
The coin has been climbing for three days, and this morning the tape turned red, while a sell-side density level sits above the price in the order book and hasn't moved for half an hour. You don't feel like going long, you feel like selling. The only problem is there's nothing to sell, the coin isn't on your balance. That's exactly where the question of how to short crypto begins.
When people search "short crypto", they're usually looking for a button. The button exists, on Binance and Bybit it's called Sell/Short, and finding it is easier than understanding what happens after you press it. What are you selling if you don't own it? How much does it cost to hold the position? Where do you get an argument for the entry? And where does the stop go if the price can multiply, while the execution of a protective order depends on the market.
We'll break down the mechanics, two ways to open a position, step-by-step instructions for Binance and Bybit, three entry scenarios with calculations, and the mistakes that blow up deposits most often. The examples are for learning purposes, the numbers are hypothetical, and the fees are set for the calculations; check current rates in your own account.
A short is a trade where you sell an asset first and buy it back later. You make a profit if the price went down between the sale and the buyback.
How do you sell something you don't have? It depends on the instrument. In margin trading, the exchange lends you the coin, you sell borrowed coins and commit to returning the same number of coins. On perpetual futures, nobody lends anything: you open a contract on a price decline, no coin changes hands, and on linear USDT contracts the result is calculated and settled in USDT.
The result of a short position on a linear USDT contract fits into one line.
Short result = (Entry price - Exit price) × Size in coins - Fees ± Funding
For a margin short on spot, the formula has loan interest instead of funding, always with a minus sign. The asymmetry with a long kicks in when the price moves far.
The percentages are calculated from the position's notional at entry, excluding fees and funding. An unleveraged long can lose at most what was put in, while a short's loss has no upper limit. The table shows the theoretical result: a real position can be liquidated earlier if the collateral runs out. Liquidation is the forced closing of a position by the exchange, and the higher the leverage, the closer to the entry price it sits. That's why a stop in a short isn't about discipline, it's about arithmetic.
Closing a short is a buy, so when thousands of short sellers exit at once or get liquidated, the price flies up through an empty order book. That's how a short squeeze works.
The two ways to open a short position solve the same task with different money.
For scalping, traders often use futures: you don't need to borrow the coin separately, stop and take-profit orders are available, and a position closed before the settlement moment neither pays nor receives funding at all. Perpetual contracts and the mark price are covered in "Crypto Futures: What They Are and How to Trade Them".
A margin short makes sense in three cases. The coin has no perpetual contract. You're hedging a coin you hold on spot. Or the contract's funding rate has gone deep negative, short sellers are paying longs 0.3% per settlement, and borrowing the coin at interest turns out cheaper. More in "Crypto Margin Trading: How It Works and How It Differs From Futures".
Learning example 1. Margin short on BTC spot, hypothetical numbers.
You borrow 0.1 BTC at 68,000 and sell it immediately, 6,800 USDT lands in your account. A day later the price is 66,640, down 2%. You buy back 0.1 BTC for 6,664 USDT and repay the debt. The difference is 136 USDT. The spot fee is 0.1% per trade, 6.80 plus 6.66, 13.46 in total. Loan interest at a hypothetical rate of 0.0015% per hour comes to 0.036% of 0.1 BTC over the day, that's 0.000036 BTC, about 2.4 dollars. Net result is roughly 120 USDT.
The mirror scenario: the price rose by the same 2%. Minus 136 on price, minus 13.74 in fees, minus 2.5 for the loan, around minus 152. And the detail that catches beginners: the debt and interest are counted in BTC, you have to return exactly 0.1 BTC plus interest, no matter what they're worth at the moment of repayment.
![[Placeholder: screenshot of the order book in the terminal with a sell-side density level above the price and an order lifetime timer]](https://api.secret-terminal.com/uploads/densitys_2_efbda1952c.png)
If futures, margin and leverage still feel hard, start with the free lesson from the crypto trading course on the Secret Terminal YouTube channel. It covers exchanges and futures from scratch, and the lesson is part of a full five-part course for beginners.
On Binance, you can open a short position in two ways. Through USDⓈ-M futures, where the trade takes a few clicks, and through margin trading, where you need to borrow the coin before selling. Let's start with futures.
