Lizard EA Review (2026): A Six-Strategy Gold Breakout EA With 25 Releases in 21 Weeks and No Live Record Left to Check

36.0 / 100

weighted verdict score on eight published criteria

0 signals

seller profile on 29 Oct 2026: the cited signal 2372821 returns 'not found'

25+ versions

released between 14 May and 4 Oct 2026, from 1.20 to 2.1

The Lizard EA verdict at a glance: an active product with no record to check

Lizard earns points for being a transparent kind of machine. A breakout EA that waits for a swing level to be crossed, with a defined stop loss and take profit on each trade, is a design a trader can reason about, and the developer answers technical questions in public. It loses points on nearly everything that can be tested. There is no verified signal, no Myfxbook record, no performance figure in the description, and a changelog that shows the strategy set, the defaults and the news logic being rebuilt every week. A buyer in October 2026 is therefore buying version 2.1 on the strength of reviews, most of which describe weeks, not seasons. The scorecard and radar show how the total was built.

What Lizard EA actually is

Lizard is an Expert Advisor for MetaTrader 5, sold on the MQL5 Market by Marco Scherer of Switzerland, trading under the name Masthetics. His seller profile shows a 4.2 rating from 91 votes, three products, zero signals, zero subscribers, 2,687 demo versions across the range and an experience label of 'no experience'. Lizard was published on 9 May 2026, showed version 2.1 and was last updated on 4 October 2026. The page lists 15 activations (a limit set by the seller, not a count of users), 6,621 demo downloads, a rating of 4.07, 71 reviews and 323 comments, and reports 34 purchases per month. Support runs through a moderator called Zolia, whom the seller described on 20 July 2026 as based in Taiwan, with responsibilities divided inside the team. A regular and well-informed defender in the comments, Mercier Guillaume Patrick, answers many questions; the page does not say how he relates to the seller, so we treat his posts as those of an interested community member.

The strategy statement is specific. The EA identifies key swing highs and swing lows on the H1 chart and places pending buy stop and sell stop orders at calculated entry points. An order is triggered only by an actual breakout; simply touching the price is not enough. Six independent strategies run at once on H1, each with its own stop loss and take profit, its own trailing system, its own magic number and its own risk weighting, from short intraday breakouts to multi-day swing setups. Strategy activation depends on the balance: up to 1,000 USD, zone A is active, and from 1,001 USD zones A and B both trade. The exit system combines break-even, trailing stop loss, trailing take profit and a step-based trail. A virtual stop loss closes positions in software, independently of the broker. Per-strategy time filters block hours where the strategy's historical profit factor was below 1.0, a news filter pauses around US Non-Farm Payrolls, pending orders are cleaned up before the daily close, and an input caps the maximum daily loss in 'prop firm' style.

How we judged it, and who has something to sell

Evidence on EAs is never neutral. Here is every party near this product and what each stands to gain, ours included.

• The developer sells Lizard (and two other products), runs the Telegram and Discord communities, and wrote on 5 October 2026 that he is no longer actively pushing sales. His profile also carries a promotion for a third-party offer. His comments are testimony from the party who earns from the sale.
• MQL5 and MetaQuotes host the Market, the reviews, the comments and the signals. The 'not found' page for the cited signal is the platform's own statement.
• Support and community helpers answer most technical questions. One of them, Mercier Guillaume Patrick, also quoted backtests and results; his relationship to the seller is not stated, so we report his posts as community testimony.
• Buyers who wrote reviews split in two. Some say they are profitable; others report losses. We could read 20 of 71. All are testimony, and some reviewers have their own products to sell.
• ForexCopilotEA sells a listing for this EA. We gain if you buy it, which is why this verdict opens with that fact and applies the same rules as every other.

Does Lizard have a live signal? The one that existed is deleted

A live record is the only thing that separates an EA that works from one that merely tests well. The Lizard page does not link one, and the seller's profile shows zero signals. In the comments the seller and his helpers referred to one repeatedly, and several of the early sales arguments were made with it: on 20 May 2026 the seller wrote that the signal 'has been running on default settings since v1.3'; on 31 July 2026 he quoted its monthly returns; and on 10 September 2026 he answered a one-star review by linking signal 2372821 under the name 'Signal Normal'. A 'High Risk' signal was also mentioned on 12 August 2026 without an identifier.

