Skip to main content

MIP-4 — Options

info

A first release is live, and it is deliberately narrow. Standard European, fully collateralized puts and calls, cleared through RFQ only. A put settles in USDC; a call settles in the underlying coin and escrows one coin per contract. Read Options for the product, and option_series for the wire.

The margined options book this page originally scoped is NOT built. The constraints below are the reason, and the shipped design answers them by side-stepping them: the chain never prices an option, so no option value ever enters committed state.

MIP-4 is the MetaFlux options product.

MIP-4 previously named a perps liquidity aggregator. That design is withdrawn and the number is reassigned to options.

Why options are the right next primitive​

MetaFlux competes on capability, not on price. Perpetuals give a trader one axis: direction, with leverage. Options add the two axes a perpetual cannot express — convexity and time — and they are what a hedger needs. A miner selling forward, a treasury protecting a floor, and a market maker running a delta-neutral book all need a payoff that a perpetual cannot build cheaply.

The margin engine is the reason this is a MetaFlux product rather than a generic one. MetaFlux already runs cross-asset portfolio margin over a governed grid of price and volatility scenarios. An options book margined inside that same portfolio, rather than position by position, is the differentiator: a covered call should not cost the same margin as a naked one, and a spread should not cost the sum of its legs.

Scope​

What the first release shipped:

  • Standard European puts and calls on assets that already carry a live MetaFlux price feed. A put is cash-settled. A call is coin-settled: it escrows and pays ONE unit of the underlying, so its underlying needs a spot token.
  • Full collateralization. The holder pays the premium; the writer escrows the worst case. Neither leg can be liquidated.
  • RFQ clearing. There is no option order book, and the chain computes no premium.
  • Settlement from a window mean of committed oracle prices, with a defer-and-widen rule and an abandonment backstop.

Deliberately out of the first release, so the first version can be proved rather than merely shipped:

  • Margined options. An option position holds its own collateral and does not offset a perpetual.
  • Portfolio margin across options and perpetuals together.
  • Cash-settled calls. max(S* - K, 0) in USDC has no finite worst case, so no cash escrow can fully collateralize it. The lane sidesteps that rather than bounding the payoff: the same call read in the underlying is worth at most one coin, so a call escrows one coin and settles in coin. See why a call escrows one coin.
  • Physical settlement, and exotic payoffs.
  • Permissionless options deployment. A series is listed by validator ⅔-stake vote. The MIP-3 pattern can follow, but a permissionless options market is a risk surface that must be earned.

The constraints that shape the design​

These are not implementation details. They decide what the product can be, so they are stated here rather than buried.

Every price must be reproducible by every validator​

MetaFlux is a chain. Each validator recomputes each block and the results must agree exactly. Any quantity that enters committed state must therefore be computed in exact arithmetic. Floating point is not usable, because two machines can round it differently and a disagreement halts the chain rather than degrading quietly.

An option value is a transcendental function of its inputs. Producing one in exact arithmetic, identically on every machine in the fleet, is the central engineering problem of MIP-4. It is being solved before anything is built on top of it.

Margin needs a value in every scenario, not just at the current price​

The portfolio margin engine asks what a position is worth after a price shock and a volatility shock. A perpetual answers cheaply: its value moves with the price, one for one. An option does not. Each scenario needs the option revalued, which multiplies the cost of a margin pass by the number of scenarios in the grid.

The grid is governed, so it can be tuned. The cost is still the budget that decides how many option series a single account can hold.

Volatility must come from somewhere trustworthy​

An option's value depends on expected volatility. Reading that from the option book itself is circular during exactly the moment it matters: in a liquidation the book is thin and moving, and a volatility read from it would feed the margin call that is causing the move. MetaFlux already guards a mark price against a wash-traded book. Volatility needs a guard of the same kind, and it is an open design question which source carries it.

What exists today​

The collateralized lane is live: the series registry, the RFQ trade path, the escrow lifecycle and the expiry settlement. See Options.

What is not built is everything that needs an option VALUE — a margined option position, portfolio margin over options and perpetuals together, and any read that serves a premium or an implied volatility. The three constraints above are open, and the shipped lane needs none of them answered.