Test the idea.
Understand the limits.

CMAC examines quantitative methods for investigating market behavior, including statistical analysis, adaptive algorithms, recursive machine learning, and AI model design. Research hypotheses, architectural designs, and model candidates are distinguished from methods supported by completed evaluation.

Research priorities

Four questions for
investment research.

These priorities describe areas of inquiry. They are not announcements of funded projects, completed studies, or validated investment outcomes.

01 / Signal durability

Does the relationship persist?

Examine whether an apparent signal extends beyond the sample that suggested it, and whether an economic explanation survives alternative tests.

02 / Risk and change

What changes the conclusion?

Study sensitivity to market conditions, volatility, liquidity, concentration, and the assumptions on which a result depends.

03 / Trading frictions

Does the opportunity survive costs?

Distinguish theoretical results from outcomes that account for fees, spreads, slippage, capacity, and practical execution constraints.

04 / Model reliability

When should a model not be trusted?

Investigate data integrity, reproducibility, overfitting, model failure, and the limits of algorithmic decision support.

Technical research and engineering

Research and development disciplines.

CMAC's technical program encompasses four related disciplines that support quantitative research and the evaluation of algorithmic methods.

Adaptive algorithm development.

CMAC investigates and designs algorithms intended to adapt to new evidence and changing conditions within defined constraints. Any claimed adaptive advantage must be established through comparative testing rather than assumed from the design.

Recursive machine learning.

Our research program includes iterative and recursively structured machine-learning methods, including hypothesis development, candidate-model generation, comparative evaluation, and controlled refinement. Model changes remain research candidates until supported by appropriate evidence.

AI model and system development.

CMAC's architecture and development work addresses AI model design, computational research workflows, evidence handling, model integration, and controlled experimentation. Architectural designs and development work do not by themselves establish production deployment or active investment capability.

Validation and assurance.

CMAC's research methodology incorporates requirements for reproducibility, robustness evaluation, out-of-sample testing, and failure analysis. Internal validation and independent assurance are separate responsibilities. Completion of either must be supported by attributable evidence.

Principles for
evaluating an idea.

Define the comparison.
State the hypothesis, relevant baseline, and evidence that would count against the idea.
Keep the test honest.
Use information available at the relevant time. Separate exploration from validation and preserve unfavorable findings.
Measure what matters.
Consider uncertainty, trading frictions, downside exposure, and the difference between statistical and economic significance.
Separate evidence from action.
A research result is not, by itself, a reason or authorization to commit capital.

Public research status

The capabilities described on this page include research directions, architectural work, and development objectives. They do not represent claims of deployed investment algorithms, verified predictive accuracy, completed independent assurance, or live capital execution. This website is informational and does not offer personalized investment advice or solicit client capital. Proprietary code, strategy parameters, confidential datasets, and protected research details are not disclosed.

A question worth testing?

Technical and research correspondence is welcome through the corporate contact channel.

Research correspondence

Test the idea.Understand the limits.

Four questions forinvestment research.