SmartChoise EA Review (2026): An M1 Gold EA for MT5 With No Track Record We Could Verify

25.5 / 100

weighted verdict score on eight published criteria

0

live records, signals or tester reports we could open and check on 29 Oct 2026

v9.7

version shown on our own store listing in an earlier site render; not matched to a vendor changelog

The SmartChoise verdict at a glance: a product name, a version number and no checkable proof

Most verdicts on this site weigh evidence against evidence: a signal against a review, a backtest against a live account. This one cannot. On 29 October 2026 we searched the MQL5 site for SmartChoise and the search page returned no readable product result; we could not identify the product ID from the seller side; and our own store page could not be fetched. The only product facts we hold come from our own listing as it appears in an earlier render of this site: the name, the platform (MT5), the market (XAUUSD), the timeframe (M1), the version (9.7) and the phrase 'neural network-powered'. Every one of those is a sell-side description and none is a measurement.

So the short answers are these. Is SmartChoise legit? We do not allege wrongdoing, and we found nothing either way; it is unverified. Does it work? We could not see any trade history, so we cannot say. How risky is it? Unknown, which is why the label is High: a gold system on one-minute bars with an undisclosed exit and sizing design has to be treated as able to lose a large share of an account until someone shows otherwise. What does it cost? We could not confirm the Market price or our own current price on the day we checked. What settings and broker? Our listing names XAUUSD on M1 on MT5; anything beyond that, such as the minimum deposit, the risk input or the broker type, was not in a source we could read.

What SmartChoise is, as far as we could confirm

SmartChoise is presented as an Expert Advisor, a program that runs inside MetaTrader 5 and places trades by itself. Our listing places it on XAUUSD, the symbol most brokers use for spot gold against the US dollar, and on the M1 chart, which means its logic reads one-minute bars. The word 'neural network' in the title suggests that some part of the entry decision comes from a trained model rather than from fixed indicator rules. We could not read how that model is built, what it was trained on, how often it is retrained, or whether it decides entries, exits, sizing or all three.

How we researched SmartChoise, and who has something to sell

Our method is the same for every verdict: open the vendor listing, its reviews and changelog, the seller profile, every linked signal and any reputable third-party review; grade each source from A to D; score eight criteria against written anchors; and say plainly what could not be checked. For SmartChoise the first step failed. The MQL5 site search we opened returned no readable result for the name, we did not find the product ID through other routes in the time available, and so the listing, reviews, comments, changelog and seller profile could not be read. General web searches were not usable on the day. We did not fill the gap with numbers from memory or from unnamed websites.

• The developer of the MQL5 product sells the EA and would usually publish the listing, the screenshots and any signals. We could not read any of it, so the developer's own case is missing from this page, which is a gap we want to close; see the right of reply below.
• MQL5 and MetaQuotes host the Market, its reviews and its signals. Platform-built signal statistics are the strongest evidence a vendor can show. None were available to us.
• Buyers who wrote reviews would provide testimony. We read none, so we cannot report a rating or a range of experiences.
• ForexCopilotEA sells a listing under this name. We gain if you buy it. That is why the disclosure leads, why our own listing is graded D, and why its title is not treated as proof of anything.

SmartChoise track record: what a signal would need to show, and what we found

The question most buyers ask first is whether SmartChoise has a live track record. We could not find one we could open. That is different from saying none exists: a signal may be linked from the Market page we could not read. The table below is therefore a template rather than a result. It lists the figures we read from every live signal in other verdicts, and records for each one that the SmartChoise value is unknown. If you find a signal for SmartChoise, these are the rows to fill in before you decide anything.

Why insist on all of these rather than a growth figure? Because a single number can be made to look good by design. A system that never closes a loss shows a beautiful balance curve until the day it does. A system that risks 10% a trade shows huge growth until the streak arrives. A system that was tuned to the last six months of gold shows a great backtest and nothing about next year. The rows above are chosen so that none of those patterns can hide. Our guide on how to verify an EA track record explains each one, and backtest versus live results explains why a tester report alone is not enough.

How an M1 gold EA makes and loses money

Without SmartChoise's own logic to read, the most useful thing we can do is explain how the type of system our listing describes generally behaves, so that a buyer knows which questions matter. A one-minute gold EA looks for small, short-lived moves: a burst through a level, a quick reversal after a spike, or a pattern a model has learned to associate with the next few minutes of price. The holding time is usually minutes, the target is usually a small number of dollars per ounce, and the number of trades is high. That has three consequences.

