McCortex OI
McCortex OI
Menú

The Mechanics Behind Every Trade

  • Machine Learning Model Training Every model is trained on historical and live price data across multiple timeframes, refining its parameters before a strategy is made available to traders.
  • Signal Generation Process Signals are produced when a model detects a pattern that meets its defined criteria, combining several indicators rather than relying on a single data point.
  • Backtesting Against Historical Data Before deployment, every strategy is run against historical market data to observe how its logic would have behaved across different conditions and time periods.
  • Execution Logic and Timing Once a signal is confirmed, execution logic determines order size and timing, sending instructions to the exchange without requiring the trader to be online.
  • Risk Limits Before Execution Risk limits are applied before an order is placed, capping exposure per strategy so a single trade cannot exceed the boundaries a trader has set.
  • Machine Learning Model Training Every model is trained on historical and live price data across multiple timeframes, refining its parameters before a strategy is made available to traders.
  • Signal Generation Process Signals are produced when a model detects a pattern that meets its defined criteria, combining several indicators rather than relying on a single data point.
  • Backtesting Against Historical Data Before deployment, every strategy is run against historical market data to observe how its logic would have behaved across different conditions and time periods.
  • Execution Logic and Timing Once a signal is confirmed, execution logic determines order size and timing, sending instructions to the exchange without requiring the trader to be online.
  • Risk Limits Before Execution Risk limits are applied before an order is placed, capping exposure per strategy so a single trade cannot exceed the boundaries a trader has set.
  • Machine Learning Model Training Every model is trained on historical and live price data across multiple timeframes, refining its parameters before a strategy is made available to traders.
  • Signal Generation Process Signals are produced when a model detects a pattern that meets its defined criteria, combining several indicators rather than relying on a single data point.
  • Backtesting Against Historical Data Before deployment, every strategy is run against historical market data to observe how its logic would have behaved across different conditions and time periods.
  • Execution Logic and Timing Once a signal is confirmed, execution logic determines order size and timing, sending instructions to the exchange without requiring the trader to be online.
  • Risk Limits Before Execution Risk limits are applied before an order is placed, capping exposure per strategy so a single trade cannot exceed the boundaries a trader has set.

How our AI works

McCortex OI applies machine learning models to crypto market data so that a trader does not have to watch every chart or type every order by hand. This page explains, in plain terms, how our AI trading works from the moment data arrives to the moment an order reaches the market. It also explains what the risk gate does before execution, and where the limits of this approach sit. Nothing here is a promise of a particular outcome; it is a description of a process.

What data goes in

The starting point for any model is market data: price, volume, order book depth and the pace at which conditions change across spot and futures markets. McCortex OI pulls this data continuously rather than at fixed intervals, because crypto markets do not pause outside business hours. The workspace also tracks account-level information, such as open positions and the risk limits a trader has already set, so that every decision is made in the context of what is actually held, not in isolation.

Data quality matters as much as data volume. Before anything reaches a model, it passes through cleaning steps that remove obvious errors, gaps and duplicate entries, since a single bad tick can distort a signal if left unchecked. The system also normalises data across different markets and timeframes so that a model trained on one instrument behaves consistently when applied to another. None of this guarantees a particular result; it simply means the inputs are as reliable as the platform can make them before any analysis begins.

Finally, this stage is ongoing rather than a one-time setup. Market conditions shift, new instruments appear and old patterns fade, so the data pipeline behind McCortex OI keeps refreshing what the models see. This is the foundation the rest of the process depends on.

What the models are actually doing

At the core of the platform, machine learning models look for structure in market data: recurring relationships between price movement, volume and volatility that have shown up often enough to be worth tracking. These are statistical patterns, not certainties, and the models are built to describe tendencies rather than to claim they know what will happen next.

Each model is trained on historical data and then evaluated against fresh data it has not seen before, a standard step that helps confirm a pattern is not simply an artefact of one specific dataset. This process is repeated regularly, because a pattern that held six months ago may weaken or disappear as market behaviour evolves. We do not publish internal accuracy figures or claim a fixed hit rate, because any single number would misrepresent how conditions change over time and across different markets.

The models themselves cover a range of approaches suited to different market behaviours, from trend continuation to shorter-term mean reversion, and McCortex OI runs several of them in parallel rather than relying on one. The output of this stage is not a trade. It is a signal: a structured assessment of current conditions that feeds into the next stage of the process, where it is weighed against risk limits and turned into an actual decision.

From signal to order

A signal on its own does not move money. Once a model flags a condition worth acting on, that signal passes into execution logic that decides whether, when and how to translate it into an order. This step accounts for the strategy a trader has chosen, the direction permitted, long or short, and the market in question, spot or futures, since each combination carries different mechanics.

Execution happens on our servers rather than on the trader’s device, which is part of why McCortex OI can keep a strategy running whether the trader is at a desk or asleep. The execution logic also considers order type and timing, since placing a large order all at once can move price against the trader, so the system may size or stage an order to reduce that effect where the strategy calls for it.

Backtesting plays a role here as well: before a strategy is made available in the library, its execution rules are tested against historical scenarios to check that the logic behaves as intended across a range of conditions, not just the ones it was designed around. This testing informs how the strategy is built. It is not a projection of how it will perform going forward, and no specific return or historical result is presented as evidence of future outcomes.

The risk gate before execution

Before any order reaches the market, it passes through a risk gate. This step checks the order against the risk limits a trader has configured, such as maximum position size, exposure per strategy and boundaries around loss. If an order would breach one of these limits, it is adjusted or blocked before it is placed, not after the fact.

This ordering matters. A risk check applied after execution can only report what already happened; a risk check applied before execution can prevent an unwanted position from being opened at all. McCortex OI applies risk limits at this earlier point deliberately, so that a trader’s stated boundaries are respected on every single order, not treated as a general guideline.

Risk limits are configurable per strategy, which means a trader running several strategies at once can set tighter boundaries on one and looser boundaries on another, depending on how much of the account they are willing to allocate to each. The gate does not evaluate whether a trade is a good idea in a broader sense; it checks whether the order complies with the limits already in place. Trading digital assets carries substantial risk, including the total loss of capital, and no risk gate removes that risk. It manages exposure against rules the trader has set. It does not eliminate exposure altogether.

What this approach cannot do

It is worth being direct about the limits of this system. Machine learning models describe patterns that have appeared in past data; they do not predict the future, and no configuration of McCortex OI changes that basic fact. Markets can and do behave in ways that have no clear precedent, and a model built on historical structure has no special ability to foresee a genuinely new event.

Past performance and any illustrative figures used to explain how a strategy works do not guarantee future results. A pattern that has repeated consistently can stop repeating without warning, and a risk limit reduces the size of a loss without preventing one from occurring. Nothing on this website is investment advice, and the decision to deploy a strategy, choose a market or set a risk limit remains the trader’s own.

We built McCortex OI to remove manual friction from execution and to apply consistent risk rules to every order, not to remove uncertainty from trading itself. Understanding that distinction is part of using the platform responsibly, and it is why this page describes a process rather than a promise.