Demo accounts and portfolios: why returns should be calculated honestly

A demo account does not show your return; it shows the return of flawless execution with no costs, no tax and no fear. An honest calculation is the same trade with everything the simulator gave away for free taken out of it — the spread, slippage, commission, tax on payouts and idle cash — and with the result measured not against zero but against an available alternative. Until that subtraction is made, a paper portfolio measures the quality of the interface, not the quality of the decisions.

What the simulator leaves out

A paper order is filled at the quote you see on the screen. A real order goes into the order book and pays to get in: a purchase is executed at the ask price, a sale at the bid price, and the difference between them stays with the market. The wider the spread and the thinner the order book in a security, the larger this unnoticed expense item becomes — and the harder it hits a strategy that trades frequently.

Next comes order size. If the volume is comparable to what is sitting in the order book, the price moves away from you while the order is being executed, and part of the order may remain unfilled or be filled later and at a worse level. A simulator usually fills the entire volume instantly and at one price. In liquid securities the difference is small; in the third tier it can eat up the whole calculated result — check the depth in the list of stocks for the names you trade on paper.

Costs that turn a gain into zero

On top of the spread come the broker's and the exchange's commissions, the charge for carrying a margin position overnight, and depositary and tariff fees. Each of them is charged on turnover, not on the result, so a high-turnover strategy pays them even in a losing period.

Tax is a separate matter. Payouts on shares and coupons arrive in a demo account in full, but in a real account they arrive after withholding, and the base for the calculation is set by law: 13%. If you reinvest the whole amount of a payout in a paper portfolio, you are reinvesting money you would not have had.

Corporate events break simple arithmetic

After the ex-dividend date the security opens lower — the market subtracts the payout that has left it. A paper portfolio that first received the payout and then recorded a "drawdown" has counted the same event twice, with opposite signs. The same applies to stock splits, conversions and bond amortisation. For the books to balance, events have to be allocated to their dates in advance: the nearest payouts are {{dividend_calendar|limit=5}}, the full schedule is in the dividend calendar, and corporate dates are in the events calendar.

Portfolio return is not the average return of the trades

Percentages on individual trades cannot be added up: a different amount of money stands behind each percentage. A portfolio in which the successful trade was made with a small share of the account and the unsuccessful trade with a large share looks profitable by average percentage and is loss-making in money terms.

There are several correct ways to measure, and they answer different questions. The money-weighted return shows what happened to your capital, and it depends on when you added funds to the account. The time-weighted return answers another question — what quality the decisions showed regardless of the schedule of contributions. If you topped up the account during a decline, the first figure will be higher than the second, and that is not an achievement of the strategy but the consequence of a fortunate contribution schedule. And in both cases the denominator must be the whole account, including the unused balance: money waiting in cash also takes part in the result — as a zero return.

What to compare against

Profit by itself says nothing until the alternative you gave up is placed next to it. The lower bar is the risk-free rate: the yield on a short government issue that you would have earned by doing nothing at all; see the OFZ market and the explainer on how the OFZ market works. The next bar is the broad market including payouts, a role played by an index fund. An active portfolio that has not beaten both bars has done work the market does not pay for.

Risk and the length of the period

Return without drawdown is half of the report, and usually the half that is more pleasant to show. Record the depth of the account's maximum decline and the time it took to return to the previous level: these are the values that determine whether you will be able to stick with the same strategy using real money. And keep the length of the window in mind: over a short stretch the result is determined mostly by noise, and projecting such a result into the future is not a conclusion but a coincidence.

Behaviour is the largest source of divergence

The main thing a simulator does not reproduce is the price of a decision. In a demo account risk is not felt, so positions are taken larger, a loss is sat through more calmly, and the exit rule is followed with ease. A real account changes the position size, the holding period and the willingness to admit a mistake, and therefore changes the result under the very same strategy. The move to real money is best started with a reduced size and the same journal, not with the same amount and new rules.

What to enter in the journal

A minimally useful record: the date and reason for entry, the date and reason for exit, the actual price including the spread, the commission, the tax withheld on payouts, the cash balance and the value of the chosen benchmark on the same dates. The reason for the decision matters more than the price: without it, months later, you cannot tell working logic from a lucky set of circumstances. If the decision relied on an issuer's report, cite the source — financial disclosures — and take the security's parameters from its profile page, for example {{instrument:SBER}}, not from memory.

What this article does not cover

It does not describe how any particular simulator is built: execution rules, the treatment of payouts and the modelling of commissions differ from service to service, and this has to be checked in the documentation of the service you use. For the definition of the term itself, see demo account; the same entry is in the glossary. The platform has no comparable figures on the gap between paper and real results, so this text gives no estimate of "how much a demo overstates" — instead it lists the items you have to subtract yourself.