The Profits Circle applies AI-powered backtesting to decades of market data, converting short-term noise into quantified, risk-weighted strategies built for household financial planning.
Access PlatformThree technical pillars underpin every recommendation the platform generates. Each is designed to be auditable rather than asserted.
Portfolio exposure is reassessed against incoming market data on an ongoing basis, flagging deviations from a household's defined risk tolerance before they compound.
Every strategy is tested against historical market cycles, including periods of contraction, before it is presented as a recommendation rather than after the fact.
Underlying models process data across asset classes and timeframes simultaneously, maintaining consistent analysis as the scope of a household's holdings grows.
The Profits Circle was built to give middle-income families access to the same class of quantified, backtested analysis traditionally reserved for institutional portfolios.
Rather than issuing generic market commentary, the platform processes each account's specific time horizon, liquidity requirements, and risk tolerance against historical precedent, producing a defined set of ranked strategy options.
More on our approach →
The platform follows a fixed four-stage pipeline. Each stage narrows the dataset and reduces the influence of short-term emotional reaction on the final output.
Market, macroeconomic, and account-level data are collected and normalised into a common structure before analysis begins.
Statistical models identify recurring structures in price behaviour and correlate them against historical outcomes.
Identified patterns are weighted against a household's stated time horizon, liquidity needs, and risk tolerance.
The model produces a ranked set of strategy options, each annotated with its backtested performance under comparable historical conditions.
| Pipeline stage | Primary function | Bias reduced |
|---|---|---|
| Ingestion | Data normalisation | Manual entry error |
| Pattern recognition | Statistical correlation | Recency bias |
| Risk weighting | Constraint matching | Emotional overexposure |
| Strategy output | Ranked recommendation | Decision paralysis |
The examples below illustrate how the pipeline distinguishes a genuine signal from ordinary market noise in two common household scenarios.
This reflects the behaviour of diversified portfolios during past periods of prolonged contraction, where staged reallocation reduced drawdown exposure relative to static allocations held unchanged throughout.
Historical data shows fixed-horizon funds are disproportionately affected by short-term noise close to the withdrawal date; the pattern recognition stage is designed to filter this noise from genuine signal.
Answers below are factual and limited to how the platform actually operates.
Account and personal data are encrypted in transit and at rest, and are not shared with third parties for marketing purposes. Internal access is limited to the functions required to generate a recommendation.
The models are trained on historical market pricing, macroeconomic indicators, and anonymised portfolio structures. No single household's data is used to train models relied upon by other accounts.
A backtested strategy has been applied retrospectively to historical market data to measure how it would have performed. Backtested performance does not guarantee future results and is provided for comparative analysis only.
Models are reviewed and recalibrated as new market data becomes available. Frequency depends on market conditions rather than a fixed calendar schedule.
All recommendations are generated for analytical purposes and should be considered alongside independent financial advice. Read the full risk disclosure →