Research and trade quantitative strategies using Python.
United States
Joined December 2018
- Tweets139
- Following73
- Followers1.7K
- Likes118
Pinned Tweet
Want a quick overview of QuantRocket? Check out this interactive demo of the user interface and core features:
quantrocket.com/#demo
Algorithmic trading is powerful but can be complicated. Good debugging tools are a must. 🐞 In this video, learn how to debug Zipline strategies in QuantRocket using the JupyterLab Debugger.
Get started with Zipline for free: quantrocket.com/code/?repo=z…
Curious about using machine learning in trading strategies? Try QuantRocket's free tutorial for MoonshotML, a machine learning backtester. Learn how to use walk-forward optimization with a random forest classifier to predict forward returns.
Tutorial: quantrocket.com/code/?repo=m…
⏰ Scheduling tasks is a basic requirement of automated trading. QuantRocket uses standard Unix cron for scheduling. Cron has stood the test of time: the first version of cron was released at Bell Labs all the way back in 1975.
Learn more: quantrocket.com/docs/#schedu…
Quants often find alpha in unique, proprietary datasets. QuantRocket’s open architecture makes it easy to import custom datasets into a SQL database for use with QuantRocket’s suite of research and backtesting tools.
Learn more: quantrocket.com/docs/#custom…
The best way to write code is interactively. Interactive programming makes it easier to develop your strategies and easier to debug them. In this video, learn how to debug a Moonshot strategy interactively in QuantRocket’s JupyterLab environment.
youtu.be/s_Wp8tsd4wk?si=qEBy…
Live trading results don’t always match backtests.
Some backtests fall apart out-of-sample due to overfitting.
Other backtests keep working but the live results don’t track the backtest. This is called implementation shortfall.
Learn more: quantrocket.com/docs/#perfor…
Event-driven backtesters like Zipline process one data point at a time, while vectorized backtesters like Moonshot process all data points at once using a library such as pandas.
Both styles have pros and cons. Learn more and try both with QuantRocket. quantrocket.com/docs/#moonsh…
💡In walk-forward optimizations, use a rolling window when you want your model to learn from recent data only. Use an expanding window when you want your model to learn from all available past data.
Learn more: quantrocket.com/docs/#ml-rol…
Are you a quant with an Interactive Brokers account?
QuantRocket's full-featured integration of the IBKR API supports advanced capabilities that platforms with more generic broker integrations rarely offer.
Learn more: quantrocket.com/interactive-…
@IBKR_QB
Customization can be essential to getting a quant platform to do what you want. Hosted platforms are risky: if no dial or button does what you want, you're stuck. QuantRocket offers an open architecture, filesystem access, and custom scripts.
Learn more: quantrocket.com/#customize
⚠️ Are your backtest results surprisingly good? Beware: you may be overfitting.
Avoid overfitting by taking large samples, choosing simple models, and not cherry-picking parameters.
Learn more in the Quant Finance Lectures (adapted Quantopian Lectures): quantrocket.com/codeload/qua…
Interested in backtesting and trading machine learning strategies?
QuantRocket supports rolling and expanding walk-forward optimization of machine learning strategies using MoonshotML. MoonshotML is compatible with scikit-learn and XGBoost.
Learn more: quantrocket.com/machine-lear…
Alphalens lets you quickly assess whether your alpha factor has predictive value. It's ideal for long-short and factor model strategies. Learn more in Lecture 38 of the Quant Finance Lectures (adapted Quantopian Lectures): quantrocket.com/codeload/qua…
The Pipeline API lets you screen, filter, rank, and analyze large universes of securities using price, fundamental, and alternative data. You can define rules and perform computations that are too complex for point-and-click screening tools.
Learn more: quantrocket.com/code/?filter…
Most traders focus on the US market — the most competitive market in the world. Yet the US stock market represents less than 50% of global market cap and only 25% of global listings. Find new opportunities in international markets: quantrocket.com/internationa…
QuantRocket retweeted
Quantrocket made the original Quantoptian backtester available again. Very cool for traders using python.
quantrocket.com/#
🆓 Want to run your algo trading on free cloud hardware?
Oracle Cloud offers a generous free tier: 4 CPUs, 24GB of memory, and 200GB of storage - for free. Your trading software must be ARM64-compatible.
Demand is high. Learn how to access capacity here: quantrocket.com/installation…
QuantRocket 2.11 has been released. Here's what's new:
💵 Model fees in Zipline:
* margin interest
* management and performance fees
* borrow fees
Cumulative fees are depicted in a new Pyfolio plot.
(1/6)
Full release notes: quantrocket.com/docs/softwar…
Full support for Apple Silicon
QuantRocket's IB Gateway container now runs natively on Apple Silicon, bringing full Apple Silicon support to QuantRocket. This means you can connect to Interactive Brokers for data collection and live trading on your new Mac.
(5/6)