On 29 October 2026 the platform returned 'trading signal not found' for 2372821. On 5 October 2026 a user asked whether it had stopped and the seller replied: 'Yeah, I stopped the signal a while ago because I'm no longer actively pushing the EA sales, and some users always seem to know everything better anyway. So I'd rather save myself the effort of running a separate terminal just for the signal.' A reviewer wrote on 2 October 2026: 'One of the main reasons I bought it was the seller's live signal, and that signal has now been deleted.' Stopping a signal is the seller's right and costs a seller a terminal and some effort. It also removes the only independent test that existed for this product, at the moment the version changed again.

Three further points about the signal history are worth knowing, because they are the reasons we do not treat the quoted monthly figures as even soft evidence. First, the seller named the signal broker as TMGM Edge (RAW) on 11 August 2026, while a community helper named VTMarket on 23 May 2026 and wrote on 5 June 2026 that the symbol suffix changed from XAUUSD.ECN to XAUUSD.STD because the broker had switched the account type 'as part of the partnership'. A user had alleged that trades were being copied from two accounts; the answer did not establish what was run. Second, the signal ran on the product's default settings, but the defaults changed through the versions (the 1.75 notes list a default risk of 5%, a stop of 3000 points and a high trail mode, while a helper wrote on 29 August 2026 that the default was risk 1%). Third, the quoted July figure of 0.13% was blamed on 'summer seasonality', which is possible and also not testable now.

How a swing breakout EA makes and loses money

A swing breakout system bets that when gold closes beyond a recent high or low, the move continues long enough to pay for the ones that fail. It wins in clean trends and loses to false breaks: price pokes through a level, fills the stop order, then reverses to the stop loss. The payoff profile is the opposite of a scalper's. Expect a modest win rate, a stop that is large relative to many individual wins, and a profit that depends on a few big runners. The edge is real in many markets and also unstable, because it needs trending behaviour and decent execution.

Lizard's own filters tell you where the fragility sits. A 'fakeout' filter, with settings from off to high that changed meaning through versions 1.30, 1.50, 1.76, 1.80 and 1.93, exists because a stop-order system fills on spikes. A helper wrote on 21 July 2026 that being 'stopped out immediately' is usually the fake-breakout filter working as intended, and a buyer wrote on 22 July 2026 that since the latest update he had 'only losses with fakeout low filter'. The time filter blocks hours whose historical profit factor was below 1.0, which is a classic optimisation step: it improves the backtest by construction and has to survive out-of-sample hours to be real. The version 2.1 note says the new time profile 'opens more trading hours based on fresh data analysis', which is a re-fit. For how to tell a robust parameter from an overfitted one, see our guide to optimising an EA without overfitting.

Gold adds its own difficulty. Spreads widen at the US data releases that move gold most, stop orders slip, and a trailing stop that sits close to price can be flicked out by a spike. Lizard's changelog shows the developer fighting exactly this: a virtual trailing fallback in 1.86 when the broker's freeze or stop level blocks a modification, a hard stop that reads the broker freeze level in 1.92, and a three-way virtual stop switch in 2.0. Those are sensible responses to real problems. They also mean the exit logic a buyer runs today is not the exit logic that produced any earlier result.

Lizard version history: 25 releases in 21 weeks

The update list is the richest evidence we have, and it cuts both ways. A vendor who fixes margin errors for 1:500 leverage (1.21), a day-switch bug (1.67), a per-EA daily drawdown bug (1.72) and a case where 'zones could place too many pending orders' (1.71) is looking after the product. A strategy whose basic structure moves every few weeks is hard to judge, and impossible to match with a record. Here is the sequence that matters.

Two statements in the comments sit awkwardly beside this table. On 23 July 2026 a helper wrote: 'We have not changed the core of Lizard EA since v1... add-ons.' The changelog records zones removed, zones added, defaults changed, presets introduced and presets deleted. On 6 October 2026 a buyer asked whether the presets had disappeared and was told to 'keep default and change SL as based on the preset before', which means the reader is expected to rebuild an earlier preset by hand. A buyer wrote on 9 October 2026 that version 1.85 was generating a profit on his account while version 2.1 produced a loss. One comparison proves little, but it is the version-drift problem in a sentence: reviews of one release do not describe the next.

