Gate AI Risk Control Analysis: Smart Stop-Loss vs Hard Stop-Loss — How They Work Together to Protect Assets

According to Gate market data, as of April 9, 2026, the Bitcoin price is $70,905.9, and Ethereum is $2,178.57. In only the past 24 hours, Bitcoin’s price amplitude has exceeded $2,300, while Ethereum’s price has fallen from a high of $2,270.47 to a low of $2,162.01. In such a high-volatility market, stop-loss is no longer a multiple-choice question of “whether to set it,” but a strategy question of “what kind of way to set it.”

Gate AI provides multi-layer stop-loss tools for crypto traders. Among the two most core concepts—“smart stop-loss” and “hard stop-loss”—users often confuse them because their functional positioning is different. This article breaks down the key differences between the two from three dimensions: underlying mechanism, trigger logic, and usage scenarios, and provides a combined-use framework based on the current market environment.

First, clarify the concepts: what do the two types of stop-loss refer to?

Within Gate’s trading ecosystem, “smart stop-loss” and “hard stop-loss” are not the same kind of function; rather, they are two mechanisms that serve different levels of risk-control needs.

Smart stop-loss—AI-driven strategy-level risk control

Smart stop-loss specifically refers to the “global stop-loss” feature in Gate AI trading robots. It is not a stop-loss line for a single trade; instead, it sets a unified loss threshold for the entire AI trading strategy. When the strategy’s overall loss reaches the preset ratio, the system automatically terminates all related trades.

The core design principle of this feature is “risk control up front”—it locks the risk boundary before the strategy is executed, rather than passively responding after risk occurs. Gate AI strategy builders themselves do not directly operate users’ assets. Users must manually review and approve every action, and assets always remain in the users’ trading accounts, with full visibility into the permission layer.

Hard stop-loss—fixed stop-loss rules at the platform layer and contract layer

“Hard stop-loss” covers stop-loss mechanisms in the Gate platform that are not AI-driven and are triggered at fixed prices, mainly including two scenarios.

First category: position stop-loss in contract trading. After opening a position, users directly preset a stop-loss price. Once the price touches it, the system automatically closes the position at market price. Gate refines take-profit and stop-loss into two modes: “position take-profit and stop-loss” (one-click close-all) and “planned take-profit and stop-loss” (partial position reduction), suitable for more granular needs across different trading styles.

Second category: the platform layer’s tiered liquidation mechanism. When the contract position’s risk ratio increases, Gate does not clear the entire position at once. Instead, it executes in batches—first closing 10% to 20% of the position to bring the margin ratio back to a safe line, thus giving the main position more time to survive. At the same time, Gate uses the mark price rather than the latest trade price as the liquidation determination baseline, fundamentally preventing false triggers caused by momentary price spikes that could lead to unintended liquidation.

Core differences: an in-depth comparison across four dimensions

The two stop-loss mechanisms differ fundamentally in how they operate, what triggers them, their risk boundaries, and who uses them. The following breaks them down step by step.

Comparison One: Operating mechanism

Smart stop-loss is a strategy-level dynamic risk control that is triggered based on the AI’s comprehensive evaluation of the strategy’s overall performance. It targets the entire investment portfolio rather than a single trade. When the strategy’s overall loss reaches 8% or 10% of the initial principal, the system automatically terminates all related trades, preventing a single loss from spreading throughout the entire portfolio.

Hard stop-loss, on the other hand, is a trade-level fixed trigger. Users preset a specific stop-loss price in advance, and when the market price reaches that level, the system automatically executes a closing of the position. The trigger conditions are clear, and the execution logic is definite.

The most fundamental difference is: smart stop-loss measures the strategy’s overall performance by percentage, while hard stop-loss measures the risk of an individual trade by price levels.

Comparison Two: Trigger conditions

Smart stop-loss is triggered based on the magnitude of the strategy’s overall loss. Taking Gate AI smart grid as an example: if you invest 5,000 USDT principal and set a global stop-loss at 8%, once the overall unrealized loss reaches 400 USDT, the stop-loss is triggered and the strategy automatically stops—rather than triggering a stop-loss for any single grid trade by itself.

Hard stop-loss is triggered based on whether the market price touches the preset price level. For example, in a BTC contract trade, if the user opens at an average price of $70,905.9 and sets a stop-loss price of $69,000, once the price touches that point, it triggers the closing—regardless of the strategy’s overall profit and loss.

Comparison Three: Risk boundaries

The design philosophy of smart stop-loss is “protecting the overall safety of the investment portfolio.” Its risk boundary is proactively set by the user in percentage terms. Gate recommends setting global stop-loss in a dynamic range of 5% to 15%, balancing returns and drawdown space.

