Leverage and margin — what you are really borrowing, and what it costs
Almost every beginner guide tells you leverage is dangerous because it "multiplies your gains and your losses." That sentence is true in the loosest possible sense and useless in practice, because it points at the wrong dial. A trader risking $50 on a trade risks $50 whether the ticket says 3× or 20×. The leverage number decides something entirely different — who gets to close the position when it goes wrong, you or the exchange — and that is the thing worth understanding before you ever touch the slider.
KEY TAKEAWAYS
- Your risk on a trade equals position size × stop distance. The leverage selector appears nowhere in that formula.
- Leverage sets your liquidation distance: roughly 1 ÷ leverage minus the maintenance margin rate, which is about 4.5% at 20× and 1.5% at 50×.
- Above a certain leverage, your liquidation price sits closer to entry than your stop-loss — meaning the exchange, not your plan, decides the exit.
- In cross margin the number on the ticket is cosmetic. Your real leverage is total notional ÷ account equity, and it applies to every position at once.
- Fees and funding are charged on notional, not on margin, so a 20× position costs 20× the fees for the same amount of your own money at work.
What is leverage actually lending you?
Leverage lets you control a position larger than the cash you put up. The cash you put up is called margin — it is not a payment, it is a deposit the exchange holds as security against your losses. If you post $500 of margin at 10× leverage, you control a $5,000 position. That $5,000 figure is your notional size, and it is the number that every profit, loss, fee and funding payment is calculated from.
Three terms do all the work in this lesson, so here they are in one place:
- Notional size — the full value of the position you control. Profit and loss, trading fees and funding are all a percentage of this.
- Initial margin — what you post to open it. Notional ÷ leverage.
- Maintenance margin — the minimum equity the exchange requires you to keep in the position. Fall below it and you are liquidated. It is quoted as a small percentage of notional, commonly around 0.4–0.5% for the smallest position tiers on major perpetual venues as of August 2026, rising for larger positions. Check your venue's own table; it is public and it varies.
On a perpetual futures venue you are not really borrowing coins from anyone in the way a margin loan works on spot. You are entering a contract whose value tracks the asset, and the exchange holds your margin as collateral against the contract moving against you. The practical difference is that there is nothing to repay — there is only a level at which the exchange takes the position away from you. That level is the whole subject of this lesson. If the word itself is still hazy, the one-page version lives in the glossary: what is leverage.
Does higher leverage mean higher risk?
No — not on its own, and believing otherwise is what makes people mis-size trades in both directions. Your loss if a trade goes wrong is fixed by two things: how big the position is, and how far away your stop sits. Write it out:
Risk in dollars = notional size × stop distance in %
The leverage selector is not in that equation. A $1,000 position with a stop 5% away risks $50 whether you opened it at 3×, 10× or 50×. What changes across those settings is only how much of your cash the exchange holds hostage while the trade runs: $333 at 3×, $100 at 10×, $20 at 50×.
The best-documented case of that behavioural step is ninety years old. Going into the 1929 crash Benjamin Graham — the man who later wrote the textbook on not losing money — held $4.5m of unhedged stock against $2.5m of capital, a shade under 2×. His picks beat the Dow in every one of the next four years and he still lost about 70% of the fund, because every 1% move in that book moved his capital 1.8%. Run the same holdings at 2.6× and the account reaches zero. That is the whole argument for treating the selector as a survival setting rather than a size setting.
So why does leverage have such a fearsome reputation? Because of a behavioural step that has nothing to do with the mechanics. When a trader with $5,000 discovers that 20× lets them post $1,000 of margin and control $20,000, almost nobody keeps their position at $1,000 and enjoys the freed-up cash. They take the $20,000. Leverage does not make people take more risk; it makes larger risk affordable, and then they take it. The danger is real. It just enters through position sizing, which is why sizing the position first and picking the leverage afterwards is the right order of operations — never the reverse.
