The engine behind every strategy

What the AI on this platform is

Stripped of mystery, the engine is a large-scale statistical system tuned to financial markets. Its job is to find order in streams of numbers: prices, volumes, volatility, correlations. It does this continuously and consistently, and it expresses its conclusions as probabilities, never certainties, which is honest, because markets do not offer certainties.

Its role is support, not substitution. The engine handles the parts of trading that humans do poorly at scale: watching dozens of instruments at once, applying the same rules every time, and acting without hesitation or hope. The decisions that matter, how much to risk and when to stop, remain with you, informed by the engine's output and your manager's explanation of it.

How the technology works, end to end

What the engine actually analyses

Price behaviour comes first: level, direction, speed and the recurring shapes movement makes across timeframes from minutes to weeks. Volume comes second, distinguishing moves with broad participation from drift on thin trade. Volatility comes third, both as a tradable condition and as a governor on position size. The fourth stream is history: how conditions resembling today's resolved across this asset and the class, which is the raw material of the probability score.

Correlations complete the picture. Assets move together and apart in regimes; when the engine sees a normally independent pair lock step, it reads regime change, tightens exposure and logs the observation. None of these streams is secret; what the engine contributes is applying all of them, consistently, at machine speed.

What the approach gives you

Who benefits from it

People whose calendars do not have room for a second profession. The engine watches the market during your meetings, your shift and your sleep, and hands you the summary. Users who want the discipline of a systematic approach without building one themselves also benefit: the rules exist, are applied identically for every account in a strategy, and are documented where you can read them.

Sceptics benefit too, and we mean that seriously. If you distrust automation, start at the smallest size, read the activity reports for a month, and question everything in them with your manager. The engine earns trust the same way a colleague does: by being checkable, not by asking to be believed.

Using it: four steps

A worked example of the engine reacting

Suppose an asset has traded in a narrow band for two days on declining volume, and your strategy is set to a conservative class. Early on the third day, price breaks the band's upper edge and volume triples within minutes. The engine scores the pattern: history shows such breaks resolving higher roughly two times in three on this asset, but it also notes that realised volatility has jumped above the strategy's tolerance.

Given the conservative class, the engine takes the modest position the class allows rather than the larger one the signal alone would justify, places the exit closer than usual because of the volatility, and logs the reasoning. Two hours later, if the move accelerates, the position pays and the report explains the sizing; if the move collapses, the close exit caps the loss and the report explains that too. The same event, run twice, would have produced two human stories; it produces one engine record.

Where the engine's limits are

Stating limits builds more trust than hiding them, so here they are. The engine sees markets through data, not through understanding: it can misread a regime change until the data confirms it, and a fast structural break can hurt before the re-scoring catches up. It optimises the strategy you chose, not your life: if your situation changes, the settings need changing with it, and the engine will not notice on your behalf.

It is also only as good as its inputs: a venue feeding delayed or thin data produces weaker scores, which is why coverage sticks to established assets and exchanges. None of these limits is fatal; all of them are reasons the manager layer exists, and why your own decisions about amounts and pauses are built into the product rather than bolted on.

Questions about the engine