Gold grid EAs and the move that does not come back

A gold grid EA is undone less by a trend than by a move that does not come back while the account can still hold the open positions, which the equity drawdown shows and the balance hides.

• A gold grid EA is not undone by a trend as such. It is undone by a move that does not come back inside the account’s capacity to hold it.
• The balance line can sit flat while the equity line falls, so grid EA drawdown hides in the equity fields of the report, not the balance ones.
• The method is two Strategy Tester windows, not one: the longest directional stretch in your own data, and the turn that ends it.

The phrase people type is xauusd grid ea trending market, and its premise is slightly wrong. A trend is not what kills a basket — one of our own grid-labelled Products is described in the Catalogue as suited to trending gold years. What ends a basket is a move that does not come back while the account can still hold it, and the turn that arrives after a long stretch of being right. This guide is about testing for that.

What a gold grid EA is actually betting on

The families are set out in gold EA strategy families explained, so assume that page’s definitions. What is worth restating is the shape of the bet.

A grid or a basket is positioned against a move and is betting that the move resolves: either price comes back to where the set of positions can be closed together, or the position it still holds is on the right side when the market turns. Neither is simply a bet that price returns to a level — that fits some grids and not others, and not a strategy that holds one position and follows the trend out.

The exposure is duration and distance, not direction: a move that keeps going, far enough and long enough that the account cannot carry the open positions until the resolution arrives. That sentence is the whole of a basket’s risk on XAUUSD.

Grid EA drawdown: why the balance curve looks calm and the equity curve does not

This is the mechanism, and the reason these strategies surprise people.

MetaTrader 5 keeps two account values: MQL5 defines ACCOUNT_BALANCE as “Account balance in the deposit currency” and ACCOUNT_EQUITY as “Account equity in the deposit currency” (MQL5 documentation). The tester’s visualisation help says which is which: “Equity — the amount of money taking into account the results of the currently open positions” (MetaTrader 5 Help).

A basket under water has realised nothing yet. Balance does not move until a position closes, so the balance line stays flat — or keeps stepping up, if other positions in the basket are closing well — while unrealised loss piles up underneath it. The report reflects that split faithfully: Balance Drawdown Maximal is the “difference in deposit currency between the highest local balance value and the next lowest account balance value”, and the equity family is computed the same way on equity values (MetaTrader 5 Help). So the balance drawdown field can read clean on a run whose equity drawdown was severe. The equity drawdown field is the one describing what you would have had to sit through; the fields themselves are unpacked in how to read a backtest report.

Two windows to run, and what to read when they finish

One run is not a test of a basket. Two are the minimum.

Window one: the largest sustained directional move in your own data. Not the period a seller picked — find the longest one-way stretch your broker’s history contains, and set the tester’s date range to it.

Window two: the choppiest reversal stretch you can find. Helix Basket’s risk line points straight at this: it is “For traders who want a diversified, hands-off gold basket and accept sharper swings when the trend turns.” A test covering only the trend has tested the easier half, and a nine-month window inside one kind of year is the weakest evidence a basket can offer.

When each run finishes, read the equity drawdown fields rather than the balance ones. Then re-run in visual mode and watch the open positions table and the equity line, not the balance curve: that is where the worst episode is visible while it happens. Then find how long it lasted and whether it sits early or late in the run — one in the last fortnight is a strategy you caught mid-drawdown.

Three of our six gold EAs are grid strategies

Three of the six live Products carry the style grid in the Catalogue: Helix Basket, Aurum Recoil and Helix Trail. Worth stating rather than hiding: Helix Trail is labelled grid although its description holds one position at a time and follows the trend. Read the description, not the label.

Helix Trail’s own risk line is “For active traders who follow the trend and can sit through deeper pullbacks.” And Aurum Recoil’s Deep Margin profile carries the bluntest note we publish anywhere: The desk’s multi-year study of this configuration on a $25,000 account ended in account loss in 2023, in 2024 and in 2025; only the 2026 window is published here.

The other three Products are not safer. They have a different failure shape.

One backtest window is one story

The published Helix Basket run is a Strategy Tester run: 1 January 2026 to 7 September 2026, XAUUSD on an IC Markets demo server, on a $25,000 deposit, History Quality 63 % because that feed serves real ticks only from 2026-04-21. Its maximum equity drawdown reads 10.71 % over 1,551 trades. That percentage is unreadable without the deposit beside it, which is the argument of gold EA account size and drawdown.

The desk’s own records are the counterweight, and they describe the families these Products come from rather than the published configurations. In the same 2026 window, a reverse-grid family run at its source defaults on a $25,000 account produced −$15,714 with 21 stop-outs. Over the earlier years, earlier versions of the trailing-grid family produced ruins in 2023 and in 2025.

The full report is at the Helix Basket backtest. Then run your own two windows: every live Product has a free Backtest Edition that runs inside the Strategy Tester on your own broker’s ticks, from the strategies page.

Before you act on any of this

Every figure above comes from the MetaTrader 5 Strategy Tester, never from live trading, and no result is promised anywhere on this site. Two departures apply to all of them: the tester charged no commission on XAUUSD, while real trading on that account cost roughly $7 per lot round-turn that the net does not deduct; and fills are modelled rather than executed by a broker. Trading gold on leverage carries a high risk of losing money rapidly — read the risk disclosure before you act on a gold grid EA of any kind.