Every position described on this site is simulated: no real order is ever placed on any venue, and no customer or personal funds are at risk. That doesn't mean the exercise is purely theoretical, or that bugs don't matter — quite the opposite.
Simulated, but with real consequences for the rules
A paper trade uses live, observed market data (funding rates, prices, order book conditions) and applies the exact same entry/exit rules a live version would use. The only thing missing is an actual order at a broker or exchange. Fills are modelled at the observed quote at the moment a rule fires — not a certified execution price, which is a real limitation this research is explicit about (see Methodology).
A real bug this process caught
A configuration value meant to enforce a minimum 22-hour hold before certain exits could fire existed in the rules file for weeks — but was never actually wired into the exit-check code that was supposed to read it. The result: positions could close within minutes of opening purely on funding-rate noise, before enough funding had accrued to cover even the fixed round-trip fee. 23 trades were closed this way before the bug was found and fixed, for a combined loss of −$219.72 — a small amount in absolute terms, but a 100%-preventable one, and exactly the kind of bug that paper trading is positioned to catch cheaply. Fixing it required nothing more than wiring an existing config value into the code that should have used it from the start.
Why this is more useful than a backtest alone
A backtest replays history against a fixed rule set and never encounters real infrastructure failures, stale data feeds, or wiring bugs like the one above. Running the rules continuously against live data — with real operational surface area, real outages, real API quirks — surfaces a different, often more important class of problem before it ever touches real capital.
What's tracked
Every simulated position records its rule version, its entry and exit reasoning, and its full accounting (funding, both legs' price movement, fees) — see When funding profit differs from total trade profit for how that accounting is presented, and the Lab Portfolio for the live record.