Two settings are done once. Transfer USDT to your futures account and check the position mode. In One-way mode, you hold either a long or a short per contract. Hedge Mode is enabled in settings via Position Mode, but the exchange won't let you switch modes with open positions or active orders.
About money. The learning calculation below assumes fees of 0.02% maker and 0.05% taker. Check your fee tier, discounts and instrument availability in your account. Funding is a periodic payment between holders of longs and shorts that keeps the contract price close to spot. Only those with an open position at the settlement moment pay or receive it, nothing accrues in the hours between settlements. Many contracts have an 8-hour interval, but the exchange can change it. Watch the timer for the specific contract. The exchange takes nothing from funding.
Now the margin short. Transfer collateral to your margin wallet, open Trade, then Margin, choose the pair and the Sell tab. Turn on Borrow mode, the exchange will borrow the coin for your order automatically, enter the amount and press Margin Sell. You close it with a reverse buy in Repay mode, and the purchased coin repays the debt. Interest is charged every hour in the coin you borrowed. Check the margin call and liquidation thresholds for your mode.
Same logic, slightly different interface. Shorting crypto on Bybit without surprises starts with checking the margin mode, the leverage, and One-Way or Hedge Mode.
For the example below we use fees of 0.02% maker and 0.055% taker. The funding payment is calculated the same way as on Binance: the position value at the mark price is multiplied by the rate, only those holding a position at the settlement moment pay or receive it, and the interval depends on the contract and can change. A short receives with a positive rate and pays with a negative one.
The examples assume maker execution of limit orders. Without Post-Only, a limit order can become a taker order, which will change the calculations.
The button exists, you need an argument. Below are three scenarios that cover most of the question of how to short crypto with a calculation rather than a "too expensive" gut feeling. Below is a hypothetical calculation scheme for the futures examples.
Short stop = Entry price + ATR × 1.5, then shift beyond the nearest density level in the order book
Costs per coin = Entry price × entry fee + Stop price × exit fee
Size (in coins) = Risk in $ / (Stop distance × k + Costs per coin)
k - a hypothetical slippage buffer: 1.2 for BTC and 1.3 for an altcoin in the examples
In the examples, risk is 0.5% of the deposit. ATR × 1.5 and the k coefficient are set for illustration, they aren't universal settings. The buffer may turn out to be insufficient. Notional divided by deposit gives effective leverage; the exchange slider separately determines the initial margin.
The classic. The price approaches from below a level that has already refused to let it go higher, and you sell from it with a stop beyond the level.
The problem with the classic is that everyone sees the level on the chart, and some of the touches end in a breakout. The filter comes from the order book and the tape. A density level in the order book is a large limit order or a cluster of orders at one price. The level is tradable if a density level sits above it and holds for minutes rather than seconds, and if aggressive buys on the tape fade as the price approaches. The tape is flying green and the density level is shrinking before your eyes? The breakout risk is growing, although the outcome isn't known yet.
Learning example 2. BTCUSDT Perpetual on Binance, 5-minute chart, hypothetical numbers.
Price is 68,380, resistance at 68,400 has rejected the price twice this session. In the order book at 68,420 there's an order for 3.1 BTC, the lifetime timer shows 25 minutes. ATR(14) on 5m is 90.
Entry with a limit order at 68,405, below the sellers' density level. The calculated stop at ATR × 1.5 gives 68,540, the next density level above sits at 68,560, so the stop goes beyond it, to 68,575. Distance is 170 points.
Deposit 10,000, risk 0.5%, which is 50 dollars. Costs per coin (68,405 × 0.0002 plus 68,575 × 0.0005) are about 48 dollars. Size = 50 / (170 × 1.2 + 48) = 50 / 252 = 0.198 BTC, rounded down to 0.19. Notional is about 13,000 dollars, leverage relative to the deposit is 1.3x.
Take-profit at 68,050, where the nearest buy-side density level sits according to the density map. Distance is 355 points, the risk-to-reward ratio on the chart is 1 to 2.1. Now the money. At the stop, 0.19 × 170 = 32.3 dollars of loss on price, plus 2.6 in entry fees and 6.5 for a market exit, about 41.4 in total, so there's still room for slippage within 50 dollars. At the take-profit with a limit order, 0.19 × 355 = 67.5 minus 5.2 in fees, about 62 net. In money terms the ratio comes out to 1 to 1.5, not 1 to 2.1, and that's exactly the difference most people don't calculate.
