Home › How-to › Paper trading

How to paper trade crypto

By TradingPrimer · Published 31 Aug 2026

Almost everyone skips this step, and the reason is understandable: practice feels like waiting. But the specific thing that ruins most first accounts is not a bad strategy — it is not yet knowing which button does what, while real money is on the line. Practice moves that discovery somewhere it costs nothing.

What does paper trading actually teach you?

It teaches the mechanical half of trading, which is larger than beginners expect. Order types, which market you are actually in, how size follows from risk, where a stop belongs, and what your own record looks like when you write it down honestly.

It also tests one thing nothing else can test cheaply: whether you can follow a written plan when nothing is forcing you to. If you cannot follow your rules with fake money, you will not suddenly start with real money — the pressure only goes up.

What it cannot teach is the part everyone wants to skip to. A simulated loss does not feel like a loss. It does not wake you at 3am, and it does not tempt you to widen a stop to avoid admitting you were wrong. So treat a good paper record as evidence that your process works, never as a forecast of your returns.

How do you set up a practice account that tells you the truth?

Two settings decide whether your practice is useful or misleading.

Start with a realistic balance. Use roughly the amount you actually expect to fund. A $100,000 practice account teaches nothing if you plan to start with $2,000, because every position size you rehearse will be one you can never place. The practice arena on this site uses a $10,000 virtual balance for exactly this reason.

Fix your risk per trade and do not touch it. One percent per trade is the convention, and the point is less the number than the fact that it never moves. That fixed rule is the thing you are here to make automatic.

Then define the test before you start, not while you are in it: one market, one setup, and a fixed sample — say 30 completed trades. A fixed sample stops you quitting after a lucky run or a painful one, which is the most common way practice gets abandoned right at the point it starts working.

Where do you get a practice account, and what does each one give you?

Every large exchange now runs a practice mode inside the normal platform rather than on a separate site. Binance and Bybit call it Demo Trading; OKX calls it Demo trading. All three are free, none of them asks for a deposit, and both Binance and Bybit state on their help pages that identity verification is not required to use one. What none of them lets you choose is the setting that decides whether your practice is honest: the starting balance.

Two of the three do not tell you the opening balance in advance — you find out once you are inside. On all three, the control that takes you back to real money sits in the top corner of the same screen you are practising on, which is either reassuring or alarming depending on how well the practice went. On the desktop web, Binance and OKX both put the entrance under Trade; Bybit puts it under the profile icon at the top right.

Button names read from each exchange's own public help pages, not from a logged-in account: Binance help pages as of 6 Sep 2026, OKX help page updated 4 Sep 2026, Bybit demo FAQ last updated 25 Nov 2025. Exchanges rename controls often — check the current page before you go looking for a button.

Read the middle column again with the first rule of this page in mind. The one platform that publishes a figure hands you 50,000 USDT plus three other assets, and the other two do not tell you in advance at all. So the default practice account is already breaking the rule before you place a trade — and the damage is arithmetic, not opinion.

Twenty-five times. Every stop distance you get used to, every size you type, every position that feels normal on the screen — all of it belongs to an account you will not have. That is this page's first mistake, trading a size you could never fund, arriving as the platform default rather than as your error.

The fix costs nothing and takes one decision. Set your risk in dollars of the account you intend to open, not as a percentage of whatever the demo credited you. Planning to fund $2,000 at 1%? Then $20 is your risk per trade on the demo too, and the rest of the balance is scenery. On a 50,000 USDT demo that means treating 4% of the balance as real and ignoring the other 96%.

Every platform gives you a button that a real account does not have. On Binance and OKX the demo balance can be reset outright — on OKX you close all positions and pending orders, then Assets → Reset. Bybit says the demo account cannot be reset, but it will top you up: Request Demo Funds becomes available once total equity falls below 10,000 USDT. That threshold is worth doing the arithmetic on. Starting from the 50,000 USDT line, you would have to lose 40,000 of it — eighty consecutive losses at $500 each — before the button unlocks. It was not built for anyone risking 1% a trade. It was built for people who blow the account up, which tells you what the platform expects most demo users to do.

And your record can disappear. Bybit's FAQ states that if you do not access the demo account for 30 days, it is refreshed and its data is cleared. A 30-trade sample taken at a few trades a week can straddle that gap. Keep the journal outside the platform — a spreadsheet, or the trading journal on this site — and the exchange can wipe whatever it likes.

Where this advice is wrong: if you are practising purely to learn which button does what, the starting balance genuinely does not matter and you can skip this whole section. It matters the moment you start drawing conclusions from the results, because a result produced at 25 times your real size is not a result about you.

How do you work out the size of a practice trade?

Size is not something you choose. It is something you calculate, and it is the last step, not the first. The order is always: risk budget, then stop, then size.

Take the trade in the picture above. The account is $10,000 and the risk rule is 1%, so the most this trade may lose is $10,000 × 0.01 = $100. The entry is $2,000 and the level that would prove the idea wrong is $1,950, so the risk per coin is $2,000 − $1,950 = $50.

Size falls out of those two numbers: $100 ÷ $50 = 2.0 ETH. That is a position worth 2.0 × $2,000 = $4,000 — forty percent of the account — controlled by a rule that can only lose one percent of it. This is the sentence worth keeping: position value and risk are different things, and confusing them is what makes beginners think a $4,000 position is reckless or a $100 risk is trivial.

Notice what happens if you move the stop instead of the size. Widening the stop to $1,900 doubles the risk per coin to $100, so the size must halve to 1.0 ETH to keep the same $100 budget. If you widen the stop and keep the size, you have quietly doubled your risk — the single most common unforced error in this whole process.

Why do paper results always look better than live results?

