Market intelligence
Rule-based setups use trend, momentum, price action and multi-timeframe context. Signals move through staged filters before they are considered for execution.
Explore a working research stack for Bank Nifty and Nifty index options: live market-data analysis, rule-based signals, session risk controls, paper execution and reproducible replay.
Explore the connected systems behind a strategy—what triggers an idea, which rules reject it, and what happens after a simulated entry.
Rule-based setups use trend, momentum, price action and multi-timeframe context. Signals move through staged filters before they are considered for execution.
Capital boundaries, session controls, stop management and market-depth checks shape how candidate trades are handled.
Recorded ticks and depth events are used for historical replay, fill simulation and repeatable strategy investigations.
Journal entries, rejected signals, exit reasons and trading costs help connect results back to the decisions that produced them.
See the research workflowTradeStack separates a Python strategy and risk engine from an authenticated Node.js trading gateway. Each service is designed to make decisions traceable—not to predict guaranteed outcomes.
Stream index and options data into a consistent analysis pipeline.
Evaluate setup logic, market context and predefined risk constraints.
Run paper execution with tracked lifecycle events and modeled fills.
Replay available recordings, review costs and adjust rules based on evidence.
We aim to publish useful, clearly labeled research—not selective screenshots of winning trades. Performance reporting remains under validation while trade records, charges and exit classifications are reconciled.
In an October 16 controlled replay, an experimental 1-minute entry filter reduced the recorded after-cost loss compared with the baseline. Both variants still lost money, and the sample was too small to justify a strategy change.
Explore the documented case studySource: internal research findings; not independently rerun for this publication. Just two closed trades per variant. Neither is profitable.
A modular research stack separates the market feed, strategy engine, risk decisions, paper execution and historical analysis.
Each component has a specific role in execution reliability and reproducible analysis.
Built on actual code, not a concept deck. TradeStack's execution and gateway components have distinct responsibilities and security boundaries.
Processes Upstox market events, evaluates strategy setups, applies market and session risk filters, simulates orders and records decisions for research.
An authenticated paper-trading workspace offering session controls, market snapshots, trade and signal history, analytics and account-scoped data access.
Read source-grounded articles about system architecture, market-data replay and experiments that challenge our assumptions.
From real-time data to risk-checked paper signals.
Read the article ↗DuckDB, Parquet, simulation clocks and fill assumptions.
Read the article ↗The October comparison, after-cost results and limitations.
Read the article ↗We prioritize honest research, reliable order and data handling, and better tools for understanding decisions.
Our own authorized Upstox account runs the private, rule-based Python engine for paper execution, event recording and risk research. Visitors cannot control the account.
Reconcile fees, exit reasons, market recordings and paper-versus-replay behavior before making public performance claims.
Show approved, sanitized status, activity summaries and validated paper-session research from our single account. No raw broker feed, private logs or trading controls.
Let registered users authenticate with their own supported brokerage accounts and run isolated paper simulations. Separate data, capital limits, permissions and sessions per user; no access to our account and no public live-order execution.
Not yet. TradeStack currently uses one authorized account for private paper-trading research. A separate, account-isolated paper-simulation workspace for registered users is a long-term roadmap idea, not a released or guaranteed feature. Visitors cannot control our account, and public live broker orders are not available.
No live-account performance is claimed here. Research and simulation results must be labeled and verified before publication.
This website's source is public on GitHub. Availability and licenses for other TradeStack components vary; not all system code is publicly available.
Follow TradeStack's LinkedIn profile, the TradeStack YouTube channel, or our GitHub website repository.
Explore the project, follow its development or share feedback with the TradeStack research lab.