Factor research
Construct and evaluate signals such as value, quality and momentum. Examine portfolio formation, exposures and the effect of implementation choices.
QUANTITATIVE RESEARCH
An investment idea.A testable hypothesis.
Examine market signals with a research process that makes the data, assumptions and results explicit. From factor construction to out-of-sample evaluation, the aim is to understand what holds up.
For investment teams, analysts and researchers with a strategy or signal to investigate.
SCOPE & STARTING PRICE
An initial study of one investment idea.
A focused research question using an agreed, available dataset. Suitable for evaluating an idea before commissioning a larger research programme.
From €1,500/ project + VAT
Request a research proposalOpens an email draft. Tell us your goal, available data and preferred timing.
What is included
- One defined hypothesis and a baseline backtest
- Transaction-cost assumptions and selected robustness checks
- Research code and a report covering results and limitations
Indicative starting price, excluding VAT. Final scope, price and timing are confirmed in a written proposal. Data licences, model/API usage, hosting and ongoing support are priced separately when needed.
RESEARCH AREAS
From financial data to strategy evaluation.
Financial machine learning
Build features and compare predictive models using time-aware validation. Keep model selection and evaluation tied to the question being tested.
Backtesting & robustness
Study returns alongside transaction costs, turnover, drawdowns and parameter sensitivity. Document data limitations and the assumptions behind each result.
PUBLIC RESEARCH & CODE
See the approach in practice.
PUBLIC CODE / intradayzoo
Testing signals at intraday frequency.
- Research question
- How do factor signals behave across intraday, overnight and stock-level prediction tasks?
- Approach
- An R research pipeline with time-ordered model evaluation and strategy tests that account for trading costs.
- Available to inspect
- Data preparation, model estimation and strategy-evaluation code. This is a public research project, not a client performance claim.
PUBLIC CODE / quant-ai-agent
From a written strategy to a research run.
- Research question
- Can a written strategy be turned into a structured, repeatable research workflow?
- Approach
- Parse the specification, check data availability, generate a backtest and test parameter variations.
- Available to inspect
- The Python workflow and documentation for generating baseline and robustness reports.
THE RESEARCH PROCESS
Every result needs a method.
- 01
Define the question
Specify the signal, investment universe, timing and portfolio rules.
- 02
Check the data
Review availability, coverage and timestamps before running the experiment.
- 03
Test the idea
Build a reproducible baseline and examine alternative parameters and periods.
- 04
Document the findings
Bring the code, assumptions, limitations and results into a research report.
BEFORE YOU START
A few practical answers.
Do we need to supply the data?
Start by describing the strategy and the data you already have. Availability, history, quality and usage rights are checked before the research scope is agreed.
What if the strategy does not hold up?
A weak or unstable result is still a research finding. The report explains the evidence and limitations so you can decide whether further work is justified.
What changes the price?
Additional hypotheses, new data pipelines, high-frequency execution modelling and larger model comparisons expand the scope. These are specified in the proposal before work starts.
Contentio d.o.o. · Croatia
Start with the question you need answered.
Send a short brief. We can then define the scope, required inputs and a written proposal.
Discuss a research project mislav.sagovac@contentio.biz
Opens an email draft. Tell us your goal, available data and preferred timing.