AI Trading

What AI trading actually is

AI trading is the use of statistical and machine-learning models to decide when to buy and sell, how much to commit and when to get out. Instead of a person watching a screen and reacting, the model reads thousands of data points every second and acts according to rules that were tested before any real capital was used.

The label "AI" is often used loosely, so it helps to be precise. The models used here are not general-purpose chatbots and they do not "know" the future. They are narrow tools trained to spot repeatable patterns in price, volume and volatility, and to size a position so that being wrong costs a small, known amount. Their edge, where there is one, comes from consistency and speed rather than prediction.

The practical benefits are discipline and coverage. A model does not get tired, does not chase a loss, does not hold a position because it hates admitting a mistake, and can watch many markets at once, around the clock. The practical limits are equally real: models learn from the past, and markets change.

Why investors use it

Most individual traders lose money, and the reasons are mostly behavioural: cutting winners early, holding losers too long, trading too often after a bad run and ignoring risk limits. A rules-based system is designed to remove those habits. Every trade has a defined entry, a defined exit and a defined maximum loss before it is placed.

A second reason is diversification of method. Because several independent strategies run side by side — some that follow trends, some that bet on prices returning to an average, some that trade breakouts — the pool does not depend on one idea being right. When one approach struggles, another may hold up, although in a market shock all of them can lose together.

A third reason is access. Professional-grade data feeds, execution technology and risk systems are expensive to build and run. Pooling capital lets smaller investors use infrastructure that would otherwise be out of reach.

The strategy families

The pool is not a single algorithm. It is a set of strategies, each with a different job, combined under one risk budget. The mix is reviewed regularly and can change as conditions change.

How risk is controlled

Risk control is the most important part of the service and it is deliberately boring. Every order is sized so that, if the stop-loss is hit, the pool loses only a small fixed fraction of its value. Several positions that tend to move together are treated as one exposure, so the pool cannot quietly pile into one idea under different names.

Beyond the per-trade rules there are limits on total exposure per market, on total exposure per day and on drawdown. If a strategy loses more than its allowed amount, it is switched off automatically and reviewed by a person before it can resume. These limits are set by the risk desk, not by the models.

Human oversight does not mean people override the models on a hunch. It means people own the boundaries: what the models are allowed to trade, how large they can get, and when to stop them. That division — machines execute, people supervise — is intentional.

What a realistic experience looks like

Expect variability. Systematic strategies tend to have losing streaks that are uncomfortable but normal, followed by stretches where they recover. A series of small losses interrupted by a few larger winning trades is a typical shape for trend-following; the opposite shape, many small wins with occasional large losses, is typical of mean-reversion. Neither is a smooth line upward.

Be wary of any provider that shows a perfect equity curve. Back-tests, which replay a model over past data, are built with the benefit of hindsight and almost always look better than live results. We show how a strategy is built and how it is risk-managed, and we do not promise what it will earn.

The right way to use this service is as one part of a wider portfolio, with money you do not need in the short term, and with a clear view of the maximum loss you could accept.

What it covers

Trend following

Rides sustained moves in currencies, indices and commodities. Enters when price breaks out of a range and exits when the move loses momentum. Tends to do well in strong, one-directional markets and poorly in choppy ones.

Mean reversion

Takes the opposite view when price has stretched far from its recent average, betting on a return toward it. Works in range-bound markets; can lose heavily if a stretch becomes a genuine new trend.

Breakout and volatility

Looks for compression followed by expansion — quiet markets that suddenly move — and for scheduled events such as central-bank decisions and employment data.

Cross-market signals

Uses relationships between markets, for example how the US dollar, gold, bond yields and oil tend to move relative to each other, to inform timing and sizing.

Digital assets

A small, capped sleeve in large, liquid digital assets, traded with tighter risk limits because of their higher volatility and round-the-clock trading.

Risk overlay

Not a strategy in itself, but the layer that nets correlated positions, scales exposure down when volatility rises and enforces drawdown limits across all strategies.

What drives the outcome

Market volatility — strategies need movement to trade, and flat markets produce few signals.
Regime changes — a model trained on trending conditions can underperform when markets turn range-bound.
Execution quality — spreads, slippage and latency erode returns, especially at high trade frequency.
Risk discipline — the cap on loss per trade matters more to long-run results than any single winning signal.

Who it is for

Investors who want systematic, rules-based exposure to markets rather than discretionary stock picking.
People who can accept meaningful short-term losses and have a horizon of at least several months.
Those who value transparency about method and risk over headline return claims.
Not suitable for money you may need at short notice, or for anyone seeking a guaranteed or fixed return.

How the process works in practice

Data ingestionModels consume tick-level prices, volume, volatility measures and scheduled economic data from multiple liquidity sources.
Signal generationSeveral independent strategies — trend following, mean reversion and breakout — each produce entries and exits. A signal is only acted on when strategies agree beyond a set confidence level.
Risk sizingEvery order is sized so that a stop-out costs a small, fixed fraction of the pool. Correlated positions are netted so one idea cannot dominate.
ExecutionOrders route automatically to the best available price, with slippage and spread tracked against a benchmark.
Supervision and reportingHuman risk managers review drawdown and exposure daily. You receive statements showing positions, realised results and fees.

Your journey as an investor

Day 0Account verified and funded. Your deposit is allocated to the pool at the next valuation point.
Week 1First statement. You see your share of the pool, current exposure by market and fees charged.
Month 1First full monthly report with realised and unrealised results and a drawdown figure.
End of holdMinimum holding period ends. You may withdraw, stay invested or add to your position.

Questions investors ask

Is a human involved?

Yes. The models generate and execute trades, but a risk desk sets the limits and can pause any strategy.

Is a return guaranteed?

No. No strategy here has a guaranteed or fixed return, and you can lose part or all of your deposit.

Which markets are traded?

Major currency pairs, equity indices, gold, crude oil and large-cap digital assets. The exact mix changes with market conditions.

What happens in a market crash?

Sharp, fast moves are the hardest environment for any model. Stops can be filled at worse prices than intended (slippage), and correlations rise so diversification helps less. Exposure limits and automatic strategy shut-offs are designed to contain the damage, but losses can still be significant.

How are fees charged?

Fees are disclosed before you deposit and shown on every statement. They can include a management fee and a share of profits above a high-water mark. Fees reduce returns and apply whether or not a period is profitable.

Is my capital segregated?

Capital is held in pooled accounts at regulated brokers and custodians. Pooling is what allows the strategies to be run efficiently; it also means you share in the pool's results, good or bad.

Can I choose which markets are traded?

Exposure limits can be discussed with the desk on request. By default the desk allocates across the full set.

Will you tell me if a strategy is switched off?

Yes. Material changes to the strategy set are described in your next statement.

Glossary

Drawdown: The fall from a previous peak in value to a later low. Maximum drawdown is the worst such fall over a period.
Slippage: The difference between the price a model expected and the price at which an order was actually filled.
Back-test: A simulation of how a strategy would have performed on historical data. It is not evidence of future results.
Leverage: Using borrowed or margined capital to take a larger position than your deposit alone would allow. It magnifies both gains and losses.
High-water mark: The highest value your account has reached. Profit-based fees are only charged on gains above it.