Where exactly does the exchange close your position?
The worked version of this section, using Binance’s published BTCUSDT bracket rates instead of a rounded 0.5%, is in how to calculate your liquidation price.
Your liquidation price is where your remaining equity in the position falls to the maintenance margin requirement. There is a clean approximation you can do in your head, and it is close enough to make decisions with:
Liquidation distance ≈ (1 ÷ leverage) − maintenance margin rate
The naive version of this — "at 20× a 5% move wipes me out" — is the one most articles print, and it is optimistic. Maintenance margin means the exchange steps in before your margin is fully gone, so the real distance is always shorter than 1 ÷ leverage. Here is the table, using a 0.5% maintenance margin rate:
| Leverage | Margin on a $1,000 position | Naive "1 ÷ leverage" distance | Actual liquidation distance | Liquidation price if entry is $60,000 |
|---|---|---|---|---|
| 2× | $500 | 50% | 49.5% | $30,300 |
| 3× | $333 | 33.3% | 32.8% | $40,320 |
| 5× | $200 | 20% | 19.5% | $48,300 |
| 10× | $100 | 10% | 9.5% | $54,300 |
| 20× | $50 | 5% | 4.5% | $57,300 |
| 50× | $20 | 2% | 1.5% | $59,100 |
| 100× | $10 | 1% | 0.5% | $59,700 |
Long position, isolated margin, 0.5% maintenance margin rate, fees excluded. Trading fees and accrued funding pull the real liquidation slightly closer to entry than shown. Your venue's maintenance rate rises with position size — read its tier table rather than assuming 0.5%.
What does the same trade look like at 3×, 10×, 20× and 50×?
This is the calculation to internalise, because it converts an abstract warning into a specific number you can check on your own ticket. Same account, same idea, same stop — only the leverage selector moves.
Setup: a $5,000 account, risking 1% ($50) on a long in BTC at $60,000, with the stop at $57,000 — 5% below entry, placed where the trade idea is genuinely wrong. Correct position size is $50 ÷ 5% = $1,000 notional, and it stays $1,000 in all four rows. Only the margin and the liquidation change.
| Leverage | Margin posted | Liquidation price | Which exit comes first | What the trade actually costs you |
|---|---|---|---|---|
| 3× | $333 | $40,320 | Your stop, with 27% to spare | $50, exactly as planned |
| 10× | $100 | $54,300 | Your stop, with 4.5% to spare | $50, exactly as planned |
| 20× | $50 | $57,300 | Liquidation, $300 above your stop | $50 of margin plus a liquidation fee — and no control over the fill |
| 50× | $20 | $59,100 | Liquidation, on a 1.5% wobble | $20 — closed out by ordinary noise while the idea was still alive |
Read the last column carefully, because there is something in it that most warnings about leverage get backwards. At 50× the dollar loss is smaller, not larger — $20 instead of $50. Isolated margin caps what a position can cost you at the margin you gave it, so cranking leverage while holding notional constant actually reduces your maximum loss per trade. That is arithmetic, not a loophole.
The trap is what it costs you in outcomes. At 50× you were removed from a correct trade by 1.5% of ordinary movement — a range BTC covers on most days without meaning anything. Do that four times in a week and you have lost $80 and taken zero of the four trades you actually analysed, while the 10× version of you took all four at the intended risk. Leverage that sits inside your stop does not blow up your account in one go. It quietly converts your strategy into a coin flip on noise, and hands you the bill in instalments.
It is worth seeing the same arithmetic run by someone who was very good at this. In 1994 Michael Steinhardt held a reported $30 billion of European bonds against $4.6 billion of capital — 6.5× leverage. The bonds moved roughly 5% against him and it took about a third of the fund. Put that 5% next to the box on your own order screen: at 20× a 5% move is not a third of the account, it is all of it. The full profile, with the numbers worked through, is the clearest case we know of leverage turning a small forecasting error into a career-ending one.