When it doesn't work. The density level gets eaten by market buys, the rest of the order is pulled, and the tape doesn't fade, it accelerates. If filled at the calculated price, the loss is about 41 dollars; slippage can increase it. The only mistake would be moving the stop higher or adding to the position because it's "too expensive" anyway.
In trader chats the scenario sounds like this: the order book above the price is empty, there are no sellers, so nobody is defending the level, let's short. In reality it's the other way around.
An empty order book above means the buyer has nothing to eat through. A series of market buys will pass a dozen ticks without resistance, and the price will fly on laughable volume. There's nothing to lean on and nowhere to hide the stop. An empty order book above is an argument against entry, not for it. In this scenario, without a significant order above the price, there's no support for the entry. You want to sell, but you'll have nothing to protect the position with.
A working order book short is built on three layers, and all three have to line up.
A separate signal that breaks the setup. The density level above was pulled before the price reached it, within seconds, without being filled. Why it was removed, the order book won't tell you, but the support is gone, and the short is canceled even if the tape is still red.
Learning example 3. A top-50 altcoin on Bybit, hypothetical numbers.
Price is 0.4520. At 0.4560 there's a sell-side density level of about 1.2 million dollars, timer at 35 minutes, and the density map shows it on three exchanges. Below the price, the nearest large buy is at 0.4400, about 600 thousand. The tape slows down near 0.4550, and the cluster at 0.4555 printed three times the average volume on a two-tick move. The buyer is being absorbed.
Entry with a limit order at 0.4552. Stop beyond the density level at 0.4585, distance 0.0033, which is 0.72%. The k buffer for an altcoin is 1.3. Deposit 12,000, risk 0.5%, which is 60 dollars. Costs per coin (0.4552 × 0.0002 plus 0.4585 × 0.00055) are about 0.00034. Size = 60 / (0.0033 × 1.3 + 0.00034) = 60 / 0.00463 = 12,950 coins, rounded down to 12,900. Notional is about 5,870 dollars, leverage relative to the deposit is about 0.5x.
The target is split into two parts. Half at 0.4470, roughly halfway to support, the other half at 0.4410, in front of the buy-side density level, not beyond it. At an average exit price of 0.4440, the ratio is 1 to 3.4 excluding fees and slippage. A market stop without slippage costs about 47 dollars, with the assumed buffer the calculation is about 60 dollars, and the actual loss can be higher.
We calculate funding for the record and with a caveat: the exchange uses the position value at the mark price at the settlement moment, not the entry notional. If the position value at settlement is 5,870 USDT, at a rate of +0.03% the short receives 5,870 × 0.0003, less than two dollars. At a rate of -0.3% the same position would pay about 17.6 dollars per settlement, almost a third of the trade's risk. This payment has to be included in the risk before carrying the position through a settlement.
In Secret Terminal all three layers sit in one window. The order lifetime timer is right next to the size, the density map pulls large orders from Binance, Bybit, OKX, MEXC and WhiteBIT into one table, the tape shows whether the buyer is fading, clusters show absorption, and the funding rate with a countdown to settlement sits in the order book header.
![[Placeholder: screenshot of the terminal before a short entry: order book with a sellers' density level and order lifetime timer, the tape and a cluster showing absorption]](https://api.secret-terminal.com/uploads/work_setup_5a921a53cc.png)
News gives a short two different windows, and mixing them up is expensive.
The first window is negative news. A protocol hack, a delisting, a large unlock, US macro data worse than forecast. Catching the first candle with a market order is pointless: market makers pull their quotes, the spread widens, and the sale fills at the worst prices of the move. The entry point comes later, on the pullback to the broken level, when the tape is sluggish and a sellers' density level sits in the order book on the retest. You short the second push, not the first.
The second window is positive news that didn't play out. A listing, a partnership, inclusion in an index. The price takes off, open interest (the total volume of unclosed contracts) rises, and twenty minutes later the momentum runs out and the high isn't renewed. The high of the news candle becomes resistance, and the argument is a weakening tape while OI has grown. Just remember that OI only shows the fact that positions were opened, not who holds them or with what leverage, so the stop still goes beyond the high.
