Evidence first. Assumptions made visible.

Four Quanta uses AI-assisted research to examine companies, compare investment opportunities and document decisions in separate paper portfolios.

How to read the research

  1. Business and evidence. Reports draw on dated filings, company releases and relevant external sources. Follow the source links and distinguish reported results from forecasts.
  2. Valuation and challenge. Scenarios combine assumptions about earnings, cash generation and valuation. A weighted outcome is an estimate, not a promised or statistically established return.
  3. Portfolio judgment. A score alone does not trigger a purchase. Allocation decisions consider other opportunities, cash, risk and existing exposures.
  4. Monitoring. Reviews revisit the evidence. Qualitative triggers require interpretation; dated decisions do not establish current holdings.

AI, accountability and corrections

AI systems assist with research, opposing arguments and editorial presentation. They can make mistakes. Four Quanta is responsible for the published presentation; automated validation does not establish that every claim is correct. Reports identify sources and dates so readers can inspect the evidence. Report an error or ask about the research.

How results should be interpreted

All portfolios use virtual capital. Confirmed paper fills, estimated costs and historical simulations are distinct. Distribution treatment and other limitations are shown on each dashboard. A short record does not demonstrate investment skill, and a portfolio stop cannot guarantee a maximum loss.

Research updates do not silently replace historical decisions. Editorial clarifications are labelled. Public pages show selected material; invited access provides the fuller research context.

Read the sample report →

Publication purpose and responsibility

Four Quanta publishes evidence and observations from experimental paper portfolios using virtual capital. Readers observe the experiment; the publication does not accept their investment capital or operate their accounts. New access and subscriptions are paused.

Four Quanta is responsible for the publication. AI systems assist research and writing; their outputs can be incomplete or wrong. Sources, research dates, assumptions, contrary evidence and material limitations should accompany each assessment. Report errors through the contact page.

Editorial record and conflicts

Reader editions omit internal buying instructions, entry ceilings, pending orders and price-trigger alerts. Original research remains in the private audit archive. An omission is not a change to the underlying strategy or evidence, and a proposal does not establish that a paper trade occurred.

Relevant publisher and contributor holdings and commercial relationships must be recorded and disclosed before a report is cleared for publication. A paper holding is not a disclosure of personal ownership. Existing reports have not yet been verified against a complete conflicts register; no claim of independence or absence of interests is made.

How to interpret results

Paper performance is not an achievable-return promise. Simulated execution, trading costs, taxes, currency movements and missing income can affect comparisons. Scenario weights are subjective assumptions, not measured probabilities; losses can exceed a bear scenario. A score is not a probability of profit. These publication changes do not establish FCA approval.