Isolated or cross margin — which should a beginner pick?
Isolated. Then read the rest of this section anyway, because the reason matters more than the answer.
Isolated margin assigns a fixed amount of collateral to one position and ring-fences it. If that position liquidates, you lose the assigned margin and nothing else; the rest of the account is untouched. The cost is that liquidation arrives sooner, because only the assigned margin absorbs the loss.
Cross margin uses your entire available balance as collateral for every open position. Liquidation moves much further away — often so far it feels like it cannot happen. The cost is that when it does happen, there is no ring fence. Everything is collateral, so everything is available to be lost.
Here is the part almost nobody says out loud. In cross margin, the leverage number on the ticket is close to meaningless. Set every position to 5× if you like — the exchange is still backing all of them with one shared pot, so the leverage that actually determines your account-wide liquidation is:
Real leverage = total notional across all positions ÷ account equity
Run that on a $5,000 account and the picture changes fast:
| Total notional open | Real leverage | Adverse move that empties the account |
|---|---|---|
| $5,000 | 1× | 99.5% |
| $10,000 | 2× | 49.5% |
| $25,000 | 5× | 19.5% |
| $50,000 | 10× | 9.5% |
| $100,000 | 20× | 4.5% |
$5,000 equity, cross margin, 0.5% maintenance margin, correlated positions moving together. Assumes positions are in the same direction — hedged positions net off and behave very differently.
Five separate $20,000 positions each labelled "5×" on the ticket is a $100,000 book against $5,000 of equity. A 4.5% move against you and the account is gone — and in crypto, correlated assets fall together, so five positions in five different alts is one position wearing five hats. This is the mechanism behind the market-wide events described in the anatomy of a liquidation cascade, seen from the inside of one account.
What does leverage cost you even when the trade works?
Every cost on a derivatives venue is charged on notional, never on your margin. That single fact is why leverage is expensive in a way the marketing never mentions.
Compare two traders with the same $5,000 account and the same $1,000 of their own money committed per trade. Trader A trades $1,000 notional. Trader B selects 20× and turns that $1,000 of margin into $20,000 notional. Taker fees of 0.05% per side are a typical retail tier on major venues as of August 2026 — check yours, the tiers differ.
| Trader A — $1,000 notional | Trader B — $20,000 notional at 20× | |
|---|---|---|
| Fee per round trip (0.05% × 2) | $1.00 | $20.00 |
| 20 round trips in a month | $20 | $400 |
| As a share of the $5,000 account | 0.4% per month | 8% per month |
| Funding, 5-day hold at 0.01% per 8h | $1.50 | $30 |
| One 5-day trade: fees + funding | $2.50 | $50 |
| Move needed just to break even | 0.25% of price | 0.25% of price |
Funding uses a 0.01% baseline per 8-hour period — the level perpetual funding tends to sit near when the market is balanced. It goes several times higher when positioning is one-sided, and it can go negative. Read the live rate before you assume, or run it through the funding rate calculator.
The breakeven row is the one that repays study. Both traders need the same 0.25% price move to cover costs, because both cost and profit scale with notional. That is why "leverage doesn't change your edge" is technically true. But look at what B is spending to stay in the game: $400 a month in fees against a $5,000 account is 96% of the account per year in frictional cost. A strategy that is genuinely profitable before costs can be dead after them, and the trader will conclude their analysis was wrong when the analysis was fine and the arithmetic was doing the killing. Every extra turn of leverage is another turn of the fee meter. For a fuller version of that argument, the order-type lesson works through what taker fees cost across twenty setups.
There is a version of this that is not a choice at all. Fix your risk per trade and your stop, and the position size — and therefore the leverage — is already decided: risking 1% of $10,000 behind a 0.15% stop requires $66,667 of notional, or 6.67× the account, while the same rule behind a 9.60% stop requires $1,042. That is why fast trading styles come with leverage attached and slow ones do not. Scalp, day, swing or position works through what each style costs once fees and funding are both counted.
