Transaction Simulation: What a Preview Can and Cannot Catch

Transaction simulation executes a transaction against current chain state without broadcasting it, then reports what would change. Instead of a hex payload and a contract address, you see that your balance would fall by a specific amount and an unfamiliar address would gain it. That translation, from what you are authorising to what it does, is the single most valuable thing a wallet can put between a user and a signature.

Why it beats reading the transaction

Wallets historically showed a destination address, a value and a data blob. Every one of those is accurate and none of them answers the question the user actually has.

A contract call's meaning lives in the interaction between the payload and the contract's code, which a user cannot evaluate by inspection. A malicious approval and a legitimate one look identical at that level — both are a function call with an address and a number.

Simulation changes the question from interpretation to observation. It does not ask you to understand the call; it runs it and reports the outcome. That is a categorically easier thing for a person to check.

What it reliably catches

Drainer contracts. A transaction presented as a mint that would in fact transfer your holdings shows up as exactly that: your balance down, an unknown address up. The mismatch between what you expected and what is shown is the alarm.

Unexpected approvals. A simulation that reports a new unlimited allowance being granted to an address, when you believed you were claiming a token, surfaces the mechanism described in every guide about approval drains at the moment it is still preventable.

Wrong-asset and wrong-amount errors. Ordinary mistakes — the wrong token, a decimal misplaced, the wrong recipient — are visible in the projected outcome even when the transaction is entirely legitimate.

Known-malicious counterparties. Most implementations pair simulation with reputation data, so an address already associated with drains is flagged regardless of what the simulation shows.

Where it does not help

State-dependent behaviour. A contract can behave differently depending on when it is called, who calls it, or what the chain state is at execution. A simulation runs against state at the moment of preview, and a transaction that lands in a different state can do something else. Contracts written specifically to pass simulation and behave differently on execution are a known technique.

Delayed effects. Simulation reports the immediate result. A transaction whose damage arrives later — granting a role, setting a parameter, queuing an action for execution after a delay — shows a benign immediate outcome because the immediate outcome is benign.

Off-chain signatures. Signing a message is not a transaction, so there is nothing to simulate in the ordinary sense. Some tokens allow allowances to be granted this way, and a simulation-first wallet is weakest exactly where users are most conditioned to click through.

Everything outside the chain. If you are being socially engineered into a transaction that does precisely what you intended, simulation confirms it accurately and you still lose the funds.

How to use it well

The value is in the comparison, not the display. A simulation you glance at is decoration; a simulation you check against what you expected is a control.

Form the expectation before you open the preview — I am swapping this for that, I should end up with roughly this much — and then read the projection against it. Any difference, including one you cannot explain rather than one that looks wrong, is a reason to stop.

• Decide what you expect the transaction to do before you look at the simulation.
• Treat an unexplained difference as a stop, not as something to reason around.
• Give message signing the same scrutiny as transactions, since it is the gap in the model.
• Do not treat a clean simulation as permission to interact with an unknown contract — it reports outcomes, not intentions.
• Prefer wallets where simulation is on by default rather than a feature you must find.

Simulation is the largest single improvement in retail wallet safety available, and it is a control rather than a guarantee. The distinction matters most for people who, having installed it, stop reading.