My EA Turned $500 Into $900 in 16 Weeks. Six of Those Weeks Made $14.68.
Published Nov 5, 2026, 09:56 AM · 10 min read
Between 17 July and 5 November 2026, Apex Drawdown Zero V9 traded a real RoboForex account from $500.00 to $900.09 — a gain of 80.02% on 113 trades, every one of them placed by the robot. No deposits after the first, no withdrawals, no manual trades, no grid and no martingale.
That is the sentence everybody screenshots. Here is the one nobody puts on a sales page: in the six weeks since I last published this account it has made $14.68. Thirty-nine trades, a profit factor that fell from 1.94 to 1.53, and a drawdown that nearly doubled. Both of those things are true about the same robot in the same four months, and you cannot understand the first number without the second.
The full MT5 report is published as-is: open the raw account report for 37376132. Every figure in this article is from that file.
The numbers, in one place
Source: MetaTrader 5 account report, RoboForex Ltd real account 37376132, exported 5 November 2026 at 11:15. Currency USD.
| Period | 17 Jul – 5 Nov 2026 (111 days, 15.9 weeks) |
| Deposit | $500.00, once. No further deposits, no withdrawals. |
| Closing balance | $900.09 |
| Net profit | +$400.09 (+80.02%) |
| Gross profit / gross loss | $1,148.76 / $748.67 |
| Profit factor | 1.53 |
| Trades | 113 — 55 long, 58 short. 100% placed by the EA, 0 manual. |
| Winners | 64 of 113 (56.6%) |
| Average win / average loss | $17.95 / $15.28 — expectancy $3.54 per trade |
| Best trade / worst trade | +$153.84 / −$53.10 |
| Longest streaks | 5 wins (+$193.06), 5 losses (−$31.78) |
| Maximum drawdown | 20.3% equity peak to trough. Worst single day −10.29%. |
| Recovery factor | 2.07 |
| Average holding time | 20.5 hours, about 7 trades a week |
| Swap and commission paid | $0.00 — this is a zero-commission account type, so your net will be lower |
Month by month: one month made 91% of it
| Month | Profit | Return |
|---|---|---|
| July (from the 17th) | +$3.44 | +0.69% |
| August | +$365.11 | +72.52% |
| September | +$29.22 | +3.36% |
| October | −$46.61 | −5.19% |
| November (to the 5th) | +$48.93 | +5.75% |
| Total | +$400.09 | +80.02% |
August produced $365.11 of the $400.09. That is 91.3% of four months of profit inside one month, and a single day — 22 August, +$149.18, +23.2% — produced 37.3% of it on its own.
Read that as a warning, not as a boast. A gold breakout system earns in bursts: it waits for the market to hand it a large directional day and takes one trade a day while it waits. Over five years of back testing that averages out. Over sixteen weeks it means your result depends enormously on whether you happened to switch the robot on in July or in September. Anyone who started this EA on 1 September is up about 3% today, trading the identical software.
The six weeks nobody screenshots
On 25 September I published this same account at $885.41. Here is that report against this one:
| 25 Sep 2026 | 5 Nov 2026 | |
|---|---|---|
| Balance | $885.41 | $900.09 |
| Return | +77.1% | +80.0% |
| Trades | 74 | 113 |
| Profit factor | 1.94 | 1.53 |
| Max drawdown | 13.5% | 20.3% |
Thirty-nine more trades bought $14.68. The profit factor did not drift, it dropped. The drawdown did not drift either.
I know exactly why, because the same thing shows in the back test over the same weeks. Gold fell from roughly 4,370 to 3,917 across September and October, and this EA's long trades are where it bleeds in a falling market: in back testing over that window the long side lost while the short side made money, and five trades went straight to the stop without ever reaching the first scale-out. The robot is not broken. It met a market shaped wrong for it, which is a thing that happens to every strategy and which the sales page of every other EA somehow never mentions.
Symbol by symbol: the part that actually matters
Before this account was opened I published a back test that said, in writing, trade gold with this EA and do not trade EURUSD. Sixteen weeks of real money later:
| Symbol | Net | Trades | Profit factor | What I said beforehand |
|---|---|---|---|---|
| XAUUSD | +$464.64 | 40 | 1.88 | The validated pair. Trade it. |
| EURUSD | −$3.08 | 44 | 0.97 | A steady bleed over a full year. Do not sell it as validated. |
| EURJPY | −$61.47 | 29 | 0.42 | Never properly tested. Unknown. |
Gold made $464.64. The two pairs I warned about lost $64.55 between them. The one that had never been tested at all is the one that lost the most, on the fewest trades.
