Here's a quick little python script to scrape historical orderbook data from Bybit.

Nothing fancy but gets the job done

Link 👇

Another StrategyPattern in action!

I refactored my F2P backtester's slippage logic, moving from a fixed percentage to a quick and dirty rolling, volatility based estimate.

Code & details in the article.

Link in replies

To build robust, realistic backtests, you must account for all trading costs like commissions, holding costs, slippage, etc.

We migrated our fees and funding logic to a more interface-based backtester, designing it with modular, swappable components.

Link in replies 👇

While migrating our pre-cost SR calculations to our new backtester we decoupled instruments, trading rules, features, signals, PNL & SR into reusable components and caught a critical parameter bug by using proper testing techniques.

Link in replies 👇

Your home IP address is like a permanent ID tag on everything you do online. While VPNs aren't perfect, here's why they're still essential for basic privacy in 2025 - and how to choose one that actually protects your data.

Read more👇

This post discusses improving the design of our backtesters position rebalancing logic by introducing proper interfaces, breaking down complex functions and implementing assert tests while decoupling components like instruments, trading rules, features, and signals.

Link 👇

This week we're exploring common OOP buzzwords like Inheritance, Composition & SOLID Principles while refactoring our backtesters data sources.

Turns out it's not really that hard to decouple components, build clean interfaces & make code more maintainable

Link in replies 👇

Working with code can be a pain. You change some things and nothing's working anymore. All of a sudden u'r wasting time on an hours long witch-hunt for bugs.

This article gives a high level introto refactoring and what it's all about using our current backtester as example.

Link in the replies 👇

A common error when backtesting is expecting to get filled exactly at the close price without accounting for slippage.

This article shows how to account for slippage during execution and PNL calculations.

(Link in reply)

How to bring transaction costs of a continuous signal down with a rebalancing error threshold.

Link in replies 👇

We've upgraded the backtest and added very simple turnover and funding calculations.

We'll improve these in the future after we've worked our way through all concepts mentioned in the original article.

Link to article is in the comments 👇

Our very first backtest.

Even though this isn't nearly the end of it, it already highlights some of the most common pitfalls - mainly cost considerations - a lot of people get wrong when backtesting

👇