The system combines technical indicators, market data, statistical models, and machine-learning components to generate research signals. Signals are not predictions of guaranteed future prices and should not be interpreted as investment advice.
What the system does
In order, the system:
- Collects market data across crypto, ETFs and fixed income and aligns it to common timeframes.
- Computes standard indicators — MACD, RSI, ATR, ADX and others — on more than one timeframe.
- Interprets them together, with cross-asset context, and states in plain language where they agree and where they conflict.
- Grades a signal when the evidence supports one, and produces none when it does not.
- Checks risk against volatility, liquidity and existing exposure before a signal is sized.
- Records and reviews every trade taken, including the indicator state at entry.
What the system does not do
- It does not predict prices or guarantee outcomes.
- It does not know your finances, objectives or tolerance for loss, and so cannot advise you.
- It does not see the future: unscheduled news, policy changes and regime shifts are outside its inputs until they have happened.
- It does not make the final decision. A person does.
How signals are graded
A signal carries a direction and a tier. The tier is a grade of signal confidence: how strongly the inputs agreed with one another when the signal was produced.
Signal confidence is not probability of profit. A top-tier signal can lose, and the one trade in our public journal is an example: it was Tier A, and it closed at a loss.
When inputs disagree
Different timeframes and different markets frequently point in different directions. The system reports this as “conflicting signals”. It is not averaged away or resolved by picking whichever input supports a trade.
How results are measured
Each closed trade is recorded with its entry, exit, holding period, leverage and return, together with its maximum favourable excursion (MFE) and maximum adverse excursion (MAE) — how far price moved for and against the position while it was open. These show the path of a trade, not only where it ended.
All trades are published, in order, with losses shown in the same detail as gains. We do not present back-tested or hypothetical results as if they were live.
Known limitations
- Indicators are computed from past prices and lag the market by construction.
- Models are fitted to historical data and can fail when conditions change.
- Data can be late, wrong or missing.
- A short track record says very little. Ours is short.