Research Overview
Understand how Statly research produces evidence, candidate bots, and launch guidance without hiding uncertainty.
The research section explains how Statly turns raw investigation into something a customer can actually act on. It is not an internal runbook and it is not a marketing claim layer. Its job is to show how evidence becomes candidate bots, what warnings mean, and where user choice begins.
Why these docs exist
For a research product, documentation is part of the trust layer.
These pages are meant to answer the questions a skeptical operator should ask:
- What exactly is being measured?
- What does the product mean by evidence?
- How is overconfidence avoided?
- Where can a feature fail even if the headline looks good?
What this section is for
Use this section to understand:
- how research runs generate evidence
- how promising results become Candidate Bots
- why
Backtest,Paper, andLiveare real customer actions - how institutional validation and customer launch choice stay separate
- which parts of the methodology are public and which remain internal operator detail
The bridge object: Candidate Bot
A Candidate Bot is the product object that connects research to operation.
It is the point where a customer can decide what to do next:
- run another Backtest
- activate in Paper
- activate in Live
The research system should guide that choice, but it should not pretend that a missing internal promotion automatically removes all user agency.
Two truths must stay separate
Institutional status
Institutional status is the internal research and governance posture.
Examples include:
discoveredvalidatedpaper_candidatepromoted_live
This describes how far a feature or pattern-under-study has advanced inside the governed research ladder.
User launch posture
User launch posture is the customer-facing suggestion layer.
Examples include:
backtest suggestedpaper suggestedlive with warninglive verified
This describes what the customer is being advised to do next inside the workspace.
What the product should do
Statly should guide, not lie.
That means:
- warnings must explain what is missing
- a stronger warning is different from a fake block
- institutional research truth should remain visible
- customer choice should remain real where the product allows it
What these pages cover next
Use the pages in this section together:
- Data Provenance And Hygiene explains what kinds of market data feed the research system and how public trust claims should stay conservative.
- How Validation Works explains the ladder from evidence to paper and live.
- Feature Evidence vs. Backtest Evidence explains why a forward-associated feature and a tradable strategy are not the same thing.
- Metrics And Evidence defines the core metrics and what they do or do not prove.
- Methodology And Risk covers point-in-time discipline, multiple testing, and the limits of the current public claims.