Lizard EA reviews: 4.07 stars, and what the readable ones say

The page headline is 4.07, from a rating count that the page data puts at 60 while the tab says 71 reviews. We could read 20 of them, because the pagination for the rest returned an error. Those 20 average 3.95 stars: twelve gave five, two gave four, two gave two and four gave one. That is a polarised distribution, and the split follows a pattern. Every low review cites live losses, early stop-outs, a failed news filter or the disappearance of the signal and the support group. The favourable ones typically cover weeks.

• Positive, with detail. A reviewer on 30 September 2026 had used it for about two months, one on demo and one live, and found both profitable. Another on 29 August 2026 reported about 5% in the first month on a funded account with strict risk. A reviewer on 22 August 2026 lost about 60 USD in July, cut risk and reported +75% in August 'with own settings'. A reviewer on 14 August 2026 wrote of 'Realistic, Solid EA with Proper Risk Control' while seeing '12 days of drawdowns'. A reviewer on 4 September 2026 said version 1.92 'has changed a lot; EA has become much more profitable'.
• Positive, without detail. 'Authentic and reliable products', 'Too Good :)', 'It's going ok', and a five-star rating with no comment from a seller of another EA who is active in the comments.
• Negative on losses. 'I lost 30% of my account in just one day' (5 October 2026). 'I have been using it for 2 months and no profit at all' (6 October 2026). On 31 July 2026 a buyer reported losing 486 USD twice at 0.16 lot and, on version 1.86, losses where the news filter 'didn't work under default settings'. He added that 'multiple users lost 25%+ of their accounts under DEFAULT settings during CPI news'. Another on 12 August 2026 wrote that the 'entry are good but it hit SL to early'.
• Negative on live against test. A German review of 10 September 2026 said tests and demo were great and live results 'very mixed'. A buyer in the comments on 24 July 2026 wrote that in the backtest 'the robot earns really well. However, the reality is quite different'.
• Negative on the signal and support. The 2 October 2026 review (four stars, but a negative text that still contains unfilled template placeholders such as '[version]' and '[X% / $X]') calls the results 'very disappointing', says 'many trades got stopped out quickly after a breakout entry' and asks buyers to 'demand a verifiable live signal first'.

The developer replies to the critical reviews, and the tone is part of the evidence. To the 30%-in-a-day review he replied that the reviewer was 'the only one in the whole community who manages to end up with losses like that. Maybe you messed around with the settings too much?' To another he wrote that the reviewer should 'stay far away from my community'. To a third he wrote, on 14 August 2026, that he was not 'your psychologist for your mental problems'. A helper added that 30% in a day 'is not possible, only if you trade with 100$ balance OR Risk 30% per trade'. The seller's reply may be right on the facts; the point for a buyer is that disputes are treated as flaming, not investigated, and that support for the people who report losses is thin. The developer's single thank-you message to a batch of five-star reviews within two minutes on 10 September 2026 reads as a routine, not as a conversation.

Be careful with our own sample. We read 20 of 71. The missing 51 may be warmer or colder than these. What we can say is that the verifiable reports of loss are specific (amounts, dates, versions), and the verifiable reports of profit are general, with a few exceptions. For how backtest and live results diverge, see our guide to backtest versus live results.

The CPI losses and the news filter, as reported

The most serious complaint is specific and dated, so we set it out as an allegation with the vendor's reply. On 12 August 2026 a buyer, Jung Hoon, wrote in the comments under the heading 'CRITICAL: News Filter Failure caused $2,400 Loss (CPI release)'. He said the news filter was on with 30 minutes before and after, yet the EA opened a buy trade exactly at the CPI release time, with about 800 USD lost across three accounts. A second user on TMGM wrote that he had a full stop loss at CPI with the filter on, and a third wrote that four trades all closed at their stops. A community helper wrote 'I also get the lost from CPI News'. The reviewer later wrote that the developer 'pushed v1.87 overnight to fix this' and 'even called victims liars'. The developer's reply was that critics are 'either fake users or just waiting for the EA to make a single loss', and later, on 14 August 2026, a denial.

