A systematic trading process requires more than a clearly defined strategy. It is equally important that market data is processed efficiently, trading rules are applied consistently and potential positions are assessed in a traceable way.
With our new scanner we have taken a further step toward automating these processes. The software analyzes the US stock market against predefined criteria, scores suitable candidates with our quantitative model Ridge 75 and calculates the resulting position sizes. The aim is to standardize the operational implementation of our strategy and reduce sources of manual error.
From market data analysis to trading decision
The scanner maps our strategy's rules-based selection process onto a standardized workflow.
- Systematic market screening: The software examines the US stock market against our defined trading parameters and identifies securities that meet the required criteria.
- Quantitative model scoring: The pre-selected candidates are scored with our Ridge 75 model. Further information on the methodology can be found in our article on the Ridge 75 model.
- Risk-based position calculation: For each candidate, the scanner determines the intended share count based on account size, the defined risk budget and the current price level.
- Pre-trade updates: The calculations are updated with the latest available price data until shortly before the intended trading time.
- Transparent exclusion analysis: In addition to the selected candidates, the scanner also documents excluded securities and the reason for each exclusion. This makes the entire selection process traceable.
- Reproducible documentation: Every scan is saved. The results can be reviewed afterward and reconciled with the underlying market data.
Data integrity and operational safety
In a quantitative trading system, reliable processing of market data is a fundamental prerequisite. Faulty or incomplete data must not enter the decision process unnoticed.
We have therefore implemented several safeguards:
- Validation of input data: Missing, faulty or unusable data lead to the affected candidate being flagged. It is not proposed as a regular trade.
- Visible system warnings: If the reliability of the model is impaired, a clear warning signal indicates it. Delayed or incomplete market data are flagged separately.
- Separation of analysis and execution: The scanner generates trade proposals but does not place orders. The trading decision and order execution remain separate from it.
- Controlled failure behavior: Problems with data sources, corrupted data sets or access errors are meant to result in a clear error message rather than an uncontrolled abort of the entire process.
These measures increase the technical robustness of the system. They cannot, however, guarantee that all data is free of errors or that operation is uninterrupted under all conditions.
Technical validation
Before deployment we subjected the scanner to extensive automated testing. The focus was not only on the regular workflow but in particular on behavior with faulty input data and technical disruptions.
The validation comprised:
- Around 1,000 automated tests, including targeted fault simulations with invalid prices, missing data and simulated outages.
- Independent reproduction of the calculations: The calculation logic was implemented separately on historical data. The results were then compared with those of the scanner.
- Reconciliation of historical results: The historical trades examined and the model results matched between the two implementations.
The independent reconciliation reduces the risk that errors in the original implementation go undetected. It is not, however, proof of future profitability or of freedom from error in every conceivable market situation.
Operational benefit
The scanner does not change the underlying trading strategy. It standardizes its operational application.
Recurring tasks such as market screening, candidate selection, position calculation and documentation are brought together in a single process. This reduces manual effort while improving the traceability of the individual steps.
The time gained can be used more specifically for monitoring the market environment, assessing exceptional situations and further developing our quantitative models.
Limits and transparency
Despite automation, technical and market-related risks remain.
- Strategic confidentiality: We do not publish the complete trading rules, model parameters and specific thresholds, as they are part of our proprietary trading methodology.
- Dependence on market data: The quality of the analysis depends on the availability and reliability of the data source used. Warning mechanisms can make data problems visible but cannot rule out faulty source data altogether.
- Ongoing monitoring: Technical validation is not a one-off conclusion. We monitor operation and adjust the software as needed.
- No elimination of market risk: Consistent application of rules does not protect against adverse price movements, execution risks or losses. Even a technically correct system can cause substantial losses.
The scanner is an operational tool supporting our quantitative trading strategy. It improves the standardization and verifiability of the process but is neither a guarantee of investment results nor a reduction of all risks.
This article is for information purposes only and constitutes neither investment advice nor an invitation to buy or sell financial instruments. Our Terms & Conditions apply.