This is the single most useful thing in this report and it has nothing to do with the profit. A back test that correctly predicts which symbols will fail is worth more than one that predicts a profit figure. The 2026 year-to-date back test on gold produced a 1.88 profit factor live against 1.94 tested, and EURUSD came in at 0.97 live against 0.94 tested. Those are different windows — the back test runs from January, the account from 17 July — so treat them as agreeing in direction and magnitude rather than as a like-for-like match. The direction is the part that matters: the model called all three symbols correctly, including the two it called unprofitable.
And the obvious correction — turn the other two pairs off — is not as clean as it looks, because position size is a percentage of the current balance, so those 73 losing trades also changed the size of every gold trade that followed them. You cannot simply add $64.55 back. What you can do is run the EA on XAUUSD only, which is what the shipped presets are tuned for and what I would do if I were starting this account again today.
Would this pass an FTMO challenge? No — and the arithmetic is worth seeing
This is the question I get most, so here is the honest answer rather than the one that sells a licence. FTMO's published limits as at 5 November 2026 (trading objectives) are:
- 2-Step Challenge: 10% profit target, then 5% in Verification. Maximum daily loss 5% of initial capital. Maximum loss 10%, static. Minimum 4 trading days.
- 1-Step Challenge: 10% profit target. Maximum daily loss 3%. Maximum loss 10%, trailing on end-of-day balance. Plus a consistency rule: your best day may be no more than 50% of total profit.
Now hold this account against them. All the limits are percentages, so the $500 size is irrelevant — the comparison is fair at any account size.
| Rule | Limit | This account | |
|---|---|---|---|
| Profit target | 10% | +80.0% | Cleared, eight times over |
| Minimum trading days | 4 | ~80 trading days | Cleared |
| Consistency (best day ≤ 50% of profit) | 50% | 37.3% | Cleared |
| Maximum daily loss | 5% / 3% | −10.29% on 1 August | Breached |
| Maximum loss | 10% | 20.3% peak to trough | Breached |
So an account that made 80% would have been closed in week three. It would have hit the 10% profit target on 1 August and been killed by the daily-loss rule on the same day.
That is not a flaw in the EA, it is a statement about the risk setting, and it is the thing almost nobody explains before they sell you a robot for a prop challenge. The risk percentage per trade is an input. Everything above was produced by a setting chosen for a personal account where a 20% drawdown is survivable, and a prop challenge is a different game with a different number in that box.
Measured on the 2022–2026 back test, the trade-off looks like this (returns compounded, drawdown measured as relative equity drawdown, which is the honest one — MT5's "maximal" figure flatters low-risk runs):
| Risk per trade | Return per year | Max drawdown | Fits a 10% prop limit? |
|---|---|---|---|
| 0.5% | +9.2% | 5.7% | Yes, with room to spare |
| 1% | +17.7% | 11.1% | Marginal — one bad run ends it |
| 2% | +37.7% | 21.7% | No |
| 3% | +62.6% | 30.7% | No |
Those are back-tested figures on XAUUSD H1 over 4.8 years, not live results, and they were modelled on 1-minute bars rather than real ticks — a real-tick re-run is in progress and I will correct this table here if it moves. The point of the table is its shape, not its decimal places: drawdown scales with the risk input about as fast as return does, and the prop-safe settings are the slow ones. Notice too that the live account's 20.3% drawdown sits beside the 2% row, while the 80% gain does not — four months of gold caught at the right moment is not an annual rate and I am not going to present it as one.
If your goal is a funded account, the honest version is: run the 0.5% preset, expect something closer to 1% a month than 70%, and understand that you are buying a tool that will not breach the rules rather than a tool that will pass the challenge next week. Anything advertised as a "prop firm EA" that also shows you an 80% chart is selling you two settings and showing you one. See also our guides on whether EAs are allowed at prop firms and passing a prop firm challenge.
What 113 trades does and does not prove
I am going to argue against my own evidence for a paragraph, because you should hear it from me rather than work it out after buying.
