Field Notes
Short notes on the gap between strategy code and a real trading system, the part most tools skip.
AI trading tools vs trading infrastructure
The hype is about AI that finds you trades. The boring truth is that AI is most useful for building and debugging the system around the trades. Don't confuse the two.
Backtesting is the easy part
Everyone obsesses over the backtest. But the backtest is the most controlled, most forgiving step in the whole process. The hard part comes after.
Why broker state is a first-class trading-system problem
Most trading-system bugs in live are really disagreements between your code's idea of the world and the broker's. Broker state deserves to be treated as a first-class concern.
Why calendar bugs look like strategy bugs
When a strategy does something inexplicable on one particular day, the instinct is to blame the logic. Often the real culprit is the calendar, and it hides well.
Why managed platforms beat DIY for some systematic traders
Building your own trading infrastructure is a legitimate choice. For a lot of traders, it's also the wrong one. An honest case for buying instead of building.
From notebook to live trading: where the workflow breaks
The Jupyter notebook is where most systematic strategies are born and where many of them get stuck. A look at the specific points where the path to live snaps.
Why your Python script worked in research but broke in live trading
The script didn't change. The world did. A short tour of why research code falls over the moment it meets a live market, and what that tells you.
The real infrastructure behind paper and live trading
Paper and live trading look like a button in a UI. Behind that button is a surprising amount of machinery. A tour of what's actually there.
Strategy code is not a trading system
A strategy is a small part of a trading system. The rest, execution, fills, broker state, logs, monitoring, is where research dies on the way to live.
The hidden checklist between a strategy idea and a live trading system
Between 'I have an idea' and 'it's trading live' is a checklist nobody hands you. Here it is, written down, so you can see the whole gap at once.
TradingView alerts are not a trading system
An alert is a signal. A trading system is everything that has to happen after the signal. Confusing the two is how automated trading projects quietly fail.
What to log before going live
The day you go live is the worst day to discover you logged the wrong things. A short, tactical list of what to capture first.
Why backtests die on the way to live trading
A backtest is a controlled experiment. Live trading is the real world. Most of what kills a strategy lives in the gap between them.