Currently, when we use isolated margin without a stop loss, there is a high chance of liquidation because only the initial margin is added at the beginning.
When we open a futures trade on Binance using isolated margin, there is an option to add margin to that trade to reduce liquidation risk or make it zero.
Can you add this function? For example, when we choose isolated margin on a bot or grid bot, we should be able to select auto add margin to trades.
For example, if we run a long grid bot using isolated margin, we can avoid unnecessary deal closures by automatically adding margin to those deals.
I am not sure if the add margin function is supported via the API, but in the Binance futures interface, this option is available.
Yes, that’s a great idea! Adding margin to isolated margin deals helps to control and reduce the effective leverage. If we could even add margin at the start of a deal, we could set the desired leverage, and the margin would reduce it below 1x. However, as soon as we use the margin for DCA on the leveraged deal, the effective leverage would increase again while averaging down the deal.
This would make an isolated margin deal behave like a single cross margin deal in a sub-account with extra margin, without the need to create a sub-account. Or does this work differently nowadays with unified accounts at many exchanges?
I think you are correct. I thought that when we use isolated margin for bots, the exchange opens separate trades for each deal. I understood that now, but only a few exchanges can open multiple standalone positions for the same pair.
Those exchanges seem high risk. Your subaccount idea also seems good, but I need at least 36 deals to run my current testing strategy. I am not sure if any exchange allows that many subaccounts.
The “Deal” Illusion: Why Our Bots and Exchanges Don’t See Risk the Same Way
As bot traders, we often visualize our portfolios as a collection of separate, independent “Deals.” We might have a “Deal A” acting as a conservative swing trade and a “Deal B” running a high-frequency scalping strategy. However, we need to be aware of a fundamental architectural conflict: our bots operate on a Logical Layer, while the exchange operates on a Physical Execution Layer.
If we don’t understand how these two layers interact, we risk unexpected liquidations that no bot-stop-loss can prevent.
1. The Core Discrepancy: Virtual vs. Physical
Our bot providers maintain a database of “Deals”—individual objects with their own entries and targets. But at the exchange level, these distinctions don’t exist. The exchange only recognizes a Position.
When our bot opens multiple deals for the same asset (e.g., BTC/USDT), the exchange simply merges them into a single bucket. It averages our entry prices and, most importantly, calculates one single Liquidation Price for the entire stack. We cannot partially liquidate “Deal A” while keeping “Deal B” alive; if that shared liquidation point is hit, the exchange wipes the entire position, and all our deals for that asset vanish instantly.
2. The “Blast Radius” of a Liquidation
The way we configure our accounts determines the “Blast Radius” of a disaster. We generally have three levels of risk containment:
Isolated Margin (Pair-Level Risk): Here, we manually allocate funds to a specific pair. This acts as a firewall, protecting the rest of our account. However, within that pair, there is zero isolation. If we run a high-risk strategy and a low-risk strategy on the same asset, the high-risk one can trigger a liquidation that destroys both.
Cross Margin (Wallet-Level Risk): In this mode, all our positions share a single pool of collateral. While this is great for capital efficiency—allowing our winning trades to support our losing ones—it creates a dangerous “domino effect.” A single runaway trade in a volatile altcoin can drain our entire wallet, liquidating every other unrelated position we have open.
Unified Trading Accounts (Systemic Risk): Modern Unified Accounts represent the ultimate lack of “Separation of Concerns.” The exchange treats our entire account—Spot holdings, Cash, and Unrealized Profits—as one giant collateral pool. In this environment, a bad Futures trade can force the exchange to auto-sell our long-term “Spot HODL” stack just to cover a margin deficit.
3. The Global State Trap: Leverage & Mode
We must remember that exchanges treat Leverage and Margin Mode as Global Properties for a pair.
Leverage Conflict: If we try to run one bot at 5x and another at 20x on the same pair, the exchange will usually update the entire position to 20x the moment the second bot sends an order. Our “safe” 5x trade is instantly exposed to extreme risk.
Mode Conflict: We cannot mix Cross and Isolated modes for the same asset. The exchange forces us into one global setting.
The Solution: True Strategic Isolation
If we want to achieve true separation—where a failure in one strategy cannot crash another—we cannot rely on margin modes alone. The only State of the Art solution is to use Exchange Sub-Accounts.
By assigning different strategies to separate Sub-Accounts with their own API keys, we create mathematical firewalls. A liquidation in Sub-Account A has zero impact on Sub-Account B. This is the only way to ensure that our “high-risk” experiments don’t accidentally liquidate our “low-risk” foundations.
If we have more sub accounts, we can make more profits. Currently, normal users are only allowed 5 sub accounts, while VIP 7 users are allowed 200 sub accounts. That means the rich will become more rich For example, we do not have to wait until the bot shows illusion profit, but we cannot cash out profit until the position averages down to break even on the exchange.
If we can have separate deals on separate sub accounts, we can cash out money easily and use it.
I think that is why exchanges do not allow multiple separate positions on the same pair. If you look at Binace bots , they only allow 3 isolated margin bots in same pair.
Exchanges are just cheating normal users and giving more profit power to super rich users.