• Costs are a large share of the edge. If the average trade aims to make a few dollars per lot and the spread plus commission costs a meaningful fraction of that, a small change in broker conditions can turn a profitable backtest into a losing live account. Our scalping broker rules guide covers spreads, minimum holding times and why some brokers restrict very short trades.
• Execution decides fills. At one-minute horizons, the difference between a fill in tens of milliseconds and a fill in hundreds can be the whole profit. A VPS close to the broker's server is normally part of the cost of running such an EA; see our MT5 VPS setup guide.
• The tester can flatter. Many one-minute strategies look much better in a test on modelled ticks than on real ticks, because modelled ticks do not reproduce the spread spikes and gaps that real gold trading has. Our Strategy Tester settings guide shows which mode to use.

A 'neural network' adds a fourth question: what the model learned from, and whether it still applies. A model trained on one period of gold can fit the noise of that period very well and then meet a different market. Developers manage this by holding out data the model never saw, retraining on a schedule and publishing out-of-sample results. We could not see whether SmartChoise does any of that. The word in the title does not tell you; the out-of-sample record would. Our guide to optimising an EA without overfitting explains why the same caution applies to any trained system.

Where does the money go when such a system loses? Usually in one of two ways. In a system with a stop on every trade, losses come as streaks: many small stops in a row during a market the model does not read well. In a system that holds losing trades or adds to them, losses come as rare, large events: weeks of small wins followed by one session that takes back months. Both can be traded for a long time; only one can empty an account in a day. Knowing which kind SmartChoise is would be the single most important fact for a buyer, and it is the fact we could not establish.

Use the calculator above with numbers from any SmartChoise signal you find, or from your own demo. The rule is simple: win rate times average win must be larger than loss rate times average loss, after costs. A system with a 90% win rate and an average loss ten times its average win is breakeven before costs and losing after them. A system with a 40% win rate and an average win twice its average loss is profitable. The win rate alone tells you almost nothing.

SmartChoise risk: why the label is High without a verified design

Our risk labels follow the design and the evidence. A product with a stop on every trade, no grid or martingale and a verified modest drawdown can earn Medium or better. A product that uses a grid or martingale without a hard stop cannot be rated better than High. SmartChoise is in neither group, because we could not read what it does. We label it High for three reasons, and we will revise it as soon as the design is documented in a source we can check.

• Unknown trade management. We could not confirm whether every trade carries a stop loss, whether positions are averaged or added to, or whether lot size grows after a loss. Any of those would change the worst case dramatically. Our guides on grid trading and martingale risks show what those designs look like in a signal.
• Gold at one-minute speed. XAUUSD can move several dollars in seconds around US data releases. A short-horizon system without a news filter or a hard stop is exposed to exactly those moments; see gold EA news trading.
• No measured drawdown. Without a verified maximum drawdown and longest losing streak, there is no evidence-based way to size a live account. A label of Medium would imply a confidence we do not have.

The recovery calculator shows why sizing matters more than the strategy for most buyers. A 20% loss needs a 25% gain to recover; a 50% loss needs 100%. On a short-horizon gold system, where the edge per trade is small, a deep drawdown can take a very long time to recover even if the edge is real. Until SmartChoise's drawdown history is public, the safe planning assumption is that a large drawdown is possible.

SmartChoise reviews: what buyers say, and why we cannot report it yet

We could not read the review section of the MQL5 listing, so we cannot report a star rating, a number of reviews, or what buyers praise or complain about. We also did not find a third-party review we could treat as independent. We do not quote ratings that appear on unnamed sites, on social media posts or in sales pages, because they cannot be traced to a verified buyer.

When you read SmartChoise reviews yourself, three filters help. First, look for the version: a review from version 7 says little about 9.7. Second, look for numbers: a reviewer who gives a broker, a risk setting, a trade count and a drawdown is worth ten who say 'great EA'. Third, look for the developer's replies to negative reviews, which show how problems are handled. Remember that reviews on any product skew toward buyers who are ahead, and that a buyer who lost money is less likely to post.

SmartChoise price: what we could confirm about the cost

We could not confirm the price of SmartChoise on the MQL5 Market on 29 October 2026, because we could not read the listing. We were told that an MT4 version is priced at around $999; we could not open a source for that figure and report it only as unconfirmed. We also could not load our own store page on the day, so we do not quote its current price here. An internal audit of our site kept in this repository records that our catalogue was at one time sold at a single flat price; we did not confirm that it applies today, and you should read the price on our listing at the time you buy.

Price is scored at 4 out of 10 for a reason that has nothing to do with the number: no price can be judged good value without a record. A low price for an EA without proof is still a cost if the EA loses the account, and a high price for an EA with years of verified history can be cheap. The other costs are predictable for any M1 gold system: a VPS, a low-spread account and the time to run a proper demo. Our guide on how much money you need to run an EA lays them out. If you buy from any seller other than the MQL5 Market, including us, read the terms on updates, activations, licensing and refunds before paying, and confirm that the version and support match what you expect.

