Common DeFi misunderstandings

DeFi becomes easier to reason about when a slogan is replaced with a position, a mechanism, and a failure path. These ten statements are common shortcuts—not claims that every token, protocol, or user behaves the same way.

For each one, keep the useful intuition, identify where it stops, and ask a question that can be answered from code, state, governance, documentation, and economic evidence.

Misunderstanding 1: “A token is the asset itself”

Why it fails: Sometimes the token is the directly controlled ledger asset. Often it is a claim on backing, pooled assets, debt repayment, a derivative payoff, or a governance process. The same token interface can represent very different entitlements and failure conditions.

Better question: What does this token entitle its holder to, who or what provides that entitlement, and how can it be redeemed or enforced?

Read Token and Asset versus claim.

Misunderstanding 2: “A stablecoin is the same as cash”

Why it fails: A stablecoin may depend on an issuer, reserves, collateral, liquidity, governance, price mechanisms, and redemption infrastructure. Market price, accounting unit, and an eligible holder’s redemption value can diverge.

Better question: What stabilizes the unit, what backs it, who can redeem, which asset is returned, and what happens if liquidity or backing fails?

Read Stablecoin and Solvency risk.

Misunderstanding 3: “A high APY means a productive investment”

Why it fails: Reported yield can combine borrower interest, trading fees, network issuance, token inflation, temporary subsidy, leverage, or a favorable measurement window. A high number does not identify the payer or the principal risk taken to earn it.

Better question: Who pays each component of return, in which asset, for how long, and which loss exposure is the holder accepting?

Read Yield strategy and Incentive.

Misunderstanding 4: “A vault removes complexity and risk”

Why it fails: A vault can simplify the user’s interface while composing strategies, external protocols, oracles, governance, fees, and share accounting. Packaging changes where complexity is managed; it does not erase dependencies or losses.

Better question: Which operations and claims sit beneath the vault share, who controls them, how is NAV calculated, and where does a failure reach the depositor?

Read Vault, Vault share, and Integration risk.

Misunderstanding 5: “Overcollateralized lending cannot lose money”

Why it fails: Fast price moves, stale or manipulated prices, insufficient liquidation liquidity, execution failures, accounting defects, or contract bugs can leave debt larger than realizable collateral. Extra collateral reduces some risk; it does not guarantee repayment.

Better question: At what measured price and threshold does liquidation start, how quickly can it execute, who buys the collateral, and who absorbs any shortfall?

Read Liquidation and Bad debt.

Misunderstanding 6: “An AMM always provides a fair price”

Why it fails: An AMM follows its formula and current reserves. Its quote can be far from another market, and the trade itself can move the execution price. Arbitrage can trade discrepancies but is not an instantaneous fairness guarantee.

Better question: Which reserves and formula create the quote, how much will this order move it, what fee and slippage apply, and what reference market can arbitrageurs actually reach?

Read Automated market maker, Price impact, and Arbitrage.

Misunderstanding 7: “Governance means decentralization”

Why it fails: Voting power may be concentrated, participation may be low, and admin keys, multisigs, delegates, timelocks, guardians, or upgrade controls may determine what is actually executed. A public vote is one control path, not a complete decentralization measure.

Better question: Who can propose, vote, queue, execute, cancel, pause, and upgrade—and how concentrated is each effective power?

Read Governance, Multisignature control, and Governance risk.

Misunderstanding 8: “On-chain means trustless”

Why it fails: On-chain execution can make some state and rules inspectable, while the system still depends on software, validators, oracle providers, bridge operators, custodians, legal entities, governance committees, or privileged keys.

Better question: Which facts and actions originate outside the contract, who can influence them, and what happens when that dependency is unavailable or dishonest?

Read Blockchain, Oracle, and Bridge.

Misunderstanding 9: “TVL measures protocol value or safety”

Why it fails: Total value locked is an estimate of assets deposited or counted under a stated method. It does not directly measure revenue, solvency, market depth, decentralization, code safety, loss buffers, or the independence of repeatedly counted claims.

Better question: What is counted, at which price, whether claims are counted again across protocols, how much liquidity is executable, and which metric actually answers the decision at hand?

Read Liquidity risk and the structured glossary.

Misunderstanding 10: “Composability only creates efficiency”

Why it fails: Composability can reduce integration cost and create useful products, but it also creates chains of dependency. A price, liquidity, accounting, control, bridge, or contract failure can propagate through lending markets, vaults, stablecoins, and derivatives.

Better question: Which external systems can change this position’s value or availability, and what exact claim or call carries a failure across the boundary?

Read Strategy and Integration risk.

A compact replacement model

When a shortcut appears, replace it with six statements:

1. Asset: the economic resource and who controls it.
2. Claim: the holder, obligor or accounting system, entitlement, and redemption condition.
3. Operation: the financial transformation and its state transition.
4. Return: the payer, source, asset, time window, and continuation assumptions.
5. Control: every actor who can change code, parameters, custody, data, or emergency state.
6. Loss: the trigger, propagation path, first loss absorber, and residual claimant.

Use the protocol-analysis worksheet to turn that replacement model into a reviewable record, then return to the guided learning paths.