- 113 trades is a small sample. Gold contributed 40 of them. Forty trades cannot distinguish a real edge from a good quarter, and anyone who tells you otherwise is counting on you not knowing that.
- One broker, one feed. RoboForex-Pro, zero commission, no swap charged on this account. On a broker with commission and wider spreads, every one of these trades is worth less. My own tick-data measurement of this feed put the spread cost at around $1.90 per round trip on gold, which is 13.5% of the risk on a standard stop.
- Four months includes one extraordinary month. Strip August out and the remaining twelve weeks made $34.98 on $500.
- It is my own account. It is real money and the raw report is published, but a vendor publishing his own account is still a vendor publishing his own account. The back tests, the per-symbol predictions and the losing months are there so you can check whether the story holds together rather than take my word for the total.
What it does prove is narrower and, I think, more useful: the software does on a live account approximately what the back test said it would, including on the two symbols where the back test said it would lose money. That is the only claim here I would defend.
How to see it yourself
- The raw MT5 report for account 37376132, exactly as MetaTrader exported it.
- Apex Drawdown Zero V9 — the EA that placed all 113 trades, with the XAUUSD presets referenced above, including the 0.5% prop preset.
- The settings guide — every input explained, including the risk input that decides which of the rows in that table you get.
- Gold Vigil — a different gold EA, free, in open beta, if you would rather test something at no cost before you spend anything.
- Five ways a back test flatters an EA — the checks I run on my own work, so you can run them on mine.
Questions people actually ask
How much did Apex Drawdown Zero V9 make on a live account?
On RoboForex real account 37376132, it turned a $500 deposit into $900.09 between 17 July and 5 November 2026 — a net profit of $400.09, or +80.02%, across 113 trades with a profit factor of 1.53 and a maximum drawdown of 20.3%. 91.3% of that profit was made in August alone; the six weeks to 5 November added $14.68.
What was the maximum drawdown on the live account?
20.3% from equity peak to trough over the 111-day period, with a worst single day of −10.29% on 1 August 2026. The previous report, six weeks earlier, showed 13.5%, so the drawdown nearly doubled in the second half of the period while the balance barely moved.
Can Apex Drawdown Zero pass an FTMO challenge?
Not at the risk setting used on this account. FTMO's limits are a 5% maximum daily loss and a 10% maximum loss on the 2-Step Challenge, and 3% / 10% on the 1-Step. This account's worst day was −10.29% and its worst drawdown 20.3%, so it would have breached both while still showing an 80% gain. The EA ships a 0.5%-risk preset measured at 5.7% maximum drawdown in back testing, which is the setting built for prop rules — at a far lower return.
Which pairs does this EA actually make money on?
Gold. On this live account XAUUSD made +$464.64 over 40 trades at a 1.88 profit factor, while EURUSD lost $3.08 over 44 trades and EURJPY lost $61.47 over 29. Both losses were predicted in the published back tests; EURJPY had never been properly tested and performed worst. The presets are tuned for XAUUSD on the H1 chart and that is the only symbol we recommend.
Is the win rate really only 56%?
Yes — 64 winners out of 113 trades. A high win rate is not the point: the average win was $17.95 against an average loss of $15.28, which is what makes a 56.6% win rate profitable. Be careful comparing this to EAs advertising 90% win rates; a system that risks a lot to make a little needs a very high win rate just to break even, so the headline number on its own tells you nothing.
Does the EA use grid or martingale?
No. Every trade has a fixed stop loss, position size is a set percentage of the current balance, and losing trades are never averaged into or doubled. The largest single loss on the account was $53.10 and the longest losing streak was five trades costing $31.78 in total.
Why publish the bad weeks?
Because a results page that only ever goes up is the clearest sign of a fabricated track record, and because the useful information in a live report is in the losing stretches. The six weeks that made $14.68, the 20.3% drawdown and the two losing symbols are in this article for the same reason the profit is: so the whole thing can be checked against the raw MT5 report, which is published in full.
The required warning, meant sincerely
Trading leveraged products carries a high risk of loss and you can lose more than you deposit. Every figure in this article describes one real account, at one broker, over 111 days — it is not a projection and it does not predict what the same software will do on your account, at your broker, in the market you will actually meet. A 20.3% drawdown happened here; a larger one is possible. Trade on a demo account until you have seen the robot take a losing streak and you know how it feels, and never fund it with money you need.
