Readable growth
Add conditions incrementally so candidates remain explainable.
INTELLIGENCE / AUTO SEARCH
Explore more combinations than a human can test.
kquant Engine progressively combines readable signals, regimes and filters, ranks candidates and separates discovery from independent confirmation.
HYPOTHESIS SPACE1,248 trials
01Trend Momentum0.9302Funding Regime0.8903Volatility Breakout0.68
explorerankfreeze
SYSTEM DOMAIN02 / DISCOVER
Search, rank and freeze a candidate
- DISCOVER
- OPTIMIZE
- VALIDATE
- RUN
- OBSERVE
Confirmation
Require candidates to hold quality outside the selection window.
Verified state
Keep incomplete evidence visibly unverified.
ML without magic
Use models to filter an economic signal, not invent one from noise.
DECISION FRAMEWORK
Did the candidate improve evidence or only the training score?
Read the candidate against its anchor, neighbouring parameters, sample size, costs and complete trial history.
Bound the search space
Compare several objectives
Freeze before opening holdout
Candidate record
Important
Historical, statistical and model results do not promise future returns. kquant provides computational tools and technical execution, but does not assess whether a specific trade is suitable for a user.