Selene
- Experts
-
Minh Truong Pham
Hi, I'm Pham, a programmer and trader from Vietnam. I write indicators and Expert Advisors for MetaTrader 4 and 5. I write every product myself from scratch, without borrowing third-party code.
How I build and sell EAs - Versão: 1.1
- Ativações: 10
Selene, XAUUSD, MetaTrader 5: $1,000 became $4,619 in a real-tick backtest from 1 January 2025 to 14 November 2026, at a fixed 0.01 lot per strategy. That's +$3,619.47 in 23 months, profit factor 1.31, 66.4% winners, deepest balance dip $469 (14.5%). Eight separate gold strategies on one chart. Stop loss on every trade. No martingale, no grid, no averaging down.
Selene costs $249. At the default 0.01 lot that backtest made $3,619.47 on a $1,000 account, so do the sum yourself. It's a backtest, not a promise, and I'll show you below where the numbers come from and where they're weaker.
At a glance
- Symbol: XAUUSD (gold), one chart for everything
- Style: breakout and momentum, pending stop orders plus one market-entry strategy
- Eight strategies running side by side, each with its own stop, its own target and its own way of protecting a winner
- Fixed lot. The same size on every trade, never raised after a loss
- Trades: about 89 a month in the test, holding from minutes to a few days
- Hidden management settings. You set lot, spread limit, which strategies run, and the time zone. The rest is locked at the tested values
The backtest, in numbers
XAUUSD, real tick data, 2025-01-01 to 2026-11-14, $1,000 start, 0.01 lot per strategy, no commission.
| Net profit | +$3,619.47 (+362%) |
| 2025 | +$2,369 |
| 2026 (to 14 Nov) | +$1,257 |
| Profit factor | 1.31 |
| Trades | 2,004 (1,214 buy / 790 sell) |
| Win rate | 66.4% |
| Max drawdown | $469 on balance (14.5%), $514 on equity (16.9%) |
| Positive months | 17 of 23 |
| Most trades open at once | 7 |
So 6 months out of 23 closed red, and the last few months of the test are among them. I'm not hiding that. The curve is in the screenshots, month by month.
I didn't want one lucky strategy carrying the whole thing, so every one of the eight had to make money on its own in this test before it stayed in. They did, from +$47 for the quietest one to +$1,253 for the busiest. And because they fire at different moments for different reasons, their bad days rarely land together: the average day-to-day correlation between any two of them is 0.11.
Where these numbers come from. The table is the same engine on tick data from one server. I also ran Jan 2025 - Nov 2026 in MetaTrader 5 on a second broker's real ticks: 2,026 trades, +$2,962 before swaps (2025 +$2,247, 2026 +$715), deepest dip $491. Two feeds, two results, same shape. Yours will be a third one, which is why I'd rather you run the tester on your own broker before you buy.
Why eight strategies and not one
A single gold system has a regime it loves and a regime that eats it alive. Selene spreads the work. Some of the eight wait for gold to break out of a range it's been building for days, some trade moves that start and finish inside one session, and one looks for strength coming back after a sell-off.
What they share is the discipline:
- every order goes out with its stop loss already attached
- profits are protected step by step once a trade runs, with rules tuned per strategy
- no new orders while the spread is wider than your limit
- pending orders that weren't triggered expire on a schedule instead of sitting there for weeks
I'm not publishing the entry rules. That's the part that took the time, and it's the part you're paying for.
Requirements
- Symbol: XAUUSD, or whatever your broker calls gold (XAUUSDm, GOLD...). Attach it to another symbol and it prints a warning and rescales its distances to that symbol's price, but nothing was tuned or tested there. Keep it on gold.
- Timeframe: any. It reads the data it needs on its own.
- Account: hedging. Two strategies can hold opposite trades at the same time and that's by design. On a netting account they'd be merged into one position, so don't.
- Capital: I'd use at least $2,000 per 0.01 lot. The test ran on $1,000; the extra room is there so you're never trading at the edge of your account.
- Broker: tight gold spread. Selene skips orders while the spread is above your limit, so a low-spread account gets you the full set of trades.
- VPS: recommended. Orders are placed around the clock, five days a week.
Inputs
| Input | Default | What it does |
|---|---|---|
| Broker server time zone | Auto | Leave on Auto for live trading. In the tester see the note below |
| Server GMT offset, hours | 0 | Only used in Fixed mode and in the tester |
| Strategies to run | 1-8 | Which of the eight run. "1-8" for all, or a list like "1,3,5" |
| Lot per strategy | 0.01 | Fixed size for every trade |
| Max spread (price units, USD) | 0.50 | No order goes out while the spread is wider than this |
| Magic base | 2026600 | Strategy N uses base + N. Keeps Selene apart from other EAs |
| Order comment prefix | Selene | Text at the start of every order comment |
The management settings (trailing, exits, daily limits) are fixed at the tested values. That's deliberate: those are the settings the numbers above belong to.
How to backtest it
Strategy Tester, XAUUSD, any timeframe, model Every tick based on real ticks, from 2025 on, $1,000 or more.
One thing to set first. Selene keeps its own clock, and in the tester it can't ask the server what time zone it's on. If your broker's server runs on GMT+2 in winter and GMT+3 in summer (most do), set "Broker server time zone" to the GMT+2/+3 option. If the server clock is plain GMT, leave Auto with offset 0. Live, Auto works it out by itself.
Questions people ask
Can I run only the best strategies? You can, with the Strategies input. I wouldn't. The mix is what keeps the drawdown where it is.
Does it trade on news? It doesn't read a news calendar. The spread filter keeps it out of the worst spikes.
Why did it skip an order today? Usually the spread was over your limit at the moment the order was due, or the account didn't have the free margin for one more pending order. Both are on purpose.
Can I raise the lot? Yes. Profit and drawdown both grow roughly in step with it. Keep the $2,000 per 0.01 ratio.
Several charts? One chart per account is enough. On a second one, change the magic base so the two don't touch each other's trades.
Is there an MT4 version? Yes, same engine and same settings, listed separately in the Market.
Updates and support? Updates are free for buyers. Write to me in the product comments or by private message here and I'll answer.
A backtest is the past. Start on a demo, then go live small. Past performance does not guarantee future results, and trading gold with leverage can lose money.