The calendar matters more than ratios. Before US macro data releases, decide in advance whether to close the position or cut the risk. A stop near the entry price doesn't guarantee breakeven. Slippage at the moment of the release can eat more than the move gives.
Two mistakes cost the most, and the rest usually stem from them.
The most common mistake looks like common sense. The coin is up 40% in a week, "this can't go on", the trader sells. It goes on. Then again.
Overbought on an oscillator doesn't confirm a reversal. Rising OI means more open contracts, and a positive funding rate shows the direction of payments, but it doesn't predict the price. Falling OI at a new high and a weak tape deserve attention; a short still needs its own signal and a risk calculation.
When vulnerable shorts pile up, a rally can trigger a chain of liquidations and a short squeeze, and its mechanics are covered in "Short Squeeze in Crypto: Mechanics and How to Trade It". How lopsided the crowd is shows up in the long/short ratio, and the metric is explained in "Long vs Short: How to Read the Position Ratio". Size is calculated from risk; a trade against a strong trend can be skipped.
A short without a stop is exposed to a sharp price rise. How fast the losses pile up depends on position size, leverage and the market move.
Liquidation doesn't replace a planned stop. Its level depends on the mode, collateral, maintenance margin and fees. For an isolated linear short with minimum initial margin, 100% / leverage gives only a rough guide to the distance, usually an overestimate. Use the exchange calculator and the liquidation price it shows.
Four rules that settle the topic.
Three more mistakes, smaller but regular. Adding to a losing short during a rally, which means increasing the position exactly when the distance to liquidation is shrinking. Entering at market during a spike with a widened spread. And an accidental flip into a long, when a closing buy in One-way mode is placed without Reduce-Only for a size larger than the position.
The order book, limit orders and entry points, which is exactly the part where stops end up in the wrong place, are covered in the free lesson of the same course.
What is shorting crypto in simple terms?
A short is a trade on a price decline. You sell the asset higher first, then buy it back lower, and the difference is yours. On perpetual futures you don't need the coin for this, you open a contract on a decline, while in margin trading the coin is lent to you at interest, and after the buyback you have to return it.
Can you short crypto without leverage?
You can keep the position's notional no higher than your own capital, for example by using 1x on linear futures. That doesn't rule out liquidation, and on margin the loan and interest still apply. Size is first calculated from risk, stop and costs.
How much can you lose on a short?
In isolated futures mode, the allocated collateral is at risk; adding margin manually and auto-add margin increase it. On cross margin, the entire futures account is exposed. A short's theoretical loss is unlimited because the price can rise indefinitely, so a stop-market order is placed at the moment of entry.
How do you short on Binance without futures?
Through margin trading. You transfer collateral to the margin wallet, turn on Borrow mode in the Sell tab, the exchange borrows the coin for the order, you sell it, and you close with a reverse buy in Repay mode. Available leverage depends on the pair and mode, and loan interest is charged in the borrowed coin.
Does a short pay funding?
It depends on the sign of the rate. With positive funding, longs pay and the short receives a payment; with negative funding, the short pays. The interval is often 8 hours, but it depends on the contract and can change. You only pay if you hold the position at the settlement moment.
What leverage should you use for a short?
First calculate size from risk, stop and costs. In the examples, notional relative to the deposit gives 1.3x and about 0.5x; that's effective leverage, not the slider value. With a fixed size, raising leverage reduces the initial margin and brings liquidation closer, but on its own it doesn't change the loss from the price move.
Short with Secret Terminal. The sellers' density level is visible with an order lifetime timer and on the density map across Binance, Bybit, OKX, MEXC and WhiteBIT, the tape shows whether the buyer is fading, clusters show absorption, and the funding rate with a countdown to settlement sits in the order book header. Before trading, check your terminal settings, closing orders and access to your exchange account.

Has 5 years of trading experience and spent 3 years as a mentor, training over 2,000 students. He is developing Secret Terminal to make professional trading tools accessible to every trader.
Was helpful
Your rating will help us improve the quality of published materials and increase their usefulness.
We publish product updates, setup guides, and practical materials on working with Secret Terminal tools

What cryptocurrency futures are, how perpetual contracts work, and how they differ from spot trading. A complete breakdo...

How to use the long/short ratio in crypto to analyze market sentiment and spot reversal points. We break down the connec...

What a short squeeze is, how to spot it and how to trade it.