Because a simulator usually fills you at the price you asked for, and a real market often does not. Your stop is an instruction to sell once a level trades, not a promise to sell at it.

The exchanges say so themselves, in the same help pages that advertise the feature. Bybit explains that demo orders do not enter the actual order book, so your fill never had to get past anybody. Binance warns of discrepancies in chart data, order book pricing and trade order execution. OKX goes furthest: its demo help page states that price movements in demo mode have no relation to live market conditions — while the FAQ lower down the same page says demo simulates real-world conditions. When an operator’s own page contradicts itself on the point, the safe reading is the narrow one: a demo rehearses procedure faithfully and proves nothing about whether a strategy makes money.

Run that difference forward and it stops looking small. Suppose your stop fills $2.80 below the level. On 2.0 ETH that is 2.0 × $2.80 = $5.60 extra, so a planned $100 loss actually costs $105.60 — 5.6% worse than planned. Here is what that does across a losing streak, holding the risk fixed at $100 per trade:

A 10-trade losing streak you had designed to cost 10% of the account actually costs 10.56%. That is not catastrophic on its own, and it is deliberately a modest example. The part that matters is where the error lands: slippage widens when markets move fast and the book thins out, which is exactly when losing streaks happen. So the error is not random noise that averages away — it clusters into your worst weeks.

The practical fix is not to distrust practice. It is to assume your live results will be somewhat worse than your paper results, and to size as though that is true. Run your own numbers with the position size calculator before you decide the plan is safe.

Where this advice is wrong: if you trade a liquid market on a higher timeframe with a wide stop, slippage of a few dollars is a rounding error and this whole section barely applies to you. It matters most in the opposite case — tight stops, small timeframes, thin coins.

What should you record after every practice trade?

The instrument, entry, stop, size, planned risk, actual fill, exit reason and result — and one line on what you noticed. Then one more field that most journals leave out, and which is the whole point of practising.

Grade every trade one of three ways: followed plan, minor deviation, or major deviation. A minor deviation is entering a little late or taking profit slightly early. A major deviation is moving a stop, adding to a loser, or trading a setup that was not in your plan at all.

The picture above shows 7, 2 and 1 out of ten — a 70% rule-following rate. That sounds like a decent school grade, and it is not one. Three trades in ten went off-plan, and off-plan trades are precisely the ones that are not capped at 1%, because the rule being broken is usually the stop. A record can show a profit and still be a failed test. Keep the journal honest: never delete a bad trade, never reset the balance after a bad run, never top up imaginary capital.

How do you know when practice is finished?

Not by the calendar. Move on when all four of these are true across your fixed sample:

1. You can operate the platform without thinking. Right market, right side, right size, no mis-clicks.
2. Every size came from the formula. Not from a round number that felt comfortable.
3. Your rule-following rate is high and honest. Aim well above the 70% in the example, with zero major deviations in the recent stretch.
4. The record survives a haircut. Recalculate assuming every exit filled slightly worse. If the result only works with perfect fills, it does not work.

Then go live small — far smaller than feels worth it. The first live phase is not for making money; it is to meet the one variable practice could not simulate, which is how you behave when the loss is real. If your behaviour changes, that is useful information, not failure. Go back to practice or cut size further.

What are the most common paper-trading mistakes?

Trading a size you could never fund. Practising with $100,000 when you will start with $2,000 rehearses a skill you will not get to use.

Changing the rules after every loss. If the plan moves each time it is tested, there is no plan to evaluate — only a sequence of unrelated trades.

Deleting the embarrassing trades — or letting the platform delete them for you. The bad trades are the data. A record with the losses removed cannot show you your real drawdown, which is the number that decides whether you can survive live trading. Every demo hands you the eraser: Assets → Reset on Binance and OKX, Request Demo Funds on Bybit. A real account has no such button, which is the whole reason a real drawdown teaches you something.

Assuming every limit order fills. If price touches your limit and you mark it filled, you are inventing entries you would not have got. Price has to trade through your level with enough volume to clear the queue ahead of you — see the order book.

Judging by the profit column. Over 30 trades, profit is mostly noise. Rule-following, planned-versus-actual risk, and worst drawdown are the numbers that predict anything.

Practising forever. The opposite failure, and a real one. Practice is a stage with an exit condition, not a hiding place.

FAQ

Is paper trading actually risk-free? No money is at risk, but bad habits are free to form — especially overconfidence from unrealistic fills. The risk is to your future account, not your current one.

How long should I paper trade? Use a sample, not a calendar: around 30 completed trades, judged on execution quality rather than profit. Stopping because a month elapsed tells you nothing.

Should I paper trade with leverage? Practise the way you intend to start, and beginners should start on spot. Leverage changes the arithmetic and adds liquidation, which a gentle simulator may not model faithfully.

My paper account is profitable — am I ready? Profit over a small sample is weak evidence. Check the four conditions above instead, particularly whether the result still holds once you assume every exit filled slightly worse than planned.

Which exchange’s demo account should I use? Whichever one you actually intend to trade on. Half of what practice buys you is muscle memory for one specific screen, and that does not transfer between platforms. The feature differences matter far less than the one thing you must override on all of them: set your risk in dollars of the account you plan to fund, not as a percentage of the balance the demo credited you.

Every figure on this page is worked from the single practice trade shown at the top ($10,000 balance, 1% risk, entry $2,000, stop $1,950) and can be reproduced by hand. The $2.80 slippage figure is an illustrative assumption used to show the mechanism, not a measured market average. The exchange demo figures and button names come from each platform’s own public help pages, dated in the table above; nothing here required a logged-in account. Published 31 Aug 2026, updated 12 Sep 2026.

← All how-to guides