AI-supported risk management
Teilprägur analyzes market data in real time and activates an adaptive stop-loss mechanism based on calculated probabilities to limit significant price losses before they impact the entire portfolio.
The system is based on predictive analytics: It does not predict price trends, but rather continuously calculates the probabilities of market movements and derives action thresholds from them.
Price, volume and volatility data is continuously recorded and compared with historical patterns in order to identify deviations at an early stage.
From the data streams, the model calculates the probability of a sustainable trend reversal as opposed to short-term fluctuations.
If the calculated probability exceeds a defined threshold, the system automatically adjusts or triggers the stop loss level.
The following overview compares typical key figures of standard market behavior with those of systematically optimized control.
| Key figure | Standard market behavior | Partial embossing optimization |
|---|---|---|
| Maximum drawdown | Often only visible after entry | Limited by upstream thresholds |
| Volatility index | Reaction usually delayed | Continuously integrated into the model calculation |
| Response time | Depends on manual observation | Automated triggering in real time |
| Basis for decision | Often emotional or delayed | Data-based and comprehensibly documented |
Teilprägur was designed for investors who manage their capital independently and value structured, verifiable decision-making principles. The system does not make investment decisions, but rather provides data-based decision support.
Every adjustment to the security mechanisms is based on documented calculation steps that can be viewed in the user account. Control over the portfolio remains entirely with the person investing it.
The technical architecture is designed for reliability — both in data processing and in protecting sensitive account information.
The hedging mechanism dynamically adapts to the calculated market volatility rather than remaining at a rigid, fixed percentage level.
All account data and transaction information is transmitted and stored according to current encryption standards.
Each automated recommendation can be viewed, adjusted or overridden by the person using it.
Data processing runs on mirrored server environments to avoid downtime during volatile market phases.
Every update to the underlying calculation logic is documented and can be traced over time.
Access to the account takes place via multi-factor authentication to prevent unauthorized access.
Data protection notice: Account-related data is processed exclusively to provide the analysis functions and is not passed on to third parties for advertising purposes.
The following scenarios illustrate how the system distinguishes between short-term noise and actual trend changes.
In the event of an abrupt price decline, the model registers the accelerated downward movement using volume and volatility data and calculates the probability of a continued movement.
In phases without a clear direction, the price fluctuates within a narrow range. The model recognizes this pattern as statistical noise with no structural trend change.
The following answers cover typical questions about data sources, model updating, and calculation logic.
The system processes price, volume and volatility data from established market data providers. These are continuously compared to reduce distortions from individual sources.
The calculation logic is checked at regular intervals based on current market data and adjusted if necessary. Every change is documented and stored in the user account in a traceable manner.
The threshold value results from the calculated probability of a sustainable trend reversal in combination with the current volatility of the respective instrument. The process is deterministic and can be traced individually for each position.
Yes. Any adjustment suggested by the system can be viewed, changed or completely disabled. The final decision lies with the user.
The model evaluates the persistence of a movement across multiple time windows. Short-term fluctuations without an accompanying volume change are classified as noise and do not trigger an adjustment.
If you have further technical questions, our analysis team is available to you via the channels specified in the contact area.
Get a non-binding overview of how risk adjustment works based on your preferred asset class. There are no obligations.