, not from a headline deposit. At a fixed 0.01 lot the deepest equity drop was $144 in 2026 and $165 in 2023, so $1,000 to $2,000 per 0.01 lot keeps the deepest drop we measured between 7% and 17% of the account. The lot size optimizer helps when you scale up.
- Lot setting – Use a fixed lot or the conservative preset and below. The default is a high-risk setting.
- Positions – Set MaxCountOrders to 1 on small or prop accounts. In our 2026 test it cut the deepest equity drop by 36% for 8% less profit.
- Leverage – Margin is small (about $0.80 per 0.01 lot at 1:500), so 1:100 is enough.
- Pair and timeframe – XAUUSD only. Any chart timeframe works, H1 is the usual choice.
- VPS – Gold Trend Lock EA should run on a VPS continuously to reach stable results, since it trades on the hour through the night.
- Broker – A low-spread ECN account on UTC+2 / UTC+3 server time, as most gold brokers use. A 250 ms delay in our test cost 3% of the profit and deepened the worst stretch. Compare brokers here.
Performance Review: Live Accounts and Our Own Backtests
1. Is the Gold Trend Lock EA profitable in our 2026 test?
From 2 January to 11 November 2026 at a fixed 0.01 lot on $10,000 (Tickmill real ticks, 1:500, default settings otherwise):
| Metric | Result |
|---|---|
| Net profit | +$1,614.11 |
| Trades | 195 |
| Profit factor | 2.55 |
| Average win / average loss | $16.18 / -$33.33 |
| Max balance / equity drawdown | $97.29 / $143.80 |
| Most losses in a row | 2 |
Every full month made money, from +$26 in April to +$264 in August, and the first eleven days of November lost $23.
The profit factor holds because full stops are rare, not because wins are large. The risk-reward ratio on a typical trade is 2.5 to 1 against, so each stop needs two and a half clean wins to repay. In 2026 the stops never came more than two in a row, so no month collected enough of them to turn red.
2. The live accounts, and which one to trust
The EA runs on three live accounts under its original name, My AI Gold. They take the same trades within a minute of each other, but they are not equally useful.
- IC Markets, $10,000, MT5 – Started on 8 September 2026 and tracked from its very first trade. This is the account we use.
- Weltrade, $300, MT4 – The account with the big growth figure (+1,801% on 12 November). It runs MetaTrader 4, while the file we post is MT5, and the MT4 version was only released on 8 November 2026. Its first 85 trades, from 18 May, were already on the account when live tracking began on 3 September, by which point $300 had become $3,373. Since then it has made +69%.
- Vantage, $300, MT5 – Runs the default lot setting, but only its last 5 of 30 trades were tracked live (from 29 October), too few to judge.
The Weltrade account also shows what a high lot setting feels like: on 5 November two stops cost it $1,290.20 in a day, 20.7% of its balance.
3. Does our backtest match the live account?
We started our test on 17 September, after the IC Markets account’s three larger trades, at the lot setting it has used since then (0.01 per $2,000) and its balance that day.
| 17 September to 11 November 2026 | IC Markets live | Our backtest |
|---|---|---|
| Trades | 28 | 28 |
| Net profit | +$945.25 (+8.9%) | +$1,280.05 (+12.1%) |
| Profit factor | 1.89 | 2.38 |
| Losing trades | 5 | 4 |
| Max balance drawdown | 3.4% | 4.03% |
25 of the account’s 31 entries since 14 September have one of ours within two minutes, and all 25 ended the same way on both sides: the same stop-outs, the same target hit and the same +$16 locks, some a few minutes apart. The gap comes from a few entries the two sides took differently: two live losses at the end of September that our test didn’t take, and a second buy on 5 November that our test took and the account skipped. Two updates, on 24 September and 9 October, changed the entry logic, and the build we tested matches the account on both sides of those dates.
4. The lot setting decides the drawdown
The same 195 trades of 2026, with lots growing with the balance at the settings a buyer is offered:
| Setting and deposit | Final balance | Max equity drawdown | Worst day |
|---|---|---|---|
| Default, 0.01 per $300, $1,000 | $119,606.50 | 40.43% | -26.8% |
| Normal-risk preset, 0.01 per $500, $1,000 | $21,043.71 | 24.63% | -16.1% |
| riskk 6, 0.01 per $2,000, $2,000 | $4,226.04 | 6.90% | -5.0% |
| Prop preset, one position, $100,000 at 1:100 | $178,365.49 | 3.63% | -1.7% |
| Aggressive preset, 0.01 per $100, $300 | reached the 100-lot cap | 89.02% | -80.6% |
The final balances are compounding, not a forecast. 2026 had no losing month at a fixed lot, and raising the lot after every win turns that into five and six figures on paper. The drawdown column is the part to plan around. The worst days were the two dates when both positions stopped together, 31 July and 5 November.
