Risk Controls
Every perpetuals protocol runs risk controls. They are not an add-on to the product — without them a leveraged market cannot stay solvent, and the traders using it honestly are the ones who pay when it does not.
LeverUp has no liquidity providers. The VMMV is the counterparty to every trade, so the protocol carries the risk an LP would normally absorb. That makes active risk management a requirement, not a preference.
Performance attracts both kinds of participant
Low latency, 1001x leverage, no LP slippage, sub-second price updates, gasless one-click execution, funds that never leave your wallet — these are what make LeverUp worth trading on. They are also what makes it worth attacking.
Any venue fast enough to carry professional trading strategies is fast enough to carry professional attacks. Building the first without preparing for the second is not an option.
Valid on chain is not the same as legitimate
A smart contract enforces rules. It does not form judgements.
- A valid signature proves where a request came from and that it was authorised.
- A successful transaction proves the call satisfied the execution conditions.
Neither says anything about whether the behaviour behind it was ordinary trading. Price exploitation inside an oracle update window, orders timed with unusual precision around price movements, coordinated addresses, the same anomalous path repeated, position structures assembled to manufacture risk-free return — all of these execute perfectly well on chain.
The principle
Code execution is deterministic. Abuse detection is contextual.
Whether a trade should have happened is a question about behaviour over time, not about whether a single transaction met its preconditions.
Why stopping a position is not enough
A reasonable question: if the protocol believes a position is problematic, why not simply close it at the market price?
Because if a suspicious gain is realised at market and paid out, the attacker keeps the proceeds, the protocol and its honest users absorb the loss, and the worst outcome of a failed attack is an early close.
If the cost of a failed attack is a normal close, the attack is close to free. Risk control exists to stop suspicious proceeds being extracted, contain the loss, protect protocol solvency, and protect everyone else trading against the same vault.
How abuse is assessed
Not by a single number. A risk decision is never triggered by one profitable trade, one loss, or one price move. It rests on the whole behavioural chain, weighed together:
order timing and sequencing · oracle update timing · how a position was built and unwound · links between addresses · trading history · multi-account coordination · returns that diverge from what the market actually did · the on-chain call, signature, and transaction record
All on-chain data is permanent. Internal risk signals and review records are retained alongside it, and where necessary the complete evidence chain can be produced for a court, a regulator, a law enforcement agency, or a formal dispute process.
Why the thresholds are not published
Execution and settlement are transparent. Contract logic is audited and verifiable. Anyone can check the chain.
Detection thresholds, address-association models, and abuse-identification rules are not published, and will not be. Publishing them hands an attacker a manual for staying just inside the line.
This is not unique to DeFi. Exchanges do not publish their fraud thresholds; banks do not publish their AML models; payment networks do not publish their risk rules. The balance to strike is verifiable execution and reviewable evidence, without handing over a mechanism that can be reversed against the people it protects.
Permissionless is not consequence-free
An open system lets anyone participate. It does not oblige the protocol to accept manipulation, fraud, exploitation of anomalous prices, coordinated multi-account activity, or deliberate abuse of its mechanics.
A mature protocol is open access, self-custodial, transparently executed, and fast — and professionally risk-managed, with a formal route to contest a decision. Those are not in tension.
Who this protects
When an abnormal gain is allowed to leave the system, the cost lands on the vault, the protocol's reserves, and every other trader sharing them.
Risk control is not the protocol against its users. It is the line between a trader taking ordinary market risk and a participant extracting value from a system flaw.
We are not protecting LeverUp from traders. We are protecting traders from abuse.
Our principles
- We will never penalise a user for being profitable.
- We will not change a risk decision because of social-media pressure.
- We will not publish risk details that would help an attacker evade detection.
- We retain a complete evidence chain for every decision.
- We will submit to judicial, regulatory, or formal dispute review.
- We will keep improving transparency, user notification, and the appeals process.
- We will keep improving speed, experience, and fund safety for ordinary traders.
What this means in practice
For almost everyone, nothing. Normal trading — including high-frequency and high-leverage trading — is not subject to any of this.
The controls you may actually meet are ordinary market limits, documented where they apply: per-pair leverage caps and minimum position size, open-interest caps that pause new positions on a crowded side, the take-profit cap, the minimum holding period on a few thin pairs, liquidation, and auto-deleveraging when a market is retired or a position falls below 1x leverage.
These limits change
All of them are on-chain configuration and are updated as markets and risk analysis evolve. The trading interface reflects the live value for the pair you are on.