The facts we can confirm from the public page are narrower. Version 1.87 was released on 12 August 2026 with the note 'Fixed news calendar filter'. On 29 August 2026 the seller's reply relayed in the comments said there was no fallback to the MQL5 calendar, that in a backtest 'you trade through all news anyway', that only NFP is blocked because he hard-coded the NFP dates, and that 'my own preference would be to remove the news filter completely'. Version 1.90 then moved the feed to ForexFactory, version 1.92 set the NFP filter to off by default, and the description today still lists only the US NFP pause. The reviewer of 2 October 2026 said the news filter stays off until the WebRequest address is added. Whether any buyer lost money because of a defective filter we cannot verify; that the filter has been a moving target is documented.

Lizard drawdown and risk: no verified figure, wide stops, large defaults

There is no verified maximum drawdown for Lizard, because there is no verified record. We therefore score the risk on the design and on the reported behaviour, and we do not infer a number. The design carries a stop loss on every trade, which in the documented versions is large. The 1.20 notes cap the 'Safe Exit' module at 50 USD at 0.01 lot; version 1.72 set the manual stop default at 5000 points, which the notes equate with 50 USD at 0.01 lot; version 1.75 set the default stop at 3000 points; version 1.86 raised the High preset's stop from 500 to 1000 points. On a gold contract, a stop in the thousands of points is a wide stop. A first-day buyer on 12 May 2026 described one strategy with 'TP 20$ SL 38$' and another with 'TP 10$ SL 200$' and called it 'big risk 1:20'.

• Risk per trade is configurable and the defaults moved. The 1.75 note lists a default risk of 5%; a helper wrote on 29 August 2026 that the default is 1%. A 1,000 USD account was reported trading 0.01 lot. If you do not know which default your build uses, you do not know your risk per trade. Check it on demo before anything else.
• Several trades can fire together. Six strategies, and up to eight zones in the 8-zone basket the comments describe, share a risk budget, but the same breakout can trigger several at once. The 1.51 'Risk Scope' option chooses per zone or per basket; the 1.71 fix concerned too many pending orders.
• Reports of large single-day losses exist. The 30% in a day, the 2,400 USD across accounts at CPI and the '25%+ under default settings' are reports by buyers. They are not verified, and the developer and a helper dispute them. They are also specific enough to be taken as a warning.
• The virtual stop depends on your terminal. The listing itself says the virtual stop 'requires the terminal or VPS to remain operational and cannot provide the same protection as a server side stop when the software is disconnected'. Version 1.92 added a hard stop alongside it. Check which is active.
• Prop-firm use is possible and constrained. There is a maximum daily loss input and a trailing 'jitter' for exit diversity, but a news failure or a stop-out cluster inside a firm's daily limit is not prevented by either.

A comment from the seller on 29 August 2026 deserves to be read as a design principle, not a joke: 'If you don't want to see any stop losses, then you should probably move on to a Grid EA, the only difference is that the stop loss there is usually your margin call.' That is a fair statement of the trade-off. A breakout EA with hard stops will take losses, in runs, by design, and many of the complaints are about the runs: a helper wrote that a critic left a one-star review after '3 stop losses', and the seller wrote that many one-star reviews 'came after the EA hit 12 stop losses'. Twelve stops in a row at a wide stop size can cost 10% or more of a small account, which is why the risk setting matters more than any filter.

Lizard EA price: 99 USD now, and a history we could not verify

The Market price on 29 October 2026 was 99 USD, the same on the German page, with a free demo and a rental option. We could not see the rental price in the page text. The price history exists only in comments and we could not confirm any of it. On 10 August 2026 a reviewer wrote that he paid 999 one day after the price was 199 and that 'now its 699'; on 28 August 2026 another wrote that he 'couldn't find anyone to buy at $699'; on 29 August 2026 a third referred to 'the 70% price cut this week'. If those figures are right, the price went from 199 to 999 to 699 and then to 99 within about eleven weeks. A critic argued on 13 August 2026 that 'the price goes up until the EA disappears'. The seller denies the allegations, and with the signal gone and sales no longer pushed, the price at least is now lower than the early claims.

