Algorithmic Trading

Rule-based systems that execute the same plan in the same way every session. No hesitation, no revenge trades, no talking themselves out of a stop loss.

The case for automation

The strategy is rarely the problem

Most traders do not fail because their method is wrong. They fail because they cannot execute it identically for two hundred trades in a row.

A rule that is followed on Monday and abandoned on Thursday is not a strategy, it is a mood. Automation removes that variable. Our algorithms take every valid signal, place the stop at the same moment as the entry, and size each position from the account's risk setting rather than from how confident anyone feels that morning.

They also do not get tired. Gold and the forex majors trade around the clock from Monday morning in Sydney to Friday evening in New York, and a system that watches all of it will see setups a human sleeping through the Asian session simply never gets to take.

Automation is not a guarantee

An algorithm removes emotional error. It does not remove market risk, and it will have losing periods — by design, since it takes every signal rather than only the comfortable ones.

Why our systems
  • Emotional decision-making is eliminated The system cannot move a stop loss because it "feels like it will come back".
  • It trades 24 hours, five days a week Asian, London and New York sessions are all covered without a gap.
  • Tested across differing market conditions Trending, ranging and high-volatility regimes — not just the period that flatters the results.
  • Built for gold and forex majors Tuned to XAU/USD and the major pairs, rather than sold as a universal money machine.
  • Monitored in real time The desk watches execution quality and can halt a system if conditions turn abnormal.
Risk controls

What stops an algorithm running away

An automated system that has no limits is not an advantage, it is a faster way to lose. These constraints are hard-coded, not discretionary.

Stop loss on every order

The protective stop is submitted with the entry, not after it. No position is ever left in the market without a defined worst case.

Exposure ceilings

The system caps how many positions may be open at once and the total risk they may carry, so a run of signals cannot compound into one oversized bet.

Daily loss limit

If losses reach the daily threshold, the system stops trading for the session and resumes the next day. Bad days are not allowed to become bad weeks.

News-event filters

Trading is suspended around scheduled high-impact releases where spreads widen and slippage makes any stop-loss level unreliable.

Correlation awareness

Gold and dollar pairs frequently move together. The system recognises when several signals are really the same trade and reduces size accordingly.

Manual override

The desk can disable any system instantly. In genuinely disorderly markets a human decision to stand aside beats any rule written in advance.

Development

How a system reaches a client account

Backtest results are the easiest number in this industry to manufacture. That is why they are the first stage of our process, not the last.

Hypothesis

A rule is written down from something the desk has observed repeatedly in live trading — a session behaviour, a level reaction, a volatility pattern.

Historical testing

The rule is tested across years of data covering trends, ranges and shocks. A system that only works in one regime is rejected here.

Forward testing on our own capital

It then runs live, with real money and real slippage, at small size. This is where most candidates fail and where the honest data comes from.

Client deployment and monitoring

Only after a live period does a system run on client accounts, starting small and monitored continuously for behaviour drift.

In practice

What running the algo means for you

The system runs on your own broker account under the same limited trading authority used for portfolio management. Nothing about the custody arrangement changes — only who decides when to trade.

  • Consistency of execution The two-hundredth trade is handled exactly like the first.
  • Human error removed from execution No mistyped lot sizes, no forgotten stops, no missed exits.
  • Complete auditability Every entry, exit and rejected signal is logged and can be reviewed with you.
  • Almost no time required from you Check the terminal when you want to. Read the monthly report when it arrives.
Algo or managed desk?
  Algo PMS desk
Decision maker Coded rules Human traders
Session coverage All sessions London & New York focus
Adapts to unusual events Only within its rules Yes, by judgement
Consistency Absolute Very high, but human
Trade frequency Higher Selective

Many clients run both, splitting capital across two accounts so the two approaches are measured separately.

Automated trading

See whether the algo suits your account

We will walk you through the risk settings, the daily limits and what a realistic bad month looks like before you commit anything.