It means selling insurance on a price move, and being paid a premium upfront for it. A covered call: the treasury holds ETH and sells someone the right to buy that ETH at a fixed higher price (the strike) on a fixed date. If ETH stays below the strike, we keep the ETH and the premium. If it rises above, we sell at the strike — we still keep the premium, but give up the upside beyond the strike.

A cash-secured put is the mirror: we hold cash and sell someone the right to sell us ETH at a fixed lower price. We keep the premium either way; if the price falls below the strike we end up buying ETH at the strike — more expensive than the market at that point, so it only makes sense at a strike where we would be happy accumulating.

An option's price mostly reflects how much movement the market expects — that expectation, expressed as an annualised percentage, is implied volatility (IV). Converting every quote into IV puts a 3-week quote, a 3-month quote, a % premium and an APR all in the same unit, so they can be compared with each other and with the live listed market. A premium that looks generous in % terms can be stingy in vol terms, and vice versa.

Three tests, in order. First: is the quote's IV above the asset's recent realised volatility? Selling movement for less than the asset actually moves is a losing trade regardless of the headline yield — quotes below realised are blocked. Second: is the quote at least 3 vol points above the listed bid — what Deribit would pay us for the identical option right now? Less than that and the listed market is the better counterparty. Third: anything between the listed bid and roughly fair value is negotiable — defensible if the provider brings real operational advantages (no margin management, on-chain settlement, size).

Covered call: the downside of the asset is still fully ours — the premium is a small cushion, not protection. And if the price flies through the strike, the upside beyond it belongs to the buyer; that regret is uncapped, which is why strike choice matters more than premium size. Cash-secured put: if the market crashes we buy at the strike, above the new market price. The right frame: only sell puts at prices where accumulating would genuinely be welcome. Both: premium is compensation for risk, never free yield. If a yield looks too good, the volatility being sold is being underpriced somewhere.

The mid is where nobody trades. If we sell, we sell at the bid — so the bid is the honest benchmark for what a provider must beat. Rule inherited from the July research: when you must sell, use the bid; when you must buy, the ask; never the mark.

Then there is no market price to compare an offer with, so the hub builds a fair price from the asset's own movement. It measures how much COW actually moved over a window you choose (the last 7 to 180 days, or the 2-year typical level), and prices the same option at that volatility. That is the premium a seller needs to be paid for the movement COW really has.

An offer below fair pays for less movement than COW delivers: the buyer gets the better side. Two honest limits: the past does not guarantee the future, and the fair price uses one volatility for every strike, while puts far below spot normally deserve extra. Treat fair as a floor, not a target.

Two red flags. Implied below realised — sellers are being underpaid for actual movement; wait. Very low vol percentile — premiums are thin in absolute terms; capping upside for a small premium right before a vol expansion is the classic treasury mistake. The best moments to subscribe combine IV above realised and a decent absolute vol level, at strikes the treasury is happy to be exercised at.