sysls

sysls is a systematic multi-asset long-short strategy framework designed to help individuals build systematic long/short portfolios across different asset classes. It provides a clean separation between:

• data connectors (market data access, option chains, Greeks, index universes)
• risk connectors (paper trading or connection to brokerages/exchanges)
• strategy definitions (user-implemented strategies)

The first supported broker/data source is tastytrade via the unofficial Python SDK, with a focus on equities, listed options, futures, and futures options. Users extend a base Strategy class, plug in a data connector and an optional risk/execution connector, and implement their own signal-generation logic (for example, use ATM call vs. put IV skew as a long/short signal for an equity universe such as the S&P100.)

Configuration and secrets are handled via Dynaconf.

Design Overview

At a high level, the framework is built arouind three core abstractions:

1. Data Connector
   • Fetches prices
   • Provides generic methods for strategies
2. Risk Connector
   • Places and manages orders, exposes positions
   • Can connect to brokerages or a local paper-trading engine
     • PaperRiskConnector
     • TastytradeRiskConnector
3. Strategy
   • Base class that users extend.
   • Requires a DataConnector, may receive a RiskConnector (default is paper)
   • Operates on a universe of symbols (e.g. sp100, sp500, ndx, rut2000) loaded via yfinance
   • Users implement generate_signals() to return a list of orders (long/short), given the current data.

Example use case:

• Reconstruct the at-the-money implied volatility skew for each stock in the sp100 (OEX).
• If the ATM call IV exceeds the ATM put IV, long the stock
• If the ATM put IV exceeds the ATM call IV, short the stock
• Send these equity orders through the configured risk connector (paper or live).

Getting Started

1. Install dependencies: using your preferred Python environment manager (I recommend Astral's uv) install dependencies listed in pyproject.toml
2. Configure settings: dynaconf init -f json at the project root. Fill in the relevant credentials in the .secrets.json file. [sample of the file goes here]
3. Write or use a strategy
   • Extend the base Strategy class in sysls.strategy.base.
   • Use the generic DataConnector methods to fetch prices, chains, or Greeks.
   • Return a list of Order objects from generate_signals. Example strategy implemented in the examples directory.
4. Run a strategy
   • Wire up a simple runner that:
     • Instantiates your chosen DataConnector and RiskConnector
     • Creates your Strategy subclass with a universe (e.g. sp100)
     • Calls strategy.rebalance()
   • Start in paper mode, then switch the risk connector when satisfied.

Contributing

Contributions are encouraged. Building and maintaining sysls is a big undertaking. Examples of useful contributions include:à

• New data connectors (e.g. IBKR, other brokers, data vendors, crypto exchanges for example through ccxt)
• New risk/execution connectors
• Example long-short strategies(equities, options, futures, multi-asset)
• Universe builders for additional indices or custom universes
• Backtesting and analytics tooling
• Tests, documentation, and CI improvements. Please open an issue to discuss major changes before submitting a PR.

Roadmap

This is an evolving project. A rough roadmap:

0. Implement all the base classes (in progress)
1. tastytrade integration (in progress)
   • Implement TastytradeDataConnector (spot, option chains, Greeks, candles)
   • Implement TastytradeRiskConnector (order placement, basic risk/position views)
   • Provide a few example strats
2. Additional brokers and data sources
   • Implement IBKR connectors (Use ibasync!!)
   • Add hooks for other retail-accessible brokers and possibly purely data-oriented connectors, e.g. DataBento or Polygon
   • Standardize connector interfaces so strategies remain broker-agnostic.
3. Vectorized backtesting layer
   • Introduce a backtesting engine that:
     • Uses the same strategy and connector abstractions
     • Can run historical simulations in a vectorized fashion
   • Support
     • Multiple universes, factor-style long-short portfolios
     • Realistic transaction cost/slippage modelling
     • Analytics on top (performance reports and diagnostics)
4. Infra and ergonomics
   • CLI for running strategies, scheduling, and switching environments
   • Documentation and cookbook examples for common strategy types

If you are interested in contributing to any part of the roadmap, feel free to open an issue and outline what you'd like to work on.