Strategy Lab is the systematic trading platform retail was missing. Institutional-grade backtesting engine. Full Python SDK with first-class access to every A-Trader app. Event Impact, Sentiment Compass, Order Flow, Correlation Hub, the Freedom Challenge Quality engine. A growing library of forkable strategies. One-click paper → live deployment. A marketplace to monetize what you build. The toolkit that turns a discretionary trader into a systematic one, or a systematic one into something institutional.
For thirty years, systematic trading has been a two-tier world. Hedge funds and prop desks get Bloomberg Terminal feeds, Refinitiv data, multi-million-dollar backtesting infrastructure, full Python stacks, and quant teams who maintain it all. Retail traders get MetaTrader's MQL5. A language whose tutorials still reference Windows XP, an event model from 2005, and a backtester famous for producing equity curves that bear no resemblance to live trading.
Every retail trader who tried to go systematic in the last decade has hit the same wall: the tools were built for a different century.
The gap isn't ability. The gap is infrastructure. Strategy Lab closes it. A real Python environment. A backtesting engine that matches live execution. Direct programmatic access to every signal source you'd build yourself if you had a team of three engineers and two years (which you don't). And a path from "I wrote my first strategy this weekend" to "I'm running a small systematic book that funds my discretionary trading", without ever needing to rewrite the code, change platforms, or hire anyone. That path didn't exist for retail before. Now it does.
Strategy Lab isn't a backtester. A backtester is one slice. The platform integrates research, development, testing, deployment, and monetization in a single environment, and connects every one of them to the live A-Trader apps that retail traders pay for separately today.
Vectorized where possible, event-driven where realism matters. Same code path as live execution. No "backtest worked, live failed" silently because of look-ahead bias or fill-model mismatches. Microsecond-resolution data where the venues support it.
Native Python, not a DSL pretending to be Python. Bring your own libraries. NumPy, pandas, scikit-learn, PyTorch, all callable from inside a strategy. Type hints. Async event handlers. The environment quant developers actually want.
Browse strategies built by other Strategy Lab users with live performance metrics. Fork to your environment in one click. Read the source. Modify. Backtest your version. Learn by doing what others have already done. The way every other developer ecosystem works.
The same strategy code that ran in backtest runs in paper. The same code that runs in paper runs live. Switch via a flag, not a rewrite. Built on A-Connect's execution infrastructure. Sub-15ms latency, multi-venue smart routing, the same plumbing institutional desks use.
Publish a strategy to the marketplace. Set a subscription price. Other traders run your strategy on their own A-Trader accounts; you collect a recurring revenue share. The first systematic-trading platform built for the creator economy.
For serious developers and proto-institutional users: direct access to our quant team. Code reviews on complex strategies. Architecture advice. Co-development on hard problems. Available as a paid tier for traders running enough capital to justify it.
Most retail backtesters lie by omission. Look-ahead bias. Unrealistic fills. Zero slippage. Missing trading costs. Survivorship bias on the data. The strategy looks brilliant in test and dies in live within a week. Our backtester was built by the team that built A-Connect's live execution. Same fill logic, same latency assumptions, same commission models. The backtest matches live because they share the code.
The point of backtesting isn't to find an equity curve that goes up and to the right. The point is to find a strategy whose live performance will match the backtest. That requires a backtester that's as pessimistic about your strategy as the market will be. Ours is.
Strategy Lab is real Python. You import numpy. You import pandas. You import your own modules. You write idiomatic code with type hints, async event handlers, and unit tests. The SDK gives you primitives for data access, order management, and app integration, and gets out of your way for everything else.
# A minimal moving-average crossover. Idiomatic, type-hinted, testable. from atrader import Strategy, indicators as ind class MACrossover(Strategy): def setup(self) -> None: self.symbol = "EUR/USD" self.fast = ind.sma(20) self.slow = ind.sma(50) def on_bar(self, bar) -> None: if self.fast.crosses_above(self.slow): self.buy(self.symbol, size="2%", stop="1.5%") elif self.fast.crosses_below(self.slow): self.close_position(self.symbol) # Run from anywhere: command line, notebook, scheduled job. if __name__ == "__main__": MACrossover().backtest(period="3y", capital=10_000)
Every developer ecosystem worth using has an open library of shared code. Strategy Lab launches with 400+ strategies written by the build team and early contributors. Each one comes with its source, its backtest history, and, if the author has chosen, its live performance. Fork in one click. Modify. Backtest your version. Learn the way developers actually learn.
Donchian-style breakout with dynamic channel width based on rolling ATR. Filters entries through Sentiment Compass and pauses around tier-1 macro events via Event Impact.
