Market intelligence
Research tools for comparing trend and volatility across market conditions. The goal is to make the inputs to a trading decision easier to inspect.
PlannedAI & algorithmic trading research
Trading software.
From research to execution.
Island Algorithms is developing automated trading systems that combine explicit trading rules with machine learning. Our first project examines when market conditions should allow a strategy to trade—and when it should stay out.
Python · Research & model trainingMQL5 · Planned execution engineMT5 · Future demo testing
Testing & performance
The software comes first.
This space will report our own system’s results once development and demo testing begin.
Pre-demo · Design stage
No trading account connected yet.
We’ll create the MT5 demo after the initial AI trading system is built. There are no live results or performance claims to show at this stage.
What future reporting will include
- Account status
- Demo or real-money account, with the source and last update time.
- Performance
- Balance, equity, closed trades and drawdown, with clear definitions.
- Research context
- Strategy version, test period, assumptions and known limitations.
Real-money results will be kept distinct from demo testing. Figures will appear only after a reporting source is connected.
Research
Our first system
A defined question, a simple baseline and an explicit test.
Design stage · No published results yet
Can a market-condition filter improve a rule-based trading strategy?
The proposed system uses machine learning to classify conditions such as trend, range and volatility. The trading rules then decide whether an entry is allowed. The model’s job is to filter opportunities; position sizing and trade execution remain governed by explicit rules.
- Execution
- Planned MQL5 trade engine for MetaTrader 5, with Python used for research and model training.
- Comparison
- The same baseline strategy with and without the model, evaluated over the same periods and costs.
- Evaluation
- Unseen test data, drawdown, trading costs, stability across market conditions and demo forward testing.
- Still to define
- The currency pair, timeframe, entry rules, training labels and risk limits before testing begins.
Development
Beyond execution
Tools we intend to develop alongside the trading system.
Trading analytics
Analysis of strategy behaviour: drawdown, costs, trade distribution and performance across different test periods.
PlannedCustom algorithm work
A future service for translating clearly specified trading rules into software, with an agreed testing process and documented limitations.
Future service · Not yet taking commissionsAbout Island Algorithms
Island Algorithms is an independent project moving from a single trading robot towards broader trading software and research. The focus is AI-assisted systems, market intelligence and analytics. Development is at an early stage; this site will grow with the work.