A DeFi app can show a swap button, a borrowing limit, and an attractive rate on the same screen. Those controls represent different financial relationships. A swap changes what you own. A loan creates a debt against collateral. Supplying liquidity puts your assets at the disposal of a market. Understanding the difference matters more than learning every crypto acronym.
The companion video introduces decentralized finance through those everyday actions. This article follows one hypothetical user through a swap and a collateralized loan, with a particular focus on the moment a seemingly manageable position can become liquidatable. It explains mechanics, not a strategy to try with money.
Separate the wallet, the website, and the contract
A wallet signs instructions with the keys controlling an address. A website makes those instructions easier to construct. The protocol’s smart contracts implement the rules that accept or reject them. The three can look like one product while carrying different responsibilities.
Suppose a site displays “borrow.” That word alone tells you little. You need to know which contract receives collateral, what asset becomes your debt, and which conditions permit collateral to be taken. An appealing interface does not answer any of those questions.
Self-custody also has a boundary. Keeping control of your wallet keys does not mean deposited assets remain freely spendable in the wallet. Once a transaction moves assets into a protocol, access follows that protocol’s rules. Likewise, granting a token allowance can give a contract permission to transfer specified tokens without asking for a fresh approval of the same scope each time.
A swap changes the pool as well as your balance
In an automated market maker, a trader exchanges assets against reserves held by a contract. Liquidity providers supply those reserves. Uniswap’s explanation describes how its pools price swaps algorithmically rather than matching a conventional list of buy and sell orders.
Imagine a simplified pool containing 10 units of token A and 10,000 units of token B. Ignoring fees, a constant-product rule keeps the product of the reserves at 100,000. Follow a purchase with 1,000 B:
- The B reserve rises from 10,000 to 11,000.
- To preserve the product, the A reserve must become 100,000 divided by 11,000: approximately 9.09 A.
- The trader receives the difference between the old and new A reserves: approximately 0.91 A.
The opening reserve ratio suggested 1,000 B per A. It did not promise a full unit of A for the trade, because the purchase itself changes the reserves.
These are invented quantities for arithmetic, not a quote for an actual token. The useful observation is that an opening price is not a promise to supply any amount at that price. Current pool designs can also concentrate liquidity into particular price ranges; the simple example is a starting model rather than a description of every market.
Price impact is the change caused by your own trade. Slippage can include further deterioration between the expected quote and execution. Network costs sit alongside those trading costs. For a deeper explanation of pool pricing, see how DeFi liquidity pools and AMMs work.
Borrowing adds a moving obligation
Now imagine supplying a volatile token as collateral and borrowing a dollar stablecoin. You still have exposure to the collateral’s price, while owing the borrowed amount plus the interest required by the protocol. Spending the borrowed token does not reduce the debt.
The borrowing limit and liquidation threshold serve different purposes. One governs how much a position may initially borrow; the other governs when it becomes eligible for liquidation. Neither should be interpreted as a recommendation for how much risk to take.
The Aave health-factor documentation provides a concrete example of the second calculation. For a single collateral asset, multiply its value by its liquidation threshold, then divide by the debt value. The actual protocol can combine several assets with a weighted threshold.
For our own hypothetical example, hold the liquidation threshold at 80% and the debt at $1,000:
| Collateral value | Value multiplied by threshold | Health factor |
|---|---|---|
| $2,000 | $1,600 | 1.6 |
| $1,250 | $1,000 | 1.0 |
At the second value, the threshold-adjusted collateral only equals the debt. A further fall pushes the ratio below 1 and makes the position eligible for liquidation under that model. The threshold and amounts are illustrative, not current terms for a specific market.
Notice what happened: the borrower need not miss a scheduled payment. The collateral value crossed a contractual boundary. Interest can also increase the debt, and the borrowed asset’s price can change. Looking only at the original dollar amount misses those moving parts.
Liquidation is an executable rule
When an Aave position becomes eligible, a liquidator can repay debt and receive collateral plus a liquidation bonus. How much debt can be liquidated depends on the current rules and position. It is inaccurate to assume that every liquidation always sells everything, or that a person will phone before it starts.
A dashboard warning is therefore a description of state, not a grace period. Adding collateral or repaying debt requires an executable transaction. That transaction may compete with other activity, encounter a fee you did not expect, or fail to arrive before a liquidation transaction. The fact that you are watching the screen does not reserve a place in the next block.
This makes a useful paper exercise: change the collateral value, debt value, and threshold separately. Work out which change pushes the ratio toward 1. The exercise reveals the exposure without requiring a wallet connection or a live position.
Trace the dependencies behind the rate
A displayed return should lead to questions about its source. Borrower interest, trading fees, and promotional token rewards are different revenues. A reward token’s quoted value can move while the headline rate looks impressive. Supplying liquidity also differs from lending: the pool’s changing composition can leave you worse off than simply holding the original assets.
The BIS analysis of DeFi risks highlights leverage, connected protocols, and concentrated governance. Its central warning is useful when reading an app: automation changes how financial activity is administered, but it does not eliminate the dependence on people, incentives, and outside information.
Price feeds, administrators, stablecoin issuers, bridges, and the underlying blockchain can each add a failure path. An audit addresses particular code and assumptions; it cannot establish that every economic outcome is favorable. Before interpreting a rate as income, identify what you supplied, what you owe, what permission you granted, and what would prevent exit. If those relationships remain unclear, the interface has not yet explained the product.
