AI-powered quantitative analysis based on verifiable historical data.
Lumiant Quant combines predictive modeling and rigorous backtesting to transform your market hypotheses into documented and measurable decisions.
Lumiant Quant combines predictive modeling and rigorous backtesting to transform your market hypotheses into documented and measurable decisions.
Lumiant Quant was designed for investors who want to confront their market intuitions with solid statistical models, rather than substitute them. Each recommendation is based on traceable historical data and explicit risk parameters.
We are not trying to automate your judgment, but to equip it. The signals produced by our models are accompanied by the assumptions behind them, so you can assess their relevance to your own strategy.
Before a signal is presented to you, it goes through a defined, repeatable and documented process. No step is left to the sole discretion of the algorithm.
Market flows, trading volumes and macroeconomic indicators are continuously collected and normalized, so that comparisons between assets are on a consistent basis.
The models identify statistical regularities over varying historical windows and then estimate the likelihood of these patterns repeating under comparable market conditions.
Each recommendation is weighted by volatility and correlation constraints, to limit excessive exposure to a single scenario.
Three complementary bricks, designed to integrate into an already existing decision flow rather than to replace it.
Intra-day volatility makes fixed analyzes obsolete. Our feeds are continuously updated, so that your decisions are based on the real state of the market and not on a photograph dated a few hours ago.
Each strategy may be exposed to several historical market cycles before being applied. This step helps identify conditions under which an approach tends to underperform.
Drawdown thresholds and correlation limits between positions are defined upstream, so that performance optimization is never done to the detriment of uncontrolled exposure.
Rather than relying on testimonials, we publish how our models perform over defined historical periods.
Cumulative performance curve of a typical strategy over a multi-year backtest period, before fees and execution slippage.
The indicators below summarize the behavior of a representative strategy over the back-testing horizon considered. They are calculated on historical data and do not guarantee future reproduction of the same results.
The performances presented come from back-testing on historical data. They constitute neither a guarantee nor a prediction of future performance. Actual results depend on market conditions, applicable fees and quality of execution.
Our tools are designed to fit into already structured decision-making processes, and not to replace them with a black box.
Access modeling methods usually reserved for professional desks, with sufficient documentation to justify each hypothesis made.
Integrate our signals upstream of your own execution rules, to enrich a system already in place rather than replacing it entirely.
Identify unclear correlations between asset classes, in order to limit unintentional concentration on the same risk factor.
An investment decision deserves to be supported by verifiable data rather than intuition alone. Lumiant Quant provides you with the models and history necessary to build this database.
Getting started takes just a few steps: creating your risk profile, selecting the asset classes monitored, then accessing your analysis dashboard.