Nobody hands you the checklist between “I have a strategy idea” and “it’s trading live.” You discover it one painful item at a time, usually in the wrong order, usually under pressure. So here it is, written down, so you can at least see the whole gap before you fall into it.
From idea to a real backtest
- State the rules precisely enough to code.
- Get clean, correctly time-stamped data for the right instruments and period.
- Run it on an event-driven engine, not a vectorized shortcut, if it’s bound for live.
- Model spread and slippage; don’t fill at ideal prices.
- Respect the calendar, holidays, half-days, expirations.
- Check for look-ahead and survivorship bias.
- Read the orders and fills, not just the equity curve.
- Hold out data you didn’t tune on.
From backtest to paper
- Run the same code on a live feed, no rewrite.
- Confirm the orders and timing match your expectations on real-time data.
- Watch it over different market conditions, not one calm session.
- Compare paper fills against your backtest’s assumptions.
From paper to live
- Decide that live is genuinely earned, not just available.
- Connect your broker and confirm the account starts in a known state (usually flat).
- Accept that the broker is the source of truth for fills, positions, and buying power.
- Know which of your risk checks enforce and which only advise.
- Have a kill you’ve actually tested, and a rule for what happens to open positions.
To operate it at all
- Log every order state, every fill, every rejection, with timestamps and IDs.
- Capture errors as structured records, not just stack traces.
- Have somewhere reliable for the run to execute and stay alive.
- Know who’s watching and who acts when something breaks.
The point of writing it down
Look at that list. Exactly one section, the first half of the first one, is about your strategy. Everything else is infrastructure, operations, and discipline. That ratio is the whole story of why good ideas don’t make it to live: the idea is a small fraction of the work, and the rest is invisible until you’re standing in front of it.
You can work this checklist yourself, item by painful item. Or you can use something that has already done most of it, so the list you’re left with is short and the part you own is the part that was ever really yours: the strategy.