At 99 USD the question is not whether the price is high but whether it buys anything you can check. It buys a version-2.1 licence, a demo, a manual and a set file that are mostly handed out privately, and a community that has moved between platforms. It does not buy a track record, because none exists. Count the other costs as well: a Windows VPS (which the listing says to strongly consider), an ECN or RAW account, and the time to test the presets. A rental, if you want to test live before paying in full, is the cheaper way to find out.

The gap in the last row matters. Our description copies the strategy text and has no performance figures, no signal link and no mention of the WebRequest setting that later versions need. Because our listing is version 1.85, it is the build that predates the news fix and the freeze-level work. If you are weighing it against the Market product, weigh the missing releases and the missing manual as costs, and remember that the Market seller's own demo lets you test version 2.1 first.

Support, the manual and the community

Documentation exists but arrives by hand. The manual was posted in the comments on 17 September 2026 as a zip file called Lizard_EA_Manual__EN-CN, and a signal set file (signal_setfile.set) on 9 May 2026. Many replies in the comments are 'I will come to you in private' (for set files, manuals, community access). Official support is Zolia, and many threads are answered by Mercier Guillaume Patrick. That works for a buyer who is willing to message, and not for one who wants to read the instructions before paying.

The community has moved. The seller's profile says that on 30 June 2026 the Discord was closed to verified buyers only because of 'too many scam attempts and cases of identity theft', with a public Telegram group as an alternative. On 23 September 2026 a comment said the Telegram group was reporting 'this user does not exist', a new Discord link was posted on 24 September, and on 28 and 30 September users asked what happened to the Discord and were pointed to the profile. A one-star reviewer on 6 October 2026 wrote that the developer 'had deleted EA and support group and disappeared'; the developer replied that a new group had been opened. We cannot reconcile those accounts. What is verifiable is that a buyer has to follow the seller's channels as they change.

Finally, a gap that affects a version-2.1 buyer directly: on 15 October 2026 a user asked, 'Where can I find the function that select only buy of sell? Did author remove this function?', and the question had no answer in the page text. Parameter lists that change with each release, and a manual dated 17 September that predates 2.1, mean the documentation is likely to lag the product.

Allegations, replies and what we make of them

The Lizard comment section contains a long argument, mostly between one critic, Ryan Woodward, and the seller and his helpers. We record it as allegation and reply, because a verdict that ignored it would be incomplete, and one that repeated it as fact would be unfair. Nothing here is proven.

• Allegation: the price rises until the EA disappears. The critic wrote on 13 August 2026 that 'the price goes up until the EA disappears'. Reply: the seller denies it. What we can observe: user-reported prices fell from 999 and 699 to 99, and the seller says he is no longer pushing sales.
• Allegation: backtests were edited after a large loss. On 12 August 2026 the critic's review said 'there's now an update with edited backtest results' and that the big loss 'doesn't show up on signal either'. Reply, the same day: 'The backtests are not edited. The only thing you can change is the Fakeout settings... If you test with 1ms latency, you'll never get realistic results.' The critic also argued that each new version hides earlier large losses and asked for past versions to be posted. What we can observe: defaults, presets and filters changed in nearly every release and the older builds are not published. We cannot say whether any change was made to flatter a test.
• Allegation: fake accounts and a purchase-for-review scheme. The critic claimed users were commenting from the same IP address and that the EA is 'a free EA if you post a 5 star review'. The seller and a helper answered that critics are fake or hoping for one loss, and a helper alleged on 5 August 2026 that some one-star reviewers demand a crypto refund. No evidence for either side was posted. We note only that the developer's thank-you reply went to a batch of five-star reviews within two minutes, and that some five-star reviews give no detail.
• Allegation: zero new activations since an update. Rebutted in the comments by a user who explained that the activation figure is a limit set by the seller, not usage. We discarded it.

Our reading is deliberately modest. The Lizard page shows a product that was changed often and defended hard, with a signal that existed and has gone, and buyers on both sides. It does not show fraud, and we do not claim it. It shows the absence of the one thing that would settle the argument. A seller who wants disputes to end can republish a signal on a verified broker account and keep it running for twelve months.