Pairs-trading mean reverter. Uses Correlation Hub for cointegration tests and Order Flow for entry-timing confirmation. Auto-disables when correlation regime breaks.
Captures the historical mean-reversion of NFP overreactions in EUR/USD. Trades only when the surprise size exceeds 1.5σ of historical distribution. Auto-throttles in tightening regimes.
Sells volatility on gold when realized vol exceeds the implied vol surface. Hedges with futures. Includes scenario test against historical crisis periods (2008, 2020, 2022).
Captures funding-rate divergences across perpetual exchanges. Multi-venue. Auto-rebalances. Risk-managed via Correlation Hub to avoid concentration risk in stablecoin exposure.
A heavily-commented walkthrough strategy. Designed to be read, not run. Each line annotated with the why. The recommended starting point for anyone new to systematic trading in Python.
Most retail systematic platforms make you rewrite for live. Different APIs for backtest vs. paper vs. live. Different fill assumptions. Different data handlers. The result: 40% of "successful" backtests die in the rewrite, and 80% of the rest die in the first week of paper trading. Strategy Lab is one code path running against four execution targets. The same Python file. Different flag. Zero rewrite.
Run against historical data. Realistic fills. Honest commissions. Walk-forward and Monte Carlo available. The truth-telling backtester. Run as often as you want.
Same code, live market data, simulated fills. Watch your strategy operate against the real market in real time without putting capital at risk. Run for a week, a month, however long you need conviction.
Flip a flag. Same code now sends real orders at deliberately small size. The most expensive scientific experiment of your life, but a controlled one. Live frictions reveal themselves at minimal cost.
If the small-size live performance matches paper and backtest, scale up. Position-size config is one parameter. The same code that ran your first $10 trade runs your $10,000 trade. Same code path, more zeros.
This is what makes Strategy Lab different from every other retail systematic platform. Each of the 14 A-Trader apps exposes a Python SDK surface. Your strategies can ask Event Impact when the next NFP is, ask Sentiment Compass whether retail is extreme on EUR/USD, ask Order Flow whether the cumulative delta supports the trade. Capabilities that retail systematic traders have never had access to, now available as one-line function calls.
Auto-flatten before major releases. Throttle around medium-impact events. Trade the post-event mean reversion. Build NFP fade strategies in 30 lines.
Fade extreme retail positioning. Confirm trend trades with sentiment alignment. Filter entries by sentiment-vs-price divergence. Contrarian edge, systematized.
Require cumulative-delta confirmation. Wait for iceberg patterns. Trade only when aggressor flow agrees with your signal. Microsecond-resolution tape reading without watching it yourself.
Throttle when portfolio correlation exceeds your threshold. Pause pair trades when cointegration breaks. Adapt position sizing to current regime. Institutional-grade exposure management.
Operate within a defined VaR budget. Auto-trim when exposure breaches portfolio limits. Build kill-switch logic on regime-conditional risk numbers. Live risk management at the strategy layer.
Auto-reduce after detecting revenge-trade patterns. Pause during identified tilt episodes. Behavioral analytics applied to systematic strategies. Preventing your own worst self from running the bot.
Publish a strategy. Set a monthly subscription price. Other A-Trader users subscribe; the strategy runs on their account; you earn a recurring share on every subscriber, every month. Live performance is verified by the platform. No fake screenshots, no curve fits, no "trust me bro" track records. The strategy either makes money in production or it doesn't.
Authors keep 70% of subscription revenue. The platform takes 30% to cover execution costs, data costs, and the verification infrastructure that makes the whole system trustworthy. No tier-1 hedge fund will publish on the marketplace, but a serious retail systematic developer running a few hundred subscribers can make this a meaningful second income, and a few have already turned it into a primary one.
The most-used parts of Strategy Lab are free. Backtesting on cloud compute. Paper trading. Reading the library. Authoring strategies and publishing them. We make money when you go live with serious capital (or when you join the Quant Pro tier for embedded support) and we want the path to "go live" to be as wide as possible.
For students, hobbyists, and anyone learning systematic trading. Full backtesting. Full paper trading. Read and fork the entire library. Monetize strategies in the marketplace.
For developers running real strategies with real capital. Unlimited live strategies. Full app SDK access. Priority backtest queue. Live data on all supported instruments.
For developers running enough capital to justify embedded support, and for the first wave of institutional users (asset managers, hedge funds, family offices) using Strategy Lab as part of their workflow.
The systematic trading platform retail was missing. Finally here, free to start, paid only when you're serious enough to run capital through it.