PILLAR II · SECOND GATE
Deep Discovery Research
The second of ABCD Fund's three investment pillars: where a screened candidate is tested against evidence until it earns conviction — or is rejected.
Deep Discovery does not ask whether a candidate looks cheap. It asks whether the case for buying it survives a deliberate attempt to prove it wrong.
- Second of Three Pillars
- Twelve Research Domains
- Evidence-Ranked, Not Assumed
Why This Pillar Exists
Where a Screened Candidate Becomes an Evidence-Based Thesis
Opportunity Screening identifies candidates worth investigating. Deep Discovery Research is where that candidate is tested: business model, management, financial position, competitive standing, valuation, and downside are each built on evidence, not assumption, until a thesis either earns conviction or is rejected.
Most screened candidates are rejected at this stage. That discipline is intentional — only ideas that survive sustained scrutiny, including a deliberate search for evidence that would disprove them, are eligible to advance to Valuation Realisation.
The Evidence Map
Twelve Domains, One Evidence Base
No single domain confirms a thesis on its own. Each of the twelve domains below is investigated in depth, and it is the evidence produced across all of them together — not one compelling data point — that a thesis for Eclipse or Nova is finally built on.
Business Fundamentals
Business Model
How the business actually earns, and whether that mechanism holds once the conditions that created the entry price change.
Management and Governance
Capital allocation record through prior cycles, incentive alignment, and ownership structure — tested against what management has done, not what it says.
Industry and Competition
Competitive position and industry structure through the dislocation: whether the gap to peers strengthens, weakens, or holds steady.
Financial Evidence
Financial Statements and Cash Flow
Reported results tested against cash generation, not reported earnings alone.
Balance Sheet and Liquidity
Obligations, debt maturity, and the capacity to service them through the stress period without a dilutive capital raise.
Valuation
The measurable gap between price and an estimate of intrinsic value, built from the evidence in every other domain — not assumed ahead of it.
Forward-Looking Risk
Catalysts
The specific event or milestone identified as the mechanism for recovery, and the timeframe attached to it.
Downside Scenarios
The bounded loss case if the thesis does not play out, tested independently of the upside case.
Scenario Analysis
Bull, base, and bear outcomes built from the same evidence base, so conviction reflects a range of results, not a single forecast.
Verification and Breadth
Legal and Regulatory Evidence
Filings, disclosures, and any outstanding legal or regulatory matter that could block a clean position regardless of valuation.
Alternative Data
Evidence outside standard filings and reports, used to corroborate or challenge the fundamental picture — not to substitute for it.
Local-Language Sources
Primary-market disclosures and reporting read in their original language, not filtered through translated summaries alone.
Weighing What We Find
Evidence Hierarchy: Not Every Fact Carries the Same Weight
Every domain above produces evidence of different reliability. Deep Discovery Research ranks it rather than treating it as uniform, so conviction tracks the strength of the underlying record.
-
Tier I
Primary Evidence
Evidence checked directly against its source. This is the evidence a thesis is built on.
- Audited Financials
- Regulatory Filings
- Legal Records
- Original-Language Disclosures
-
Tier II
Supporting Evidence
Evidence that corroborates the primary record. Directional, and expected to agree with Tier I — not a substitute for it.
- Management Commentary
- Competitor Analysis
- Alternative Data
- Scenario Modelling
-
Tier III
Lower-Confidence Opinion
Evidence not yet checked against a primary source. Tracked, but never sufficient on its own to support a thesis.
- Analyst Commentary
- Sell-Side Sentiment
- Secondary Reporting
- Unverified Translations
When the Evidence Disagrees
Contradiction Does Not Get Filtered Out
A thesis is expected to be tested against evidence that could disprove it. What a piece of contradictory evidence does to a thesis depends on what it contradicts and how well it is confirmed.
Weaken
Confidence Lowered
A single contradictory data point, on its own, does not overturn a thesis. It lowers confidence in the affected input and triggers closer review of that specific claim.
Revise
Thesis Rebuilt
Confirmed contradictory evidence that changes a material assumption sends the affected domain back for re-testing, and the thesis is rebuilt from the corrected input.
Pause
Held for Resolution
An unresolved conflict between primary evidence and the working thesis holds the candidate out of Valuation Realisation until the conflict is resolved.
Stop
Thesis Ends
Verified evidence that contradicts a required qualifying factor — recoverability, a bounded downside, or a clean legal position — ends the thesis regardless of remaining upside.
Accountability
Data Systems Assist Research. People Verify and Decide.
Deep Discovery Research draws on fundamental analysis, alternative data, and quantitative or scenario-based tools, including AI-assisted research aids that help organise and surface information across the domains above. These tools assist research and monitoring — they do not verify material facts or make the investment decision.
Every material fact underlying a thesis — financial, legal, or evidentiary — is checked against primary sources by the investment team before a candidate can advance. Evidence that contradicts the thesis is reviewed, not filtered out. The relevant Eclipse or Nova investment lead makes the final call on whether a thesis is confirmed.
Where Human Review Applies
Applied Across Two Mandates
How Deep Discovery Applies to Eclipse and Nova
Both of ABCD Fund's intended launch strategies apply the same evidence base and the same evidence hierarchy at this stage. What differs is the emphasis within the research.
Crisis Upside · Concentrated
Eclipse Fund
Every Eclipse equity position that clears Opportunity Screening receives a full Deep Discovery pass, drawing on company visits, channel checks, and alternative data alongside the twelve research domains, before a thesis can advance.
Absolute Return · Diversified
Nova Fund
Nova's equity sleeve applies the same evidence hierarchy, combining fundamental, macro, sentiment, and quantitative evidence before a position is eligible for Valuation Realisation.
Next Stage
Moving to Valuation Realisation
A thesis that survives Deep Discovery Research — evidenced, tested against contradiction, and confirmed by the investment team — advances to Pillar III: Valuation Realisation, where entry price, position sizing, and the exit protocol are set.
Investor Relations
See How Deep Discovery Builds Conviction
Request an investor briefing to see how the evidence hierarchy and contradiction review apply to current candidates across our funds.
Request Briefing