SynapTick
Systematic trading research and experimentation platform.
Multi-strategy research and paper execution for US session ETFs — from hypothesis discovery and backtesting to guarded paper execution and outcome-driven learning.
SynapTick treats trading automation as a systems engineering problem: separate discovery, evaluation, and execution, enforcing explicit promotion rules, and closing the loop on real outcomes from paper trading and simulation. The goal is to avoid brittle single-strategy bots and instead maintain a reproducible research pipeline for testing and promoting new ideas.
Highlights
- Strategy families — Opening-range-style playbooks, momentum with VWAP confirmation, and mean-reversion frameworks share the same pipeline: one queue, one execution path, shared guardrails.
- Deterministic core — Scans, plans, and broker-facing logic live in versioned code; agents assist with research and narrative, not unchecked order placement.
- Paper-first — Designed around paper brokerage and local futures simulation with explicit execution modes (off, dry-run, paper-active).
- Closed-loop learning — Trade archives and structured paper/sim outcomes feed rolling performance views and signal salience, with post-close updates and optional health checks to automatically disable underperforming strategy families.
- Progressive risk — Symbol allowlists, caps on open positions and trades, daily loss limits, and optional auto-disable paths limit blast radius while experimenting.
How it fits together
At a high level: context (thesis, regime, insights, events) informs hypotheses, which enter a backtest queue. Passing ideas can graduate into a strategy registry. During the session, a scanner and strategy layer produce order intents; a single execution worker applies limits and may submit paper orders or run the futures sim. Outcomes are logged and aggregated so performance feeds back into confidence and documentation — not ad hoc edits.
flowchart LR
A[Context / Market Thesis] --> B[Hypotheses]
B --> C[Backtest Queue]
C --> D[[Strategy Registry]]
subgraph Execution Loop
D --> E[Session Scanner]
E --> F[Execution Worker]
F --> G[Paper Orders]
F --> H[Futures Simulation]
end
G --> I[Outcomes]
H --> I
I --> J[Learning / Performance Analysis]
subgraph Research Loop
J --> B
end
Stack (high level)
| Layer | Notes |
|---|---|
| Language / runtime | Python for workers, scans, and backtests |
| Brokerage (paper) | Alpaca paper for equities / ETFs |
| Simulation | Separate futures-sim venue with its own limits |
| Orchestration | Scheduled jobs and role-scoped agents (e.g. discovery vs executor) |
| Data | Append-only logs, feature rows, and outcome sinks for learning |
Status
Active research and paper trading — not a product, not a signal service. The goal is reproducible experimentation and honest accounting of results in constrained, non-live environments unless you deliberately change that policy.
Links
- System architecture & design rationale: SynapTick design overview
- Contact: Use this contact form!
Disclaimer
SynapTick is for education and research only. Nothing here is investment, tax, or legal advice. Past backtest or paper performance does not guarantee future results. Trading involves substantial risk; use paper and simulation until you fully understand behavior and limits.
SynapTick is a personal project showcase. Naming and scope reflect the author’s stack, not an offer to manage capital or provide recommendations.