Your collateral ratio shows 350%, but liquidation can still happen, even within minutes. This is not a contract bug; it's because many people only look at the collateral ratio while ignoring the real liquidation trigger conditions. Next, I will break down the entire process and guide you step by step to check whether your current position is truly safe, and why those seemingly "extremely safe" numbers can be deceptive.
A leading global cryptocurrency platform,suitable for both beginners and experienced traders.
New user benefit: 20% off trading fees upon registration!！
Step 1: Discard the "Collateral Ratio" Concept Immediately and Focus Only on Health Factor
You need to stop judging safety by the ratio of collateral value to borrowed value, and instead use the protocol's native Health Factor (HF) as your standard.
- Open the page of the lending protocol where you currently have collateral (such as Aave, Compound, or an on-chain yield aggregator).
- In the "Dashboard" or "Your positions" section, find Health Factor. Most protocols display HF near your net worth.
- If the interface only shows a collateral ratio, calculate manually: Collateral × Liquidation Threshold ÷ Total Debt. In Aave V3, ETH's liquidation threshold is 82.5%. When you borrow assets worth 1 ETH and deposit 3.5 ETH as collateral, HF = (3.5 × 82.5%) ÷ 1 = 2.8875. Many beginners think the collateral ratio is 350%, which is simply a misleading number.
Your health factor must stay firmly above 1.3—if it drops below 1.1, the liquidation engine can trigger at any time; equal to or below 1 means you enter the liquidation queue directly, no matter what "collateral ratio" your page shows.
Step 2: Check If the Liquidation Threshold Is "Discounted" Because of Asset Type
Find out the liquidation threshold for each asset you have deposited as collateral, rather than relying on a vague average.
- Go to the protocol's "Risk Parameters" page. Taking Aave V3 as an example, each collateral type has its own LTV and liquidation threshold.
- Record your position composition:
- If it is a stablecoin (e.g., USDC), the liquidation threshold can be as high as 85%–90%.
- If it includes less liquid tokens or long-tail assets, the threshold may be only 40% or even lower.
- Calculate the weighted true threshold. For example, if your collateral is 2 ETH (threshold 82.5%) and 1 unit of some altcoin A (threshold 45%), the combined threshold after weighting is far lower than the 82.5% you assumed.
You now have a per-asset liquidation threshold list for your current position, and the weighted threshold has been factored into your personal HF calculation, instead of blindly using the market price of collateral divided by debt.
⚠️ Common failure reason: treating the protocol's "Max LTV" as the liquidation line. LTV is only the borrowing limit; the liquidation threshold is usually a few percentage points higher than LTV, and for high-risk assets this gap is even smaller. Once the price slips slightly, it can punch straight through the liquidation threshold, giving you no "buffer time."
Step 3: Watch Out for the Illusion of the Pricing Unit—Are Collateral and Debt the Same Currency Pair?
Check whether there is hidden foreign exchange exposure between the denomination of your collateral and the borrowed asset.
- Suppose you deposited ETH on some chain, but the borrowed asset is USDT. Many people think the collateral ratio is calculated independently based on ETH/USDT. In reality, debt is always liquidated in the denomination of the borrowed asset.
- If USDT suddenly strengthens against ETH (i.e., ETH price drops sharply), even if your amount of ETH stays unchanged, the total debt value swells when measured in ETH, causing HF to plummet. This is why positions that borrow stablecoins with volatile collateral can collapse, no matter how high the collateral ratio appears.
- Conversely, if you borrow ETH against stablecoin collateral, a seemingly high collateral ratio can be wiped out if ETH surges, as the debt value skyrockets.
You can set up a simple "debt/collateral" exchange rate monitor on any market data page (such as CoinGecko): take the price pair of debt asset / collateral asset. When that pair fluctuates more than 5% in 24 hours, manually recheck your HF immediately.
Step 4: Verify Whether the Oracle Feed Price Is Lagging or Open to Manipulation
Examine the type of oracle your protocol relies on, along with its update frequency and deviation tolerance.
- Find the protocol's documentation or a third-party analysis page (like the Oracles section on DeFiLlama) to confirm whether it uses Chainlink, Pyth, or a self-built TWAP oracle.
- For most protocols based on Chainlink, price update triggers are generally a deviation threshold (e.g., 0.5%) and a heartbeat time (e.g., 1 hour). In extreme market conditions, a price can easily drop over 20% within that hour, while the oracle has not yet updated. The "high collateral ratio" you see at that moment is an outdated illusion.
