Two hundred and thirty-five trades. Two of them were short. Both lost.
That is the entire short book of my gold EA, and I am opening with it because it is the fastest way to explain what this product actually is. It is not a system that trades gold. It is a system that buys gold, waits, and does nothing at all the rest of the time.
Let me be precise about what those two trades prove before anyone builds a theory on them: nothing. Two trades is an anecdote, not evidence. You cannot conclude anything statistically from a sample of two, and anyone who tries to sell you a lesson drawn from two data points is selling you a story.
What is worth discussing is the design decision behind the other 233, because directional specialisation is a genuine choice with real costs, and almost nobody explains why they made it.
Quick Navigation
- Why gold does not go up and down the same way
- What specialisation actually costs you
- The argument against my own design
- The numbers, since I opened with one
- What it is and who should skip it
Gold does not go up and down the same way
The core reason to specialise on the long side of this instrument is that its moves are not symmetrical, and the asymmetry comes from what gold actually is.
Gold produces nothing. It pays no dividend, no coupon, no rent. Nobody buys it expecting the metal to work. It is held as a hedge, against currency debasement, against falling real yields, against the possibility that something breaks. That gives its demand a very particular shape.
Fear arrives faster than it leaves. When the reason to hold gold appears, the buying is urgent, concentrated and violent, because the people who want the hedge want it now. When the reason fades, nobody sells in a panic: positions are unwound gradually, over weeks, because there is no forced exit. So the up-moves tend to be fast and directional, and the down-moves tend to be slow, choppy grinds.
That distinction is not philosophical, it is operational. Fast directional moves are what trend-following logic is built to capture. Slow choppy grinds are what shreds it. The same system applied to both sides is not one system, it is a good system and a bad system stapled together, and the bad half quietly eats the good half’s profits.
There is also a structural floor under the demand that has no symmetric counterpart on the supply side, and the fact that gold is priced in a currency that has historically lost purchasing power over long horizons. Neither of those is a prediction. They are reasons the drift has a direction.
What specialisation actually costs you
This is the part that gets left out of every “specialised EA” pitch, so here it is plainly.
You sit out entire phases of the market. When gold is falling or going sideways for months, the EA does very little and your capital does very little with it. There is no clever hedge, no reversal mode, no “adapting to conditions”. It waits.
Some months are nearly empty. Nine trades this month, on an account averaging about a day per position. If your definition of a working EA is one that is always busy, this will feel broken to you while it is behaving exactly as designed.
The FOMO is real. Gold drops five hundred points, everyone in every chat is talking about the short they took, and your EA is flat. Watching a move you are structurally not allowed to participate in is harder than it sounds, and it is the moment most owners of specialised systems start “improving” them.
The argument against my own design
Steel-manning this properly, because you deserve the counterargument before you spend money.
A long-only gold system is a directional bet on gold’s structure. If gold entered a genuine multi-year bear market, this EA would spend that period doing very little and occasionally getting hurt on failed long attempts. It has no mechanism to profit from sustained decline, and it is not going to grow one.
So the honest framing is: this is not an all-weather system and I will not pretend it is. It is a specialised exposure that performs when its instrument does what it has structurally tended to do, and it should be sized in your portfolio the way a directional bet is sized. Which is to say: as one component, not as the whole thing.
If a single specialised EA is going to be your entire account, you have a concentration problem no product can fix, and that is true of mine as much as anyone’s.
Why “it did nothing” is the feature
The instinct when an EA is quiet is to assume it is broken. With a specialised system it is usually the opposite: the quiet periods are the discipline working, and they are the reason the active periods are worth anything.
The practical version of this: judge a specialised system on how it behaves when its conditions appear, not on how often they appear. An EA that takes nine trades and wins eight is doing its job. An EA that takes ninety trades because you widened its filters to make it feel productive is now a different system, and not a better one.
The numbers, since I opened with one
From the live account at the time of writing, and these move, so check the account before deciding anything: 235 trades, 222 of 233 longs won, profit factor 3.89, average trade length about a day.
And the two numbers that matter more than any of that, repeated here because they are the ones that should drive your decision: maximum drawdown 38.80%, and an average loss of $77.03 against an average win of $17.55. A high win rate with a loss profile like that means one bad trade removes several weeks of good ones. That is the trade-off you are accepting, and if it does not sit right with you, this is the wrong product and no growth figure should change that.
What it is and who should skip it
DoIt Gold Guardian for MT5 (MT4 here) is $399. Set your risk percentage and attach it: every position gets an automatic stop, a take profit and a dynamic trailing stop. No martingale, no grid. Low-spread broker, $200 minimum deposit, at least 1:30 leverage, VPS recommended.
Skip it if you are running a prop firm challenge, and I mean that literally enough to have written a separate post explaining why I do not recommend it for them. Skip it if a 38.80% drawdown would make you switch it off. Skip it if you need your EA to be busy. And skip it if you were planning to put your whole account behind one specialised bet.
The close
The two losing shorts are a curiosity. The 233 longs are the product. What you are actually deciding is whether you want a system that does one thing on one instrument and stays out of everything else, including the moves that would have been fun to be in.
Most people say they want specialisation and then get restless when it means waiting. Being honest with yourself about which of those you are is worth more than any statistic on this page.
Trading involves substantial risk of loss. Past performance does not guarantee future results. Nothing here is a forecast about the future direction of gold. All figures are a snapshot of a live account at the time of writing and will have changed: verify on the live account.
