02 / CONTROLLED EXECUTION
Backtest Runner
Market, period, sizing and friction are explicit before a historical run begins.
BACKTESTAPP / QUANTITATIVE SOFTWARE ENGINEERING
A real research workflow.
A real research workflow.
Designed for evidence.
From a market-data catalog to controlled historical runs and reviewable results. This case study presents the interface, engineering decisions, and delivered software capabilities behind a local research system.
PYTHON + FASTAPIDETERMINISTIC REPLAYLEDGER ACCOUNTINGVERSIONED RESEARCH
RESEARCH
BACKTEST
REVIEW01/05
BACKTEST
REVIEW01/05
THE PRODUCT / PUBLIC INTERFACE STUDYEngineering, not just
Engineering, not just
an equity curve.
Five authentic screenshots from the BacktestApp application, with only account identity, run identifiers and selected trade details obscured. The visible UI and historical research screens are from the original software—not recreated mockups.
AUTHENTIC APPLICATION CAPTURE View full size ↗01 / Research Workspace
PUBLIC-SAFE UI RECONSTRUCTIONData context, recent execution history and evidence-first research monitoring.
BASED ON SUPPLIED APPLICATION CAPTURES
THE WORKFLOW / FOUR ADDITIONAL VIEWSSee how a research idea
See how a research idea
becomes a reviewable result.
03 / RUN OBSERVABILITY
Research Run
Visible progress, phase transitions, and report-oriented research workflow.
04 / RESULT INTERPRETATION
Results Analysis
Actual results screen with equity, drawdown, modeled fees and a recorded trade count. Its legacy evidence tab explicitly marks detailed evidence as unavailable for that historical run.
05 / RESEARCH GOVERNANCE
Strategy Library
Versioned ideas, baseline comparisons, and evidence status—not unverified rankings.
These are five original BacktestApp captures supplied by the project owner, not generative redesigns. Private account labels, run identifiers and transaction details are covered with opaque masks. The remaining software UI, visible metrics and historical research charts are preserved. These are retained historical simulation screens, not claims of live trading. Each screenshot opens at its original resolution. Verify the five published asset hashes ↗
ENGINEERING VALUEWhat the interface
What the interface
reveals about the system.
01 / DATAMarket-data provenance
Imported histories are evaluated under explicit source, chronology and coverage rules.
02 / LOGICCausal research contracts
Closed observation timing and deterministic replay reduce ambiguity in historical decisions.
03 / FINANCELedger-grounded results
Execution, fees and portfolio state are reconcilable instead of being inferred from a chart alone.
04 / REVIEWEvidence-aware workflows
Run history, configuration and result interpretation support structured technical inspection.
GO DEEPER / RESEARCH PAPER 001Not just what we built.
Not just what we built.
How we reasoned about it.
The publication documents system architecture, causal execution discipline, a reproducible synthetic accounting example and selected verification evidence. The proprietary mechanism stays private.