Clear boundaries. More useful evidence.
A technical portfolio should show what was built without quietly upgrading a simulation into a track record. Here is how to read the material on this site.
Different records.
Different claims.
The label describes the source of the evidence. It is not a ranking of how impressive the project is.
- Research profile
- A description of a question, implementation area or archived project. It makes no current performance or deployment claim.
- Archived replay
- A simulation of a specified configuration on historical inputs. It is not a broker-observed return or an independent validation.
- Paper / shadow
- A forward observation without the same execution or capital conditions as a real-money account. Any example needs its own dated source.
- Live record
- Broker-observed activity requiring reconciliation, cash-flow treatment and a clear measurement period. This website does not publish a verified live track record.
- Independently verified
- A claim requiring a named independent reviewer and an explicit scope. No such status is claimed for the strategy examples on this website.
The questions a
risk review asks.
-
What exposure is being taken?
Consider aggregate factors, concentration, leverage and correlated instruments—not just a position-by-position view.
-
What can the execution model miss?
Review liquidity, gaps, spread, slippage, financing and the difference between a quoted price and a possible fill.
-
What happens when inputs fail?
Specify missing-data behaviour and stop conditions. An absent observation must not silently become a permissive decision.
-
What is the authoritative record?
Separate simulated NAV, paper observations and broker statements. Deposits and withdrawals are not strategy returns.
Private methods.
Honest descriptions.
Monolith withholds detailed execution rules, sensitive parameters and infrastructure information where publishing them would expose proprietary work. Withholding information is not a reason to invent returns or imply an audit that never happened.
Public case studies focus on the research question, the technical choices and the limits of the evidence. A source-reported metric is identified as such; a recomputed metric states what was recomputed.
Historical notebook entries are preserved as dated research writing. Their deployment and performance statements are not maintained as current status. Known high-risk claims are flagged at the article level.
Research.
Not an offering.
Monolith Research is an independent research project, not an incorporated investment firm. It does not manage client money, offer investment products or provide investment advice. This website is for research, technology, employment and collaboration.
For research enquiries ↗