Nina Systems Portfolio

A robot here is an Expert Advisor (EA): a program that opens and closes trades on its own. I run four of them together:

- Two trailing robots on gold (the ticker XAUUSD). A trailing robot stays with a winning trade and slides the exit up behind the price, so a good move gets locked in. These usually hold a trade for one to three hours.
- One grid robot on gold. A grid places several trades at different prices instead of one, and holds them a little longer.

- One grid robot on AUDCAD, the Australian dollar priced against the Canadian dollar.

WHY FOUR ROBOTS INSTEAD OF ONE

They rarely have a bad day at the same time. When one is struggling, the others usually are not. How much two systems move together is called correlation, and keeping it low is the entire idea. Spreading the account across four low-overlap systems makes the combined swings smaller than any single one of them alone. In my testing, the four together dropped about half as much as they would have if their worst moments had all lined up.

HOW I PICKED AND VETTED THEM

I did not just switch four robots on and hope. Each had to pass a few filters before it earned a place:

- Reputation. Each is a robot other traders have bought and rated, not something I threw together overnight.
- Long, clean backtest. I tested each from 2017 to today on tick data rated 99% modeling quality, so the results do not lean on cheap, approximated prices.
- Slippage stress. I re-ran them with slippage set two to three times worse than normal, to keep only what still holds up on bad fills.
- Spread stress. I checked them against wider spreads as well, and I keep watching spread on the live account.
- Monte Carlo. I did not trust one run of history. I reshuffled it thousands of times to see how bad a stretch can get, and I size the account from that instead of from the best case.

THE HONEST PART: HOW FAR IT CAN FALL

Drawdown is how far the account falls from its highest point before it climbs back. It is the part that hurts, and it is the number most signal pages bury. I lead with it instead.

"Worst so far" is the deepest fall across about nine and a half years of testing, from 2017 to 2026. "Bad-run model" comes from reshuffling that history thousands of times to see how ugly a stretch could get if the bad days bunched up (a method called Monte Carlo).

- Start with $1,500: worst so far around 71%, and close to the full account in a bad-run model.

- Start with $2,500: worst so far around 43%, around 63% in a bad-run model.

- Start with $5,000: worst so far around 21%, around 31% in a bad-run model.

Read that first line carefully. At a small balance like mine, a single violent move in gold can take most of the account. Two things you should hold together. In your favor, these drops have been sharp but short: the worst one in the whole test, the gold crash of January 2020, was back to its old high within a day or two. Against you, on a very small account a fall that deep can trigger a margin call and close the trades before they get the chance to recover, so the recovery you see in the test on a larger balance is not promised on a tiny one. That is the deal you accept for fast growth, and I accept it on my own account every day. Put in more money and the same swings become a smaller share of it, which is why $5,000 rides far calmer than $1,500.

WHO THIS IS FOR

Copy it if you want to grow a small account,you understand it can drop by a third or more along the way, and you will leave it running through that instead of closing in a panic. Skip it if a drawdown that deep would wreck your week. There is no shame in that, and the calmer signal I am building later will suit you better.