Why your exchange caps the leverage a call asks for

A channel posts a long with an entry, a stop, a target and one more line: leverage 20x. Two people follow it. The first has a small account and opens a position worth a few thousand dollars at 20x with no friction at all. The second wants the same trade at a size that matters to them, sets the same 20x, and the exchange either refuses the order or requires far more margin than the arithmetic suggested. Nothing was misconfigured. The two accounts were never going to get the same trade, because on major derivatives venues the leverage you are allowed is a function of how large the position is.

This is a mechanic worth understanding before you subscribe to anyone who publishes leverage numbers, because it is one of the few places where the gap between the call and your fill has nothing to do with your discipline.

Why does the exchange decide your leverage instead of you?

Because leverage on a perpetual futures venue is not a slider you own. It is a ceiling the venue sets against the notional value of the position, and the ceiling comes down as the position goes up.

Binance states it plainly in its help centre article on the leverage and margin of its USDT-margined futures: "The maximum amount of leverage available depends on the notional value of your position — larger positions allow for lower leverage." The same page describes how the margin that keeps a position alive is computed: "Maintenance Margin calculations are done via a 'Bracket' setup and are always calculated the same way, regardless of what leverage you choose." The brackets are steps in notional value, and the page adds that the maintenance margin "is calculated based on your positions at different notional value tiers."

Bybit builds the same idea and gives it a name. Its risk limit page for perpetual and expiry contracts opens with dynamic leverage: "This means that as traders hold larger contract values, the maximum allowable leverage decreases." The page runs its own worked example on a BTC contract, in which an account set to 90x is capped at a maximum position value of 2.6 million USDT, and to open beyond that the trader has to drop to 80x, which raises the allowed position value to 3.2 million USDT. Those are the exchange's illustrative figures on a page last updated in July 2026, for one contract, and they are exactly the kind of number that moves.

Both descriptions point at the same thing from two directions. Binance describes brackets that raise the maintenance margin rate as notional grows. Bybit describes tiers that lower the maximum leverage as contract value grows. Either way the venue has decided in advance that a large position is not allowed to be as thinly funded as a small one.

The numbers differ per venue, per contract and over time. Binance publishes the current table on its leverage and margin page for perpetuals; Bybit publishes the equivalent parameters per trading pair. Neither set is universal, and neither is permanent. Do not carry a bracket you read once onto a second exchange or a second coin.

What does that do to a "20x, risk 2%" instruction?

It makes it non-portable, which is the part that matters to a follower.

The instruction reads like a complete specification. It is not, because it silently assumes an account size. At a small notional the leverage line is achievable and the trade behaves the way the poster described. At a larger notional the venue may cap you below the number in the message, and a lower cap on the same notional means you post more margin for the same trade. That is arithmetic, not a policy: a position of a given size funded at a lower multiple simply requires more of your capital to sit behind it.

So the same message produces two different trades. Not two different opinions about the market, and not a difference in the follower's execution. The difference is the venue. One account is running the trade the channel described; the other is running a more heavily funded version of it, with a different amount of free capital left over and different room before trouble. If you have ever compared results with someone following the same feed and found the numbers do not line up, this is one of the reasons, and it sits underneath the ones people usually blame.

Leverage on its own is also not the risk. It becomes risk only in combination with a stop distance and a position size, which is the decision the message does not make for you. That is the subject of position sizing when the trade idea is not yours, and what a run of losers costs at different multiples is covered in what leverage does to a losing streak. This article is only about the ceiling the venue puts on top of both.

Why does adding to a winner change the position you already have?

Because the bracket is measured against the position, and the position just got bigger.

A trade going your way is the most natural moment to add. It is also the moment when the notional can cross into the next step, and at that step the venue applies a higher maintenance margin rate. A higher maintenance margin requirement on the same holding moves the price at which the venue closes it. Nothing about the original position changed; the classification did. If you want the mechanics of where that price comes from and why it sits closer than the headline multiple implies, we covered it in liquidation explained.

Bybit is explicit that the classification tracks orders as well as fills: "As the position value increases, the system will automatically adjust the user's risk limit based on both the position value and the active order value." In its own example, an order that would push the combined total past the maximum allowed for the chosen leverage is blocked outright rather than filled at a worse tier. That behaviour differs by venue, and the practical consequence for a follower is the same either way: a scaling-in instruction that assumes every add is identical to the first one is describing a smaller account than yours.

Can the venue move the brackets while you are holding?

Yes, and it does not always tell you loudly.

Binance reserves the right in the same help centre article: "Without prejudice to Binance's right to adjust margin requirements from time to time in accordance with the Clearing Rules, where a Maintenance Margin Rate adjustment of each tier is less than 0.5%, Binance may, in its sole discretion, apply the adjustment without issuing an announcement." Read that as a standing condition on any bracket table you have written down.

Bybit describes its own adjustment regime on the risk limit page: it conducts routine assessments of market liquidity, says the risk limit may be adjusted in the event of significant changes in market conditions, and lists what such an adjustment can change — initial margin rate, maintenance margin rate, maximum leverage allowed and position limit. The same page adds a line worth reading twice: "Bybit's email notifications are triggered only for large-scale adjustments and are not sent for risk limit adjustments." It also documents a buffer period during which an account that cannot support the new parameters is restricted to reduce-only orders until margin is added.

OKX publishes its changes as announcements. One from May 2025 adjusted the position tiers of more than sixty perpetual contracts "between May 12, 2025 (UTC) and May 16, 2025 (UTC)" and told users directly: "Users' maintenance margin ratio may increase because of tier adjustments and market volatility. To prevent forced liquidations caused by these adjustments, users are encouraged to decrease leverage by either increasing margins or closing positions."

Every one of those is a venue-specific rule that can be changed by the venue that wrote it, and none of them describes the others. The pattern across all three is the point: parameters tighten when markets get volatile, which is precisely the period in which open positions are largest and least convenient to reduce.

What does a channel's leverage number actually tell you?

Less than it appears to, in two separate ways.

The first is the venue mechanic above. A screenshot showing a position opened at some multiple tells you nothing about what your account would be permitted to do at your size, on your exchange, on that contract, today. The number is a fact about someone else's account state, not a setting you can reproduce on demand.

The second is what the number is inside the message. In our archive, as of August 2026, leverage is the most reliably published field in signal posts — 3,700 of 3,988 indexed signals state one, with a median of 10x — and it is usually a fixed channel setting applied to every call rather than a decision about the specific trade. We wrote that up in what leverage do signal channels actually ask for. A constant that never varies with the setup carries very little information about the setup.

What is actually in your control here?

Three things, all of them available before you pay anyone.

Read the table for the contract you will actually trade, on the venue you will actually use. Both Binance and Bybit publish per-contract parameters, and they are the only version that applies to your order. Look at the notional you realistically intend to run and note what leverage is available at that step, not at the smallest one.

Know your size before you judge the instruction. A feed that publishes fixed high leverage is describing trades that are comfortable at small notional and progressively less available as you grow into them. That is a compatibility question between the feed and your account, and it is answerable in advance rather than discovered during a position.

Recheck after the venue announces changes, and treat silence as no guarantee. One exchange has told you in writing that small maintenance margin adjustments can arrive without an announcement, and another has told you its notification emails do not cover risk limit changes. A table you read last quarter is a memory, not a parameter.

None of this makes a good feed bad or a bad feed good. It means one specific line in the message — the leverage number — is not an instruction your account can necessarily execute, and that when your results diverge from someone else's on the same call, the exchange belongs on the list of suspects before your discipline does.