Live trading means placing real orders in financial markets with actual capital at risk. Executed trades change an account’s cash, positions and obligations. For a quant strategy, this is where a forecast becomes an implementable transaction: the system must obtain fills, account for costs and manage the positions it actually holds.
“Live” describes the economic commitment, not the speed or degree of automation. A manually submitted monthly rebalance can be live trading. An algorithm processing real-time prices in a paper-trading account still uses simulated execution. A live order can also be rejected or remain unfilled; submitting it does not establish that the intended position exists.
The price assumption becomes a cash ledger
Consider a wholly fictional illustration, not an actual account or performance record. An illustrative fill ledger buys 10 shares at USD 100.10 per share and later sells all 10 at USD 102.00. Total explicit fees across the two trades are USD 2. Assume no distributions, funding charges, taxes or other cash flows; all fills are stipulated complete. Cash amounts below are changes relative to the starting balance, not total account equity.
| Item | Research assumption | Illustrative fill ledger |
|---|---|---|
| Entry price, USD/share | 100.00 | 100.10 |
| Purchase cash flow, USD | −1,000 | −1,001 |
| Sale cash flow, USD | +1,020 | +1,020 |
| Gross trading profit, USD | 20 | 19 |
| Total explicit fees, USD | −2 | −2 |
| Net trading profit, USD | 18 | 17 |
For the fill ledger, gross profit is USD and net profit is USD. Reconstructing the cash movements gives the same result: USD. The ending share position is zero, so this simplified example has no remaining position to value.
Wholly fictional illustration, not an account statement or actual trading result. Complete fills and the same exit are stipulated. The research entry is an assumption, not an attainable-price guarantee.
The USD 1 difference comes entirely from paying USD more at entry under these assumptions. It is a slippage comparison only after the research reference price and its timing are specified. It does not establish whether the difference came from the spread, delay, market movement or the order’s own impact. Nor should that dollar be deducted again from the fill-based profit: the higher purchase price already contains it. The separate USD 2 fee is an additional transaction cost.
This tidy reconciliation depends on identical quantities and exits. With a partial fill, different exit time or an open position, comparing the final profit numbers alone mixes execution, exposure and valuation effects. Start by matching the decisions and resulting holdings.
Actual execution tests a different claim
A backtest asks what a specified model would have done on historical data under its execution assumptions. Paper trading can test incoming data and order-processing logic, but its fill rules remain simulated. For example, Interactive Brokers documents that its paper trades neither execute on an exchange nor settle at a clearing house, and that paper fills use the top of the book rather than deep-book access. Those are that provider’s simulator conventions, not a description of every paper platform.
Live trading adds evidence about actual execution and operations at the size traded. It also creates risks that a simulated balance cannot absorb:
- Capital and execution: prices can move against held positions, orders can fill partly, and liquidity can disappear. A limit order constrains its execution price but may not fill; a market order does not lock in a displayed quote. FINRA’s order guide explains this trade-off for securities orders.
- Account and custody: real cash and assets must be held, recorded and reconciled through the relevant broker, custodian and settlement arrangements. Access to them depends on those arrangements; a correct strategy forecast does not remove counterparty or operational exposure.
- Operating state: stale data, duplicate instructions, disconnections and uncertain order status can leave an account holding something other than the model’s target. A cancellation request is an instruction to cancel, not evidence that every outstanding order has ceased to exist.
The practical record therefore includes decisions, submitted orders, acknowledgements, fills, cancellations, positions, cash and charges. Comparing those layers reveals where implementation first departed from the research process. The Paper Trading and Staged Deployment lesson develops this reconciliation and a policy for progressing between stages; the full lesson requires membership, with a public preview available.
A real fill does not validate the strategy
One completed trade shows that a particular transaction occurred. Even a profitable live period does not by itself establish durable alpha, eliminate selection bias or demonstrate capacity at a larger allocation. Scaling can change participation, costs and feasible positions. Keep economic evaluation separate from evidence that the order system works.
For systematic deployment, define who responds to breaches, which conditions block new orders and how existing positions will be managed. As one jurisdiction-specific operational reference, FINRA’s 2015 guidance for firms using algorithmic trading strategies discusses limited-size pilots, heightened monitoring, reconciliation and mechanisms to disable an algorithm. It supplies useful control principles in its own supervisory context, not a universal legal checklist for all traders.
Stopping new orders and closing positions are different actions. A shutdown can prevent additional intended risk while leaving existing exposure, outstanding instructions or in-flight fills to resolve. Live trading is therefore an ongoing responsibility for capital and account state, rather than a final certification attached to a successful backtest.
Sources
- Ernest P. Chan, Quantitative Trading: How to Build Your Own Algorithmic Trading Business, first edition (Wiley, 2009), Chapter 5, pp. 79–80 and 89–91: execution systems, paper testing and divergence from expected performance.
- FINRA — Order Types: execution-price uncertainty and the risk that a limit order remains unfilled.
- Interactive Brokers — Paper Trading vs Live Trading: the provider’s execution, clearing and fill-simulation limitations.
- FINRA — Regulatory Notice 15-09, 26 March 2015: supervision and control practices for firms engaging in algorithmic trading strategies.