Box spread
A box spread combines a bull call spread and a bear put spread on the same two strikes and expiry, so it pays exactly the strike width at expiry whatever the underlying does; its price today therefore implies an interest rate.
Four legs, one fixed payoff
A long box on strikes K1 < K2 is two vertical spreads bought together:
- a bull call spread: buy the K1 call, sell the K2 call;
- a bear put spread: buy the K2 put, sell the K1 put.
Read another way, it is a synthetic long at K1 (long call, short put) plus a synthetic short at K2. By put-call parity the underlying cancels out: at expiry the box pays K2 − K1 wherever the price settles. Above K2 the calls pay the width; below K1 the puts do; in between, each spread pays its part. Its delta, gamma and vega are all close to zero. What is left is time and money, which makes it the options market's own loan, sensitive almost only to rho.
The implied rate
Paying C today to receive the width W at expiry, D days away, is lending C at a simple annual rate of:
rate = (W / C − 1) × 365 / D
The buyer of the box lends; the seller of a short box receives C now and owes W at expiry, so the seller borrows at the same rate. A box priced above its width implies a negative rate: the lender is paying for the privilege.
A worked example
A 90-day box on BTC with strikes $90,000 and $110,000 pays $20,000 at expiry. Suppose, for illustration, that its four legs' mids add up to a cost of $19,760 (live boxes on these venues often price closer to their width, as the next section explains):
- At mid, the rate is (20,000 / 19,760 − 1) × 365 / 90 ≈ 4.9% a year.
- Now execute it. Each leg crosses half its bid-ask spread, say $50 a leg, and pays a trading fee of 0.03% of the underlying, about $30 a leg on a $100,000 forward. Four legs add $200 of spread and $120 of fees: the box costs $20,080, more than it pays. The lender's rate becomes (20,000 / 20,080 − 1) × 365 / 90 ≈ −1.6%.
- The borrower, selling the box, receives the bids less fees: $19,760 − $200 − $120 = $19,440, and pays (20,000 / 19,440 − 1) × 365 / 90 ≈ 11.7%.
The same box quotes a 4.9% rate at mid, but no one trades at mid. Four legs mean four spreads and four fees, and on a short expiry those costs are a large share of the tiny carry. A wider box (more payoff for the same four fees) and a longer expiry (more days to spread them over) bring the execution rate closer to the mid rate. A box under about a week out annualizes a few dollars of costs into absurd percentages.
Boxes on crypto venues
The options on these venues are European and cash-settled, so a box cannot be exercised early and its payoff really is fixed. Two caveats:
- Inverse options. On Deribit and OKX, the BTC and ETH options BOSS follows are settled in the coin. A box there costs coins today and pays (K2 − K1) / S coins at expiry, where S is the settlement price: an unknown number of coins, but worth exactly K2 − K1 dollars then. BOSS's dollar prices value a coin premium at the expiry's forward, so for a box the live table below re-values its cost at today's spot instead (cost × spot / forward): the dollars you give up today against the dollars you get back. That dollar rate should track the annualized basis, shown beside it. On Bybit, whose options settle in stablecoins, the payoff is a fixed dollar amount (see inverse vs linear).
- The carry is already in the forward. The interest rate crypto prices into options shows up as each expiry's forward premium over spot, and the venues price options on that forward without discounting the payoff. So a box at mid costs close to its width in BOSS's forward-valued dollars: on Bybit that implies a rate of roughly 0%, and on Deribit and OKX, re-valued at spot, roughly the basis. The basis reads the carry directly, with no spread to cross, so it is the cleaner number; at execution, four legs' spreads and fees make short-dated boxes' rates deeply negative.
A box also ties up margin on its short legs, and the venue's risk is still there: a fixed payoff is only as good as the venue that pays it.
Live on BOSS
A long box at each listed expiry at least 7 days out, built around the at-the-money strike at the widest width the builder allows (four strikes either side): what it pays, what it costs, and its implied annual rate at mid and at execution with fees, next to the expiry's annualized basis. On Deribit and OKX the cost is re-valued at today's spot, so the rate is the dollar rate of lending through the box. The gap between the two rate columns is the four legs' spreads and fees.
| Expiry | Strikes | Pays at expiry | Cost at mid | Rate at mid | Cost at execution + fees | Rate at execution | Annualized |
|---|---|---|---|---|---|---|---|
| 09 Oct 2026 (7d) | $82000.00 / $90000.00 | $8000.00 | $7971.17 | +18.6% | $8444.62 | -271.1% | +5.1% |
| 16 Oct 2026 (14d) | $82000.00 / $90000.00 | $8000.00 | $7941.05 | +19.2% | $8410.19 | -126.4% | +5.5% |
| 23 Oct 2026 (21d) | $82000.00 / $90000.00 | $8000.00 | $7919.62 | +17.6% | $8367.25 | -76.0% | +5.4% |
| 30 Oct 2026 (28d) | $82000.00 / $90000.00 | $8000.00 | $7962.62 | +6.1% | $8324.16 | -50.6% | +5.4% |
| 27 Nov 2026 (56d) | $83000.00 / $91000.00 | $8000.00 | $7898.11 | +8.4% | $8238.14 | -18.8% | +5.5% |
| 25 Dec 2026 (84d) | $80000.00 / $96000.00 | $16000.00 | $15795.94 | +5.6% | $16200.52 | -5.4% | +5.4% |
Rate = (pays at expiry ÷ cost − 1) × 365 ÷ days to expiry. A negative rate means the box costs more than it pays: lending at a loss.
Common mistakes
- Reading the rate at mid prices: crossing four spreads and paying four fees can turn a positive rate negative.
- Annualizing a box a day or two from expiry, where a few dollars of costs become hundreds of percent a year.
- Treating a box on coin-settled options as a fixed dollar loan without noting that it pays a variable number of coins.
- Forgetting that the short legs need margin and that the payoff depends on the venue being there to pay it.
Where this shows up on BOSS
Educational content, not investment advice. See the disclaimer.