Loading market quotes…

Subscribe free
Market Fieldnotes11 minFreeGBC Research

Paper trading: what a simulation can and can’t test

In short

Paper trading means trading a simulated account with no real money. It tests the mechanics well: tickers, quantities, order types, recorded exits, routine and journal. It is much weaker on fills (some simulators fill a limit order as soon as the price touches it), costs, gaps and how you behave with real money. In a hypothetical example, a month that ends at +$580 in the demo account becomes −$113 under conservative assumptions.

On the morning of August 1, 2012, in the first 45 minutes of trading, Knight Capital’s order router sent more than 4 million orders while trying to fill 212 customer orders. It traded more than 397 million shares and left the firm with a loss of more than $460 million. According to the SEC, which settled the case the following year with a $12 million penalty, the cause was badly deployed new code that triggered a defective function, so the system could no longer recognize when an order had been filled and kept sending more.

It was an operational error, the kind you find by rehearsing a procedure before trusting it with real money. In a personal account, operational errors are smaller and more mundane: the wrong ticker, an extra zero in the quantity, a stop that was never entered. Those are exactly what paper trading is good at exposing.

The rest, starting with how you will behave when the loss is yours, a simulation does not measure. And a positive virtual balance can hide an account that, with real fills and costs, would be losing money.

What is paper trading?

Paper trading is simulated trading. You pick stocks, send orders and manage positions as you would for real, but with virtual money. The name dates from when traders logged hypothetical trades by hand on paper. Today many brokers and platforms offer a demo account linked to market prices.

A simulation tells you fairly precisely whether you can execute your plan without mistakes. It tells you much less about whether the plan makes money, because that takes realistic fills, real costs, more trades than you place in a month and the pressure of your own capital.

What can a demo account test?

Tickers, quantities and order types

A demo account is the right place to pick the wrong ticker between two similar names, type 1,000 instead of 100, or confuse a stop with a stop-limit. One extra zero on a $40 stock turns a $4,000 position into a $40,000 one. Deliberately practice a limit order, a cancellation and a partial fill, and watch how the platform handles each. The differences between order types are in our guide to market, limit, stop and stop-limit orders.

The routine

A simulation also tests whether you can keep a schedule: check the market, write the plan before the order, record the outcome. Run it on days with no opportunities too, and keep those days in the log. A simulation with endless retries, or one where every loss quietly disappears, mostly tests your skill at editing the log.

The journal

Save the entry reason, planned holding period and exit conditions before you see the result. Editing your notes after a rally erases the evidence of what you knew at the start. If you use TC2000, save the chart next to the entry note in the same workspace you will use with real money (ours, refined over about two years, is available as a TC2000 setup), because rehearsing the routine on tools you won’t use live tests a different routine. The full structure is in our guide to the trading journal.

What can’t a simulation test?

Fills

In a real market, a buy limit order at $40 joins the queue behind orders that arrived earlier at the same price. If the stock touches $40, trades 800 shares and bounces, whoever was at the back of a 1,600-share queue gets nothing. Some simulators fill the order as soon as the displayed price touches the limit. Interactive Brokers, for example, states that fills in its paper account are simulated from the top of the book, with no access to deeper book data. The fills a simulator gives away are also the most profitable ones, because a stock that touches your limit and bounces right away is exactly the trade that works best.

The demo account fills an order the market would have left in the queueThe demo account fills an order the market would have left in the queue. Buy limit at $40.00 for 100 shares; the price touches the limit once and bouncesFIG. 1 · SAME ORDER, TWO ASSUMPTIONSGB.CAPITALThe demo account fills an order the market would haveleft in the queueBuy limit at $40.00 for 100 shares; the price touches the limit once and bouncesTHE SIMULATORLimit $40.00touches onceFilled on the touch: 100 shares at $40.00Exit at $41.50: +$150 in the demo accountTHE REAL MARKETBuy queue at $40.00: 1,500 shares arrived beforeyours800 traded at $40.00 before the bounce700 still ahead of you in the queue100 yours, at the back of the queueThe price bounces before the queue reachesyou: not filledReal result: $0Hypothetical example. Some simulators fill on the touch or from the top of the book, without rebuildingthe queue.
Figure 1 A limit order the price touches once: filled for the simulator, still in the queue for the market.

Before reading on, open your demo account or your paper trading sheet and count how many limit orders show as filled at a price the stock touched only once without trading through. Those are the trades to recheck first.

Costs and slippage

If the simulator doesn’t charge commissions, spread and currency conversion, add them yourself. A simulated trade that makes $8 with $10 of omitted costs is a $2 loss. The same goes for stops, which in a real account become market orders and fill at the available price, often a few cents worse than the level. The full list is in our guide to trading costs.

Gaps

If you log trades by hand, using the stop level as the exit price can understate a loss after an earnings gap. A triggered stop executes at available prices, not at a guaranteed opening price. In the exercise below, the opening price is a modeling assumption. In a demo account linked to live prices, a gap only shows up if one happens during your test, so it pays to simulate one on purpose, as we explain in our guide to earnings gaps.

Emotions

In 2002 the economists Charles Holt and Susan Laury published an experiment on risky choices in the American Economic Review. When the prizes were real and scaled up by factors of 20, 50 and 90, participants chose far more cautiously; when the prizes were hypothetical, raising the stakes barely changed their behavior. Losing a virtual $500 and losing money you need for a bill do not create the same pressure, and no simulation can tell you whether you will honor the stop when the loss is yours.

How much does the result change under realistic assumptions?

Take a hypothetical month of paper trading: 20 trades of 100 shares, 9 winners of $150 and 11 losers of $70. The demo account shows +$580, or $29 a trade. Before seeing that number, though, you had written four conservative rules, and now you apply them to every trade.