The design philosophy of hard stop-loss is “locking in the maximum loss of a single trade.” Its risk boundary is determined by the price difference. In addition, Gate’s contract system includes a price-spread protection feature: when the take-profit or stop-loss is triggered, if the price difference between the latest trade price and the mark price exceeds the system threshold, the order will be automatically rejected from execution, effectively preventing abnormal stop-loss caused by malicious price spikes.

Comparison Four: User and scenario

Smart stop-loss is configured by users of Gate AI trading robots and is suitable for scenarios that run automated trading strategies, such as AI smart grids, AI enhanced DCA, AI trend tracking, and so on. It answers the question: whether the “entire strategy” needs to stop.

Hard stop-loss applies to all manual contract traders who open positions, as well as traders who seek precise price control. It answers the question: how much this trade can lose at most.

Summary of differences

How to combine them: the three-layer safety cushion framework

Smart stop-loss and hard stop-loss are not replacement relationships; they can be coordinated to build a multi-layer risk-control system. Within Gate’s trading framework, take-profit/stop-loss and liquidation mechanisms form a complete closed loop of “active defense + passive backstop.” Based on this idea, it is recommended to build a risk-control structure from the following three levels.

First layer: Active risk control—hard stop-loss locks single-trade risk

This is the most basic protective layer. Whether it’s manual trading or AI-assisted trading, set a position stop-loss or planned stop-loss at the time of opening. Using current market data as an example, BTC’s 24-hour amplitude is about 3.3% ($70,461.3 to $72,857.1), ETH’s 24-hour amplitude is about 5.0% ($2,162.01 to $2,270.47). You can reference this amplitude to set a reasonable buffer for your stop-loss levels.

For traders holding BTC long positions, set the stop-loss price 2% to 3% below a key support level. For ETH contract holdings, since its volatility is higher than BTC’s, you can loosen the stop-loss range accordingly. At the same time, it is also recommended to enable price-spread protection to prevent abnormal spikes from triggering the stop-loss.

Second layer: Strategy backstop—Gate AI smart stop-loss protects the overall portfolio

When running an AI trading strategy, enable global stop-loss at the same time. Gate AI recommends setting the global stop-loss between 8% and 12%. This range both allows enough tolerance for intraday volatility and cuts off risk in time if the strategy deviates systemically. For example, a BTC grid strategy can set global stop-loss at 8%, an ETH swing strategy can set it at 10%, and a GT strategy can set it at 6% to 8%. As of April 9, 2026, the GT price is $6.48, with a 24-hour change of -2.11%. As a platform token, its volatility is relatively smaller than that of mainstream coins, so you can set a relatively tighter stop-loss range.

The relationship between global stop-loss and single-trade stop-loss is: even if an individual trade has not yet reached its stop-loss price, if the overall strategy loss exceeds the threshold, the system will still terminate the strategy run. This prevents the common problem of “holding individual trades through losses while the overall continues to bleed.”

Third layer: Profit protection—move profits to an insurance box to lock in gains

Gate AI’s “move profits to an insurance box” feature is an effective supplement to the stop-loss mechanism. After enabling it, part of the profits generated by the strategy will be automatically transferred to the spot account, preventing profits from being given back when the market reverses. Users can set a fixed percentage of profits to transfer automatically, or set different transfer percentages based on yield tiers—for example, transfer 20% of profits when the return rate reaches 5%, and transfer 30% when it reaches 10%.

This layer works in coordination with stop-loss: with stop-loss protection limiting downside risk, and the profit-to-insurance-box feature locking in upside gains, together they form a complete two-way protection.

Summary of the combined-use framework

Taking a user who runs an AI smart grid strategy on Gate and also performs manual contract trading as an example, a complete risk-control configuration should include the following elements:

• Manual contract positions: For each open trade, set a position stop-loss; the stop-loss price should reference the underlying’s 24-hour amplitude and key support level
• AI smart grid strategy: Enable global stop-loss, set the threshold at 8% to 10%, and enable smart backtesting to validate the effectiveness of parameters
• All AI strategies: Enable the move-profits-to-insurance-box feature, and set tiered profit transfer ratios
• Extreme market protection: Rely on Gate’s tiered liquidation and mark price mechanisms as the final passive line of defense

The core logic of this framework is: use hard stop-loss to lock the floor for single-trade risk, use smart stop-loss to control the overall strategy risk, use profit protection to lock in realized gains, and use platform mechanisms to guard against systemic black swans.

Conclusion

In the high-volatility environment of the crypto market, a single risk-control tool often cannot cover all types of risk. Gate AI’s smart stop-loss and hard stop-loss provide protection from two dimensions—overall strategy and individual trades. Only by using them together can you form a truly robust risk-control closed loop.

Under the current market structure, where BTC has a market cap of $1.33T and a market share of 55.27%, and ETH has a market cap of $271.24B with a share of 10.58%, the volatility characteristics of mainstream assets and small-to-mid market-cap assets differ significantly. This makes it even more necessary to flexibly allocate these two stop-loss tools.