Find
Identify a market pattern, anomaly, or question worth investigating.
MARKET DATA HUNTER
MARKET IDEAS. TESTED.
RESEARCH STANDARD
An idea isnot a result.
Market Data Hunter separates finding an interesting pattern from deciding what that pattern actually tells us.
The process is designed to make it progressively harder for a hypothesis to survive by chance, hindsight, or changing rules.
Evidence before conclusions
Rules before outcomes.
THE RESEARCH LIFECYCLE
Start curious.
Become skeptical.
Every investigation can begin with an observation. What matters is what happens next: define it, test it, challenge it, freeze it, and let future evidence answer the remaining question.
Define
Turn the idea into explicit data, signal, ranking, portfolio, and evaluation rules.
Test
Measure the historical relationship and document both supportive and adverse evidence.
Challenge
Ask whether chance, instability, or research choices could explain the observed result.
Freeze
Lock the surviving methodology instead of optimizing around the historical outcome.
Track
Collect qualifying future observations without moving the rules or backfilling missed evidence.
THE CRITICAL SEPARATION
Development is not validation.
BEFORE THE FREEZE
Historical evidence
Historical data helps us develop the hypothesis, examine its behavior, expose weaknesses, and decide whether the idea is worth carrying forward.
THE QUESTIONDoes this idea deserve another test?
RULES FROZEN
AFTER THE FREEZE
Prospective evidence
Once the methodology is frozen, only qualifying observations captured without knowledge of their eventual outcomes can add to the pristine prospective record.
THE QUESTIONWhat happens when the rules can no longer move?
RESEARCH STANDARDS
Make hindsight expensive.
No process can remove every research judgment. The goal is to make those judgments visible and to prevent later outcomes from silently rewriting earlier decisions.
Define the rules
Signal definitions, filters, ranking conventions, portfolio construction, horizons, and benchmarks should be explicit enough to reproduce.
Preserve the evidence
Weak periods, failed tests, data problems, and inconvenient results belong in the research record rather than disappearing from it.
Challenge chance
When appropriate, statistical tests ask whether the observed relationship could plausibly arise after the signal-return link is broken.
Check stability
A full-sample average can hide changing behavior. Temporal analysis asks whether the relationship persists across distinct periods.
Freeze before tracking
A surviving hypothesis enters prospective tracking only after the methodology is locked against further historical optimization.
Never backfill the future
A missed prospective observation remains missed. Reconstructing it later does not turn historical knowledge into prospective evidence.
PROSPECTIVE EVIDENCE
The evidence must existbefore the answer does.
Prospective validation is not simply a later backtest. The research state must be captured before the outcome is known, and the frozen rules must determine what qualifies.
01
Pre-outcome snapshot
Required inputs and the research decision must exist before the forward outcome becomes known.
02
Immutable record
The qualifying snapshot is preserved rather than silently revised after later information arrives.
03
No missed-date reconstruction
A date that was not captured prospectively cannot be promoted into the pristine sequence later.
04
Material changes reset the sequence
Changing the methodology materially creates a new research specification rather than rewriting the original one.
FAILED AND INCONCLUSIVE RESEARCH
A failed idea
still teaches us something.
If only favorable studies remain visible, the public record becomes more impressive and less informative.
Market Data Hunter is intended to preserve investigations that weaken, fail statistical challenge, become unstable, encounter unresolved data limitations, or simply remain inconclusive.
View Research LibraryMETHODOLOGY IN PRACTICE
Trend Quality
Our first published investigation moved through historical development, statistical challenge, and methodology freeze. It is now awaiting qualifying prospective observations.
View InvestigationCURRENT STATUS0
Pristine prospective observations
AWAITING OBSERVATIONS