The Recovery switch on the normal-risk preset finished 10% higher with the same 24.63% equity drop: with at most two losses in a row there was little to recover from. The prop preset stayed inside a 5% daily and 10% total loss limit all year. Its news filter can’t run in the tester, so that part is untested.
5. Earlier years
Real ticks on our test broker start in January 2026, so earlier years ran on generated ticks. On 2026 the two models gave the same 195 trades and generated ticks showed 9% more profit, so read the older rows as slightly optimistic. Fixed 0.01 lot on $10,000:
| Year | Trades | Net profit | Profit factor | Most losses in a row | Max equity drawdown | Losing months |
|---|---|---|---|---|---|---|
| 2022 | 198 | +$1,134.42 | 2.17 | 3 | $108.35 | 1 |
| 2023 | 173 | +$696.38 | 1.74 | 5 | $164.66 | 2 |
| 2024 | 223 | +$1,956.14 | 3.61 | 3 | $108.00 | 0 |
| 2025 | 261 | +$1,832.22 | 2.20 | 3 | $151.24 | 0 |
| 2026 (real ticks) | 195 | +$1,614.11 | 2.55 | 2 | $143.80 | 0 |
Every year made money at a fixed lot, and 2023 was the weak one. With gold under $2,100 a $40 move took longer, trades lasted a day or more and the Friday close did a third of the exiting. The EA’s settings were last re-optimized in September 2026 over a period we can’t see, so the older years may be part of what it was tuned on.
Gold Trend Lock EA
An MT5 XAUUSD robot that opens hourly trades with the half-day trend, risks a fixed $40 per 0.01 lot and locks wins at +$16. Our 2026 real-tick test made +$1,614 per 0.01 lot with a 2.55 profit factor.
Pros
- A $40 stop loss on every trade, no grid and no averaging
- Our real-tick test matched 25 of 31 live entries to within two minutes
- Positive at a fixed lot in every year from 2022 to 2026
- Works on any chart timeframe, and the prop preset stayed under 2% per day in 2026
Cons
- One full stop costs two and a half normal wins
- The default lot risks 13% per stop, 27% when both positions stop together
- The headline live account is mostly history from before live tracking, on MT4
- The news filter can't be tested, and the "AI" shows in nothing it does
Summary
Our backtest takes the same trades as its $10,000 live account, and its fixed-lot results were positive in every year we tested. However, wins are locked at $16 against $40 stops, nothing in its behaviour needs the AI in its name, and the default lot risks about 13% of the balance per stop. Run it at 0.01 per $1,000 or less.
Built by ForexCrackedWe tested 1,200+ indicators. Then we built our own.Six years of reviewing EAs and indicators went into FXC Fusion: two engines, the seven-panel cockpit, MT4 and MT5, from the team behind this site.
Your download will start automatically. If it doesn’t, click here.Right now a download can take up to a minute to start. Please wait a moment, or use click here.
How it compares to other “AI” gold robots we have tested
| Gold Trend Lock EA | Mavrik Scalper EA | Gold Cortex EA | |
|---|---|---|---|
| What the log shows | Hourly entries with the 12-hour trend | Entries on a 15-minute timetable, one direction a day | A neural-network model inside the file, replaying a stored sequence in the tester |
| Stop loss on every trade | Yes, $40 | Yes, $38 | No (most positions had no server-side stop) |
| Typical win vs loss | $16 vs a $40 stop | About $1 vs a $38 stop | $3.99 vs -$17.07 (averages) |
| Positions at once | Up to 2 | Up to 4 | Up to 5 |
| Profit factor in our test | 2.55 (2026) | 1.93 (Apr-Sep 2026) | 1.76 |
The three tests cover different windows, so read the numbers as a rough guide. Of the three, this one has the least lopsided payoff and is the only one we tested across several years.
Frequently Asked Questions (FAQ)
Conclusion
Gold Trend Lock EA is a plain fixed-stop gold robot, and our backtest takes the same trades as its cleanest live account. Its fixed-lot results held up in every year we tested, 2022 to 2026. However, the risk sits in the lot setting it ships with and in the second position that doubles the bad days, not in the strategy. Run the Gold Trend Lock EA free download at 0.01 lot per $1,000 or less, with one position on smaller accounts, and judge it over months.
Discussion no comments yet
Is this still working?
Nothing here yet
No one has asked anything about Gold Trend Lock EA yet. First question gets answered.