DOJO — No-Code & API-First Statistical Enumeration Engine
First platform doing exhaustive statistical research at this scale. No-code UI plus a programmatic API for LLM agents and automated pipelines. Privacy-first quantitative research tooling by Student One. Statistical exploration tooling — not investment advice.
Why DOJO Exists
Traditional platforms optimize — they search for a single 'best' indicator configuration. The problem: that 'best' result is almost always overfitted. DOJO takes the opposite approach: exhaustive enumeration. Instead of finding one answer, it tests every answer and returns the complete solution set, then runs survivors through an 18-gate statistical cascade.
What DOJO Does
- 1,960,000 parameter pairs per sweep — every (period_A, period_B) combination from 1 to 1,400
- Native 1-minute resolution — no resampling into 5m/15m/1h bars, zero interpolation artifacts
- 18-gate statistical cascade — Deflated Sharpe (Bailey & López de Prado), BH FDR, Romano-Wolf, Hansen's SPA, PBO with 3,432 combinatorial splits, walk-forward survival, hit-rate, percentile-floor robustness, regime-aware filters
- 30+ indicator families — RSI, MACD, EMA, CCI, Ehlers Super Smoother, FRAMA, MAMA, Fisher Transform, Hilbert Transform, and more
- No code required — or fully scriptable — use the no-code UI, call the programmatic API from LLM agents (Claude, GPT, Gemini) and CI pipelines, or wait for the CLI on the roadmap
- Audit-ready output — CSV, JSON, Parquet, Arrow IPC. Your thresholds, your interpretation.
Privacy & Security
We retain nothing, deliver everything. Ephemeral compute architecture: user input data is not persisted beyond the job lifetime. Bring-your-own-key data sourcing (Tiingo, Polygon) keeps vendor credentials session-scoped — never stored server-side. Output artifacts are yours, exportable in machine-readable formats designed for downstream research and audit trails.
Two Engines
Gemina™ — Cross-asset lead-lag detection at minute resolution. Six directional tests — does A lead B? does B lead A? — exhaustively enumerated at every minute of every trading session across every lookback window. Surfaces asymmetric lead-lag behavior that machine learning systematically misses because single-window ML structurally cannot test all of them.
PermuCheck™ — Single-asset indicator-parameter sweep. Two indicators, every lookback period from 1 to 1,400, one-minute resolution. Returns the complete solution set after the 18-gate statistical cascade.
The Problem with Indicator Optimization
Standard backtesting platforms resample 1-minute data into 5m, 15m, or 1h bars before applying indicators. Every resampling step introduces interpolation artifacts, phase errors, and information loss. When you then optimize across only the standard timeframes, you're searching a tiny fraction of the parameter space and calling it 'the best.' DOJO tests every integer period from 1 to 1,400 on the canonical 1-minute axis using DSP-based windowing — no data loss, no resampling errors, no optimization bias.
Who Uses DOJO
- Quantitative researchers and academic groups studying market microstructure
- Hedge funds, family offices, and proprietary research desks
- Registered Investment Advisors (SEC, FCA, SEBI) building research workflows
- Independent analysts evaluating indicator robustness rigorously
Compliance Posture
DOJO is statistical research and exploration tooling. It is not investment advice, not a recommendation engine, not a backtester, not an order-management system. No performance forecasts, no return claims, no signal interpretation is provided. Output artifacts are statistical objects; downstream interpretation and any operational use are the user's responsibility.
Free Sandbox
Free research-tier access lets you evaluate the engine before committing to a paid plan or enterprise engagement.
Statistical exploration tooling — not backtesting, not investment advice. © 2026 Student One Causal Networks Private Limited (U62013PN2025PTC244699)
Student One Research Portal · sales@studentone.tech