DeFi risk rarely arrives as one dramatic event. It usually accumulates in a ratio.
Aave calls that ratio the health factor. It compares collateral value, liquidation thresholds and borrowed value. A higher value gives more room. A lower value moves a position closer to liquidation eligibility.
That sounds simple. It is not a promise of safety. Collateral prices move. Borrow balances accrue interest. Oracles update prices. Liquidity can change when many positions need to act at once.
That uncertainty creates a useful Yes/No/Maybe lens. Yes can mean a position stays above a defined threshold. No can mean the position crosses it. Maybe can describe a published middle band, such as a health factor that remains above one but enters a monitoring range.
Maybe does not make a leveraged position safer. It makes the question more precise.
The Yes, No and Maybe mechanic
Oddup Markets uses three outcomes. Yes and No split 90% of the pool. Maybe wins 10% of the pool reserve. The reserve is not the whole pool.
For a DeFi market, the contract must define the event before anyone reads a chart. It should name the protocol, chain, asset pair, measurement source, time window and threshold.
- Yes: the defined position remains above the stated health-factor threshold at the settlement time.
- Maybe: the position enters a stated monitoring band but does not cross the liquidation boundary.
- No: the position reaches the defined liquidation condition or fails the contract's stated test.
The middle band needs a number and a clock. “Nearly liquidated” is not a settlement rule. Neither is “looks healthy”. A clean contract could define Maybe as a health factor from 1.00 to 1.10 at a specified snapshot. That is an illustrative market design, not a live Oddup quote.
What Aave health factor measures
Aave defines health factor as the numeric representation of a borrow position's safety. Its documented formula is:
Health factor = total collateral value × weighted average liquidation threshold ÷ total borrow value.
A health factor below one makes a position eligible for liquidation under Aave's documented rule. The metric therefore measures distance to a boundary. It does not predict the next asset move.
Liquidation thresholds vary by collateral asset. Aave Governance determines those parameters. The weighted average combines the thresholds across collateral in the position.
That weighting matters. Two accounts can hold the same collateral value and debt value. Their health factors can still differ because their collateral mix differs.
The Aave FAQ gives a worked example with $10,000 of ETH collateral, an 80% liquidation threshold and $6,000 of USDC debt. The calculation is 10,000 × 0.8 ÷ 6,000, producing a health factor of 1.333.
That number is a documented example, not a recommendation. It also shows why a single collateral price is not enough. The threshold and debt balance matter at the same time.
Why the buffer can shrink without a trade
A position can move closer to liquidation even when the borrower does nothing.
If collateral value falls, the numerator falls. If borrowed value rises through accrued interest, the denominator rises. If both happen together, the health factor can decline faster.
Aave states that borrow positions have no fixed repayment period while they remain properly collateralised. It also notes that accrued interest can reduce the health factor over time.
This creates a key distinction between a static balance sheet and a live position. A snapshot can show a comfortable ratio. The same account can have less room after price movement, interest accrual or a collateral-parameter change.
Volatility makes the distance harder to interpret. A health factor of 1.20 is not equally comfortable for every collateral pair. Correlated assets may move together. Uncorrelated assets can create a wider joint-risk surface.
Aave's FAQ states that there is no exact universal “safe” health factor. The right margin depends on volatility and correlation. Treat any single comfort number as a risk preference, not a protocol guarantee.
Oracles turn prices into a protocol event
Health factor calculations depend on the prices used by the protocol. Aave documents an oracle for each reserve. Governance selects the oracle used for that reserve.
The Aave oracle documentation describes Chainlink Price Feeds and correlated-asset price oracles among the oracle types used on production Aave markets. The feed is not a trader's preferred exchange screen. It is protocol infrastructure.
Chainlink's Data Feeds documentation explains that data feeds aggregate data from multiple sources and publish an onchain result for smart contracts. The practical lesson is simple: define the source before defining the outcome.
A prediction market that references an Aave health factor must therefore specify the chain and data source. It should also specify the observation time or observation window. “At the end of the day” leaves too much room for dispute without a timezone and a source.
Oracle design does not remove market risk. It defines how the protocol converts market information into a value that smart contracts can use.
Worked example: separate the boundary from the band
Consider an illustrative Aave-style contract. It is designed to explain the mechanics, not to forecast a live position.
Question: Will the defined borrow position keep a health factor above 1.00 at the stated snapshot?
Inputs: $10,000 collateral value, an 80% weighted liquidation threshold and $6,000 borrow value.
Starting calculation: 10,000 × 0.80 ÷ 6,000 = 1.333.
Now define the three outcomes:
- Yes: the health factor is above 1.10 at the snapshot.
- Maybe: the health factor is at least 1.00 but no more than 1.10.
- No: the health factor is below 1.00 and the position is eligible for liquidation.
The numbers in the band are illustrative. A real contract would need an approved source, chain, asset identifiers, timestamp and treatment of missing data.
Suppose collateral value falls while the threshold and debt balance stay constant. At $8,250 of collateral, the calculation becomes 8,250 × 0.80 ÷ 6,000 = 1.10. The position has not crossed one. It has entered the defined Maybe band.
At $7,500 of collateral, the calculation becomes 7,500 × 0.80 ÷ 6,000 = 1.00. The position sits at the boundary. A production contract must say whether “below one” means strict inequality and how rounding works.
At $7,400 of collateral, the calculation becomes about 0.987. Under Aave's documented rule, a health factor below one is eligible for liquidation.
The example shows the value of separating three ideas. The first is the current ratio. The second is the monitoring band. The third is the liquidation boundary. They are related, but they are not the same event.
Liquidation is an action, not just a label
When a position becomes eligible, liquidation mechanics still matter. Aave's documentation describes a liquidator repaying part of the debt and receiving a corresponding amount of collateral plus a liquidation bonus.
The Aave Pool documentation describes a close factor that limits how much debt a liquidator can cover in the relevant implementation. It also explains that a successful liquidation raises the health factor.
A market should not treat “eligible” as identical to “fully closed”. Those are different states. A contract must define whether it settles on eligibility, a liquidation transaction, a debt balance or a final account state.
This distinction also matters for timing. A health factor can cross a boundary before a liquidation transaction confirms. The source and timestamp decide which event the market observes.
Three checks before reading a DeFi market
1. Read the threshold
Do not compare health factors without checking the denominator, collateral mix and liquidation thresholds. A number without its inputs is incomplete.
2. Read the oracle rule
Identify the chain, oracle and observation method. A screen price and a protocol price can differ without either being “wrong”. They answer different operational questions.
3. Read the settlement clock
Check the timezone, snapshot window, rounding convention and treatment of missing data. Those fields define the event more reliably than a headline.
Why this matters for prediction traders
Aave health factor is useful because it turns a complex position into a traceable ratio. It is risky when traders treat that ratio as a guarantee.
The Yes/No/Maybe lens adds discipline. Yes can describe a defined buffer. No can describe a crossed liquidation boundary. Maybe can describe a published middle band. Each outcome needs an exact rule.
Yes and No split 90% of the pool. Maybe wins 10% of the pool reserve. It is not a guaranteed win, a tie or a promise of protection.
The practical edge is not guessing the next candle. It is knowing which value the protocol will use, which threshold matters and when the contract settles.
In DeFi, distance is information. It is never certainty.
Compliance note: This article is for educational purposes only. It is not financial, legal or tax advice. DeFi and prediction markets involve risk. Outcomes are not guaranteed. Read the protocol and market rules before participating. Oddup does not make price predictions or recommend any individual trade.