Setting Lizard up on a broker: what the listing asks for

The listing asks for MetaTrader 5, a hedging account, ECN or RAW pricing, leverage of 1:100 or more, a minimum balance of 250 USD and a VPS. The EA runs on a single XAUUSD H1 chart. Several of those items matter more than they seem. A hedging account is needed because six strategies hold their own positions. Leverage matters because the 1.21 note records a margin fix for 1:500 and higher leverage. Spread matters because the EA places stop orders and a user asked on 20 July 2026 whether the spread filter removes pending orders when the spread exceeds the limit; the question was not answered in the page text. A cent account matters because the 1.72 note records a virtual stop fix 'on brokers with a freeze level (Cent Account)'.

Balance decides what trades. Up to 1,000 USD only zone A runs; the 1.75 note says all zones are active from 500 USD, while the description says the second zone becomes active at 1,001 USD. A user wrote on 20 July 2026 that a small balance does not unlock all zones. The practical reading is that a 250 USD account trades fewer strategies at the broker's minimum lot, which raises the risk per trade relative to the balance, exactly as other EAs on gold warn. A stop of 50 USD at 0.01 lot is 20% of a 250 USD account, which explains the 'big risk' comment from the first day. The minimum balance is a floor for the software, not a safe starting capital.

Gold session behaviour matters as well. The EA's time filters exclude the hours with historical profit factors below 1.0 per strategy, and the 2.1 'New' profile widens them. A buyer in a broker time zone different from the vendor's should check that the filters use server time and confirm by logging a few orders. For the broader picture of how breakout and scalper designs compare, see our strategy comparison and our gold EA category.

The score, criterion by criterion: 36.0 out of 100

The total is 36.0 out of 100. Evidence is 3 because the only independent record was deleted, so what remains is checkable vendor material, which the anchors place in the 3 to 4.5 tier, and no grade A source supports it. Risk is 4, not 5 or more, because the controls are documented but not verified and the reported single-day losses are specific. Profitability is 3 because the only edge on offer is unverified and mixed. Trust is 3 for open integrity disputes and a deleted signal, offset by detailed technical replies. Fit is the highest score at 5, because the requirements are plain and workable for a motivated buyer. Because the two heaviest criteria are weak, the product lands in the band Avoid for now: demo only, until a verified record exists.

Who Lizard EA fits, and who it does not

• It fits an experienced gold trader who understands breakout systems, accepts runs of consecutive stop losses, will test on a demo through a major news release, and will size the risk down himself.
• It does not fit anyone who is buying on the strength of past monthly returns, because the signal that produced them has been removed.
• It does not fit a trader with a small account who wants all six strategies: the lower balance zones and the minimum lot make the risk per trade larger than it looks.
• It is a poor choice for prop-firm accounts with tight daily limits unless the news handling has been tested by you, since the filter has changed in nearly every version.
• Be cautious with our own v1.85 listing if you need the news fix, the freeze-level stop or any documentation. It has none of them.

How it compares with other EAs we have scored

Next to the other gold EAs we have scored, Lizard is the one with the least to check. Gold House and Pulse Engine each have a platform-built signal we could read line by line, even where the numbers were modest; Lizard's was deleted. Gold Beaver also carries a stop on every position and many presets, but its signal is live. What Lizard shares with them is a preset-heavy design, and what it adds is a changelog that moved faster than any of them. Its 36.0 reflects an inspectable idea and a missing record. For how the designs differ, see our EA category pages for gold and scalpers.

Questions to put to the developer before you pay

1. Will you publish a new signal on a verified broker account, on a named version and preset, and keep it running for twelve months?
2. What was the broker, account type and symbol of signal 2372821, and why was its symbol changed from XAUUSD.ECN to XAUUSD.STD in June 2026?
3. What are the default risk per trade, the stop size and the preset of version 2.1, and what is the maximum drawdown in a tick-based test with your own settings?
4. Which versions had the news filter failure reported at the 12 August 2026 CPI release, and what exactly was fixed in 1.87?
5. Why is the NFP filter off by default in 1.92, and why does the description still advertise it?
6. Will you post earlier versions or a full changelog of default settings, so that buyers can reproduce an older result?
7. How many of the six strategies are active at each balance level, and is the 'eight-zone' basket described in the comments the same as the six strategies in the description?
8. How does your answer to a reported loss differ from a support process, and where do buyers get help now that the Discord and the Telegram group have changed?