Technical architecture of the analysis engine
Each recommendation goes through four distinct stages, from raw data collection to historical validation, before being presented to the user.
Data ingestion
Continuous collection of market feeds, macroeconomic data and price series from public and private financial sources.
Standardization
Raw data is standardized, outliers filtered, and time series aligned to a common base.
Predictive modeling
Statistical models identify correlations between assets and estimate likely market scenarios.
Validation by backtesting
Each strategy is simulated over historical periods distinct from the training phase before being made available.
Analysis tools for diversification
Each module addresses a specific dimension of the portfolio — allocation, risk, correlation — in order to produce quantified recommendations.
Dynamic allocation
Adjusts the portfolio allocation based on the volatility measured on the assets held.
Systemic risk scoring
Assigns each position an exposure rating to market, credit and liquidity shocks.
Scenario simulation
Projects the impact of macroeconomic variations on the overall value and risk of the portfolio.
Correlation detection
Identifies highly correlated assets to limit unintended concentration of risk.
Backtesting reports
Documents the historical performance of each strategy tested, period by period.
Rebalance Alerts
Reports deviations between the actual allocation and the user-defined target allocation.
Risk is not an abstract variable
The engine segments the portfolio into homogeneous risk classes, measures the cross-correlation between positions, then calculates a net exposure to market, credit and liquidity risk.
The trigger thresholds are adjustable by the user and applied identically to each recalculation, without intermediate manual intervention.
Analyze patternsConcrete application for active profiles
Three representative profiles of users seeking to diversify their income in addition to a main activity.
Diversification of employee savings
Objective: to distribute savings built up over several years beyond the products offered by the employer, with monthly rather than daily monitoring.
Testing new asset classes
Objective: evaluate, via backtesting, the interest of integrating commodities and currencies into a portfolio hitherto concentrated on equities.
Cross-validation of internal models
Objective: to compare personal management hypotheses with the backtesting results produced by the engine, before any allocation decision.
Data sources and validation method
The most frequently asked questions concern the origin of the data and the reliability of backtesting.
Where does the data used by the platform come from?
Data is aggregated from recognized financial data providers and public markets, with continuous updating of price series and macroeconomic indicators.
How is backtesting carried out?
Each strategy is simulated over a historical period distinct from that used for training the model, in order to limit the risk of overfitting on already known data.
Does past performance guarantee future results?
No. Backtesting measures the robustness of a strategy over market cycles that have already occurred. It does not constitute a guarantee of future performance and should not be construed as such.
Which asset classes are covered by the analysis?
Stocks, bonds, ETFs, currencies and commodities. Coverage depends on the data sources available for each instrument.
Do you need advanced financial knowledge to use the platform?
A basic understanding of risk and allocation is recommended. The reports are designed to remain readable without portfolio manager training.
Go from an intuitive decision to a quantified decision
The Mont Rendoire terminal provides access to the allocation, risk scoring and backtesting modules described above, with monthly monitoring of your positions.
Access the terminal