Yoga Den analyzes market data in real time and tests each strategy against multiple historical periods before deploying it. No intuition, but modeled decision-making.
Three mechanisms form the basis of every recommendation the platform generates.
The system recognizes patterns in historical price and volume data and translates them into scenarios with an associated probability. Each scenario is continuously adjusted based on new market input.
Volatility thresholds and downside scenarios are calculated per position. When a set threshold is exceeded, exposure is automatically reduced, without manual intervention.
Analyzes run in parallel across multiple asset classes. New datasets are added without interrupting or slowing down existing model processes.
Each strategy goes through four set steps before it is released for use.
Structural and alternative data sources are merged and normalized.
Algorithms are trained on historical series and multiple market regimes.
Each strategy is backtested on periods outside the training data.
Only strategies that pass the backtest go to production.
Backtesting tests a strategy against historical market data that falls outside the training period. This makes it visible how a model would have behaved during previous corrections, interest rate changes and volatility peaks.
Strategies that deviate from the pre-set risk limits under these circumstances will be adjusted or rejected before going live. This process is repeated periodically on new dates.
The same model supports three different decision issues.
For individual investors, the model continuously calculates the ratio between expected return and risk per position. In the event of deviations, the user receives a rebalancing proposal with underlying motivation.
The system compares current market movements with expected patterns based on historical data. Statistically significant deviations are flagged for review, not automatically performed.
| Signal | Deviation | Status |
|---|---|---|
| Volatility index | +2.3σ | Assessment |
| Volume pattern | +1.1σ | Ignored |
| Correlation break | +3.0σ | Assessment |
Business operators use the same data structure to inform budget and capacity choices. The model calculates scenarios based on demand fluctuations and cost price developments.
| Segment | Assignment | Trend |
|---|---|---|
| Operating capital | 42% | Stable |
| Growth initiatives | 31% | Rising |
| Risk buffer | 27% | Stable |
The view below shows the structure of the dashboard in which key figures, signals and historical series come together on one screen.
| Data set | Series | Status |
|---|---|---|
| Stock indices (EU/US) | 10 yrs | Active |
| Fixed income securities | 8 yrs | Active |
| Alternative data sources | 4 yrs | In validation |
Answers focused on integration, security and model operation.
All data traffic between user and platform is encrypted. Market data and model output are stored separately from account data, so access to analytics does not provide insight into personal user information.
Models are recalculated every 15 minutes based on the most recent market data. In the event of significant market movements, an interim recalculation can be triggered, regardless of the fixed interval.
The platform offers API access for reading signals and portfolio data. Export to common spreadsheet and BI formats is possible for those who want to combine analyzes outside the dashboard with internal data.
After registration you will receive access to the backtest results and an overview of the active models before you decide on further implementation.