The same month recalculated, hypothetical example with 100 shares per trade
Conservative ruleEffectResult
Demo account result20 trades+$580
A limit fills only if the price trades one cent through it3 winning trades of $150 disappear: −$450+$130
Commissions of $2 per order17 trades, 34 orders: −$68+$62
Slippage of $0.05 a share on stops11 stops of 100 shares: −$55+$7
Stops fill at the open after a gap1 stop, $1.20 past the level: −$120−$113
Same month: +$580 in the demo, −$113 under conservative assumptionsSame month: +$580 in the demo, −$113 under conservative assumptions. 20 simulated trades of 100 shares, recalculated with rules set before seeing the resultsFIG. 2 · FROM VIRTUAL BALANCE TO REAL TERMSGB.CAPITALSame month: +$580 in the demo, −$113 under conservativeassumptions20 simulated trades of 100 shares, recalculated with rules set before seeing theresultsDemo account result, 20 trades+$580Three limit orders filled onlyon the touch−$450Commissions: 34 orders × $2−$68Slippage on 11 stops: $5 each−$55One $1.20 gap past the stop−$120Result under conservativeassumptions−$113Hypothetical example: commissions, slippage and the gap are teaching assumptions, not broker data.
Figure 2 The four conservative rules applied to the same month of simulated trades, in the order they were written.

The month goes from +$580 to −$113, and the share of winning trades from 45% to 35%. The four rules describe ordinary conditions in a real account, with no extreme scenarios. When you decide them matters too. Alternative assumptions have to be written at the start and applied to every trade. Chosen after a loss, they become a way to build the record you would like to have.

How long should you paper trade?

Long enough to identify and correct the operational errors that appear during the test. There is no fixed duration that proves you are ready for live trading. To prove that a method makes money, a month of simulation isn’t enough, and the reason is statistical. With 20 trades and a 45% win rate, the 95% confidence interval runs from about 26% to 66%, so the figure is consistent with a very good method and with a losing one. With 100 trades the interval narrows to between 36% and 55%, still wide. How win rate, risk-reward and expected result fit together is in our guide to win rate and risk-reward.

That is why, in the protocol below, the protocol below focuses on recorded operational errors, with the virtual balance as secondary information.

Four phases, one test before real moneyFour phases, one test before real money. You move to a real account when operational errors have been at zero for ten sessionsFIG. 3 · THE 30-DAY PROTOCOLGB.CAPITALFour phases, one test before real moneyYou move to a real account when operational errors have been at zero for ten sessionsBEFORE YOU STARTWrite the rulesHow the simulatorfills,hypotheticalcosts, capitalequal to the realone, a fixed time.WEEK 1Drill themechanicsLimit and cancel,stop andstop-limit,partial fill,look-alike ticker,lost connection.WEEKS 2-3Train theroutineA written planbefore everyorder, a completejournal, no-tradedays included.WEEK 4Make itrealistic, thentestConservativerecalculation, onesimulated gap,error count, finaltest.GB Capital teaching outline. The thresholds are conservative choices to adapt to your method.
Figure 3 The 30-day protocol: each phase has a test for moving on, and the last checks whether the procedure and records are complete.

How do you move from a demo account to real money?

The first month with real money is another test, with one new variable. Use a reduced size, for example a quarter of what you plan, for the first 20 trades, and measure one thing: in how many trades you followed the written plan, compared with the simulation. If you followed it 20 times out of 20 in the demo and 15 with real money, you know what to work on before sizing up. The checks to run before the first real order are in our 10 rules for your first swing trade.

In the cost example, the $1.20 gap adjustment is additional to the $0.05 already charged to stops: total slippage on that particular exit is $1.25. If the simulator already includes spreads or commissions in fills and balances, do not charge them again. The test checks procedure. Even thirty error-free days do not establish future profitability.

Frequently asked questions

Is paper trading worth it?

Yes, for mechanics and routine: tickers, quantities, order types, exits, journal. It is far less useful for judging whether a method makes money, because simulated fills are often more generous than real ones, costs may be missing and the pressure of real money is absent.

How long should you paper trade before going live?

Use the demo to identify and correct operational errors. The protocol here covers 30 days of exercises, repeating the steps where errors remain. Judging profitability takes many more trades: with 20 trades and a 45% win rate, the 95% confidence interval runs from about 26% to 66%.

What is the difference between paper trading and backtesting?

A backtest applies fixed rules to past prices, usually with software. Paper trading happens in real time, on prices that arrive session by session, and also tests your routine. Both depend on assumptions about fills and costs, which should be written down before you look at the results.

Why are demo account results better than real ones?

For four recurring reasons: limit orders filled on the touch without respecting the queue, omitted costs, stops filled at their level even after a gap, and the absence of emotional pressure. In a hypothetical 20-trade example, the four corrections take the result from +$580 to −$113.

Can you paper trade with a spreadsheet?

Yes. Log the date and time of the plan before the order, the price at which the order would have filled under conservative rules and the actual exit, including gaps at the open. A spreadsheet won’t train you on the platform, so the order drills still belong in your broker’s demo account.

Sources and further reading

Related research
Stop-limit order vs stop order: what each one guaranteesMarket Fieldnotes · 05.10.2026Trading journal: what to record, plus a free templateMarket Fieldnotes · 05.10.2026Swing trading with a full-time job: a practical review routineMarket Fieldnotes · 28.10.2026SMA vs EMA: moving average math, lag and crossoversMarket Fieldnotes · 05.10.2026Famous traders: 10 lessons beyond the profit rankingsMarket Fieldnotes · 05.10.2026 Full archive →
FREE · EVERY SUNDAY

Get the next Radar
before Monday’s open.

The first name, the reasoning, and our public research.

FREE · EVERY SUNDAYGet the watchlist