Paper trading a multi-account workflow before going live

Paper trading is often described as a way to test a strategy. That is only half of it. If your plan involves several accounts, several brokers and one order ticket, the first thing to test is the workflow itself: what you send, where it goes, what comes back and what happens when something fails.

This guide is a test plan for that workflow. It assumes no live money and no API keys for the first pass.

What paper trading can and cannot test

Paper trading is good at catching configuration and process errors. It is poor at reproducing market microstructure.

Paper environments differ by broker. Some simulate fills generously; others reject orders the live system would accept. Treat the results as a test of your process, not as evidence of future performance.

Step 1: start with the built-in demo

TraderDesk includes demo accounts that need no API key. They are the lowest-friction way to learn the layout: where the order ticket lives, how accounts are selected, where results appear and how positions are shown.

Spend the first session doing nothing but navigation. Find the kill switch. Find the log. Find the risk settings. Open the options chain and load a strategy template without sending anything. Familiarity reduces the chance that you discover the interface for the first time during a live order.

Step 2: add a broker paper account

Once the demo layout is familiar, add a paper or sandbox account from a broker you actually use. Alpaca paper keys and the Tradier sandbox are common starting points; other brokers vary. Use the broker setup page for the exact credentials.

Keep the paper account as its own labelled connection. Never put a paper account and a live account on the same ticket while you are testing. The whole point is to make the environment obvious.

Step 3: define the scenarios before you start

A paper-trading session without a scenario list becomes clicking. Write down the cases you want to cover, then work through them:

1. One account, one simple order. Market and limit, with a day time in force.
2. Several accounts, one order. Confirm each account receives a separate result.
3. A rejected order. Try an instrument or size the account cannot accept, and watch how the rejection is reported.
4. A closed market. Send an order outside the session and observe the broker's response, with the optional market-hours block on and off.
5. A risk limit breach. Set a low maximum value per order, then try to exceed it.
6. The kill switch. Halt routing, confirm orders stop, then restore it deliberately.
7. A multi-leg options basket. Build a defined-risk structure, review payoff and Greeks, and send it if the paper account permits it.
8. A log export. Reproduce an issue, find the lines in the log and export them.

Step 4: record expectations, not just outcomes

For each scenario, note four things: what you sent, what you expected, what happened and where you saw it. The gap between expectation and outcome is the useful part. If you expected three accounts to accept and two did, the reason matters more than the count.

Keep the log lines for anything unexpected. The event log records data requests, order actions and warnings; exporting the relevant rows is faster than trying to remember the sequence later.

Step 5: rehearse the live sequence at live speed

Paper trading is often done slowly, with pauses to read the screen. Live order entry is not always slow. Run the sequence at a realistic pace: select the accounts, check the notional, send, read the results. If that feels rushed in paper, it will feel worse live.

This is also where you learn whether your account labels and order defaults help or hinder. Rename accounts if a label is ambiguous. Set the risk limits you actually want, rather than leaving every field at "No limit".

Step 6: compare paper with a small live test

When the workflow is clean in paper, the next step is a deliberately small live order — one you are prepared to lose — used to compare the real result with the paper result. Check the fill, the fees, the position and the log. Do not scale up until the small live test matches your expectations.

If it does not match, stop and investigate. A difference in fill price may be normal market behaviour; a difference in the account that received the order is a configuration problem.

A go-live checklist

• Paper and live accounts are separately labelled and never mixed on one ticket.
• Every order type you plan to use has been tested in paper.
• A rejection has been produced and understood.
• The market-hours block and the kill switch have both been tested.
• Risk limits are set to values you can explain.
• The log has been searched and exported at least once.
• A small live order has been reconciled against the broker statement.
• You know which broker to contact for an order that TraderDesk cannot cancel.

What paper trading will not tell you

It will not tell you whether a strategy is profitable, and it will not tell you how you will behave when a position moves against you. It will not reproduce a broker outage, a halted symbol or a liquidity vacuum. Those are reasons to keep position sizes modest and risk controls on after go-live, not reasons to skip the paper stage.

Frequently asked questions

Next: read the multi-account order workflow or how to read a payoff graph and Greeks.