- If the protocol uses DEX TWAP or spot median, the possibility of manipulation by flash loans or whales in a short time is higher, especially on chains with low liquidity.
Document the "maximum update interval" and "deviation threshold" of the oracle used by your current position. If the interval is ≥30 minutes and your collateral ratio seems safe, you must reserve an extra buffer of at least 50% for price swings; otherwise, do not assume this number is real-time.
⚠️ Risk reminder: In 2023, a perpetual protocol's oracle pricing error caused many users' collateral ratios to drop from 200% to zero instantly, wiping out all margin. If you rely solely on on-screen numbers without verifying oracle health, the responsibility for a zeroed-out account lies entirely with you.
Step 5: Always Avoid the "Same-Asset Cyclical Collateral" Liquidation Trap
Check whether you used borrowed assets to swap back into the same collateral to increase your position—the so-called looping strategy.
- Look at your wallet's transaction history: did you borrow ETH from the protocol, then swap it for more ETH on a DEX and redeposit it as collateral?
- This operation amplifies liquidation risk non-linearly. When ETH drops, the collateral value across all leverage layers shrinks at the same time, and HF declines much faster than with a single-layer position.
- If you need to add collateral through a centralized exchange, you can buy coins on OKX and transfer them to your on-chain wallet, or use Binance's fast withdrawal to avoid network congestion.
Visualize all paths of cyclical collateral and make sure the total leverage ratio is within your acceptable range. The safest approach is to unwind at least one layer of the loop, returning HF to the level of a single-collateral position.
Step 6: Set Up Automatic Stop-Loss and a Backup Funding Channel
Build a real-time protection mechanism that does not rely on the protocol itself.
- Use a DeFi vault management tool (such as DeFi Saver or Flashbots' automation service) to configure HF auto-recovery: when the health factor drops below 1.2, it automatically repays debt or adds collateral.
- Also keep a reserve of USDC/ETH in a centralized account, ready at any time, and pre-authorize an on-chain contract to use those funds for repayment.
- Set up on-chain monitoring (via Tenderly, wallet notifications, or custom scripts) so that once HF falls into the 1.15 zone, a top-up transaction is triggered immediately.
You should now receive wallet push notifications, and have a backup fund covering at least 20% of total liabilities, capable of pulling the health factor back up in a single transaction.
How to Verify You Are Truly Safe
After making the above adjustments, don't leave the page right away.
- Wait at least 15 minutes, refresh the protocol page, and confirm that the Health Factor has stabilized above 1.5 (some protocols may show temporary deviations due to update intervals).
- Go to the oracle price feed page (such as Chainlink Data Feeds or Pyth's official site) and check the real-time price, confirming that the deviation from the price your position uses is less than 1%.
- Finally, check the last 10 transactions related to the protocol's liquidations on a blockchain explorer to make sure no abnormal liquidation events have occurred in the past few minutes.
A leading global cryptocurrency platform,suitable for both beginners and experienced traders.
New user benefit: 20% off trading fees upon registration!！
FAQ
Q1: If I see a "liquidation warning" in my wallet but the page still shows I can act, how much time do I have?
There is no universal time frame. Most protocols have no grace period at all; the transaction gets packaged and liquidated in the next block after the health factor falls below 1. The only safe move is to add collateral or repay debt immediately, rather than gambling on a rebound.
Q2: Why would someone with a collateral ratio above 1000% still get liquidated?
Usually because the protocol temporarily lowered the liquidation threshold for the collateral, or the on-chain oracle was attacked/erroneous, causing the system to misjudge. Such cases are uncommon but fatal, and can only be avoided through diversification across multiple protocols and oracle monitoring.
Q3: Is swapping collateral into stablecoins the best way to avoid liquidation?
Not necessarily. If a large amount of stablecoins serves as collateral but the stablecoin depegs (e.g., to $0.95), the liquidation line can be quickly breached. You need to use at least two stablecoins with different mechanisms and maintain a health factor above 2.
Q4: I maintain a high collateral ratio. Do I still need to worry about liquidation penalties?
Liquidation penalties do not actively harm you, but once a liquidation occurs, you not only lose collateral but also pay an additional 5%–15% liquidation penalty (see the specific protocol's risk parameters for details). This has no direct relation to your collateral ratio; it is simply set by the protocol.
Q5: Is there any protocol where I can completely ignore the liquidation threshold?
No. All lending protocols have liquidation logic; otherwise, bad debt would accumulate. Even a protocol with a special design must rely on some kind of price assessment to close positions, just that the trigger condition may be disguised under a different name.