When is "just use low leverage" the wrong advice?
It is the right default and it is not a law. Three cases where it misleads:
- Low leverage in cross margin can be more dangerous than high leverage in isolated. A 3× ticket on a cross-margin account holding six correlated positions is exposed to the whole balance. A 20× isolated position sized at $1,000 can cost you $50. The dial is not the risk; the table above is.
- Capital efficiency is a legitimate reason professionals run higher leverage. If your stop is genuinely tight — 1% on an intraday setup — then 10× still leaves your liquidation at 9.5%, nine times further away than your stop. Holding $200 of margin instead of $1,000 to run that trade is not recklessness; the other $800 is earning yield or funding another idea. The rule is not "leverage low," it is "keep liquidation at least three times further from entry than your stop."
- The advice assumes a flat maintenance margin rate, and yours is not flat. Every major venue raises the maintenance requirement as position size grows through tiers. A position large enough to sit two tiers up may face 1% or more, which pulls liquidation meaningfully closer than the 0.5% used throughout this lesson. If you size up, re-read the tier table — do not reuse the buffer you calculated when you were smaller.
The one thing to do after reading this lesson: open a position ticket on your exchange — no need to submit it — enter the size you would actually trade, and compare the liquidation price it shows you against where your stop-loss would go. If the liquidation is nearer, lower the leverage until it is at least three times further away. That single check, done before every trade, removes the entire failure mode this lesson describes.
What do people get wrong at this stage?
Picking leverage before position size. Size comes from your risk and your stop; leverage is whatever setting keeps liquidation out of the way afterwards. Doing it in the other order is how a $50 risk becomes a $500 one. Using 1 ÷ leverage as the liquidation distance. Maintenance margin means the exchange closes you earlier than that, and fees and funding pull it earlier still. Leaving the account in cross margin because liquidation looks far away. It is far away right up until it isn't, and then it takes everything rather than one position. Treating a leverage increase as free because "I'll still use a stop." Above the crossover in the table, the stop is never reached. Ignoring funding on multi-day holds. At high notional it is a rent payment on a position that may not be going anywhere. Adding margin to a losing position to push the liquidation away. This converts a defined loss into an undefined one and is the single most reliable way to turn a bad trade into a bad month. Assuming the maintenance rate stays constant as you scale. It does not, and the buffer you verified at small size quietly shrinks as you grow.
FAQ
Does higher leverage mean higher risk? Not by itself. Risk on a trade is position size × stop distance, and leverage appears in neither. What leverage changes is your liquidation distance. It becomes dangerous when the extra buying power tempts you into a larger position — which is a sizing decision, not a leverage one.
At what leverage does liquidation happen before my stop-loss? Whenever liquidation distance (roughly 1 ÷ leverage − maintenance margin rate) is smaller than your stop distance. With a 0.5% maintenance rate that is about 9.5% at 10×, 4.5% at 20× and 1.5% at 50×. A 5% stop is safe at 10× and already useless at 20×.
Should a beginner use isolated or cross margin? Isolated, almost always — it caps what one position can cost you. Cross margin pushes liquidation further away but makes your whole balance collateral, and in cross the leverage on the ticket is cosmetic: your real leverage is total notional ÷ account equity.
How much does leverage actually cost in fees and funding? Both are charged on notional. At 0.05% taker, a $1,000 position costs $1 per round trip and a $20,000 position costs $20. Add 0.01% funding every eight hours and a five-day hold on $20,000 costs another $30 — $50 of cost against $1,000 of posted margin before price does anything.
See where your liquidation actually sits
The distance between your stop and the exchange’s liquidation level is the number this lesson turns on — and it only exists on a real order ticket.
Compare the three exchanges we use →
That page carries referral links and says so up front — including what each venue does badly.
Keep the whole course next to your charts
The whole slide course — ten free PDF parts, 351 pages.