SmartChoise settings and broker: what is known and what to ask

Our listing gives two settings facts: the symbol is XAUUSD and the chart is M1. Everything else a buyer needs, such as the risk input, the lot method, the maximum number of open trades, a news filter, trading hours and the minimum deposit, was not in a source we could read. That is not unusual for a product description, but it means the settings are unknown until you open the EA's inputs yourself or the developer documents them.

1. 1

   Step 1: Match the symbol name

   Brokers name gold differently (XAUUSD, GOLD, XAUUSD.m and so on). Check that the EA accepts your broker's symbol and its contract size, or it may size trades wrongly. Our XAUUSD EA settings guide covers contract sizes and digits.

2. 2

   Step 2: Check the spread at the hours it trades

   On M1, the spread at the moment of entry matters more than the average. Watch your broker's gold spread during the sessions the EA trades and during news.

3. 3

   Step 3: Find the risk input and set it low

   Look for a fixed lot, a percentage risk or an 'auto lot' input. Start at the smallest setting and raise it only after a demo record justifies it. Our lot sizing guide explains the methods.

4. 4

   Step 4: Look for stops and position limits

   Confirm in the inputs and on a demo that every trade gets a stop loss and that the number of open positions is capped. If not, treat the EA as a grid.

5. 5

   Step 5: Run it on a VPS

   One-minute systems need a stable connection and low latency. Follow the MT5 VPS setup guide.

6. 6

   Step 6: Test on real ticks first

   Run the Strategy Tester in 'Every tick based on real ticks' mode over several years and compare it with a demo run. Our guide on reading an MT5 backtest report shows what to look for.

The SmartChoise score, criterion by criterion: 25.5 out of 100

The total is 25.5 out of 100. Evidence and risk carry 40 of the 100 points, and SmartChoise earns 9 of them: evidence at 1.5 because the only source is our own listing, graded D, and risk at 3 because the design is undisclosed in anything we could read. Profitability, at 2, reflects the absence of any measured edge. Fit and cost are the strongest criteria at 4, because the market, timeframe and platform are at least stated. The score does not say SmartChoise loses money. It says that, on 29 October 2026, a buyer had nothing independent to rely on, and the method does not give points for claims.

What would move the score? A live MQL5 signal or Myfxbook record of several months on the current version would lift evidence to grade A and could add several points on its own. A documented stop-loss policy with a verified drawdown could move risk toward the middle of the scale and the label toward Medium. A readable listing with a price, a changelog, a free demo and a developer who answers would lift trust and support. Each of those is in the developer's hands, and we will re-score when any appears.

Who SmartChoise fits, and who it does not

• It may fit an experienced MT5 user who wants to evaluate a short-horizon gold EA on a demo account, has a low-spread account and a VPS, and is willing to build their own record before trusting it with money.
• It does not fit anyone who needs evidence before buying, who cannot run a multi-week demo, or who would size a live account from the listing rather than from measured drawdown.
• It is risky for prop-firm challenges. Daily loss limits leave little room for an unknown worst case, and some firms restrict very short holding times; see EAs for FTMO-style challenges.
• It is a poor choice for a first EA. Start with a product that publishes a verified record so you can learn what normal behaviour looks like; our first 30 days with an EA guide explains the routine.

How SmartChoise compares with other gold EAs we have scored

The gap between SmartChoise and the better-documented gold EAs we have reviewed is the evidence, not necessarily the strategy. Gold House, a gold breakout EA, links two platform-built live signals with hundreds of trades each, which gives a buyer something to check. Smart Gold Hunter and Gold Snap were scored on the same eight criteria with the same rules. SmartChoise's 25.5 sits well below them because nothing behind it could be opened, not because of a measured loss. For the broader picture, see our best gold EA for MT5 guide, the strategy comparison of grid, martingale, scalper and trend designs, and our gold EA category.

For a fuller version of this list, see our EA buying checklist and the guide to whether forex EAs really work.

Questions to put to the SmartChoise developer before you pay

1. What is the MQL5 product ID for the MT5 version, and is there a separate MT4 product?
2. Is there a live MQL5 signal or Myfxbook record for version 9.7, and which broker and settings does it use?
3. Does every trade carry a stop loss? Does the EA ever average, add to a losing position or increase lot size after a loss?
4. What does the neural network decide, what data was it trained on, how often is it retrained, and are out-of-sample results published?
5. What are the recommended minimum deposit, risk setting, broker type and maximum spread?
6. What was the largest drawdown and longest losing streak in testing and in live trading, and on which version?
7. Is there a news filter, and does the EA trade around major US data releases?
8. Which sellers are authorised to sell SmartChoise, and what updates and support does each purchase include?