Calendar spreads — selling time against time
Ingrid rents 5 rooms and sublets 10. A calendar spread sells fast-burning short options against a slower long base, but you can only sublet as many rooms as you rent. Four checks before every order, and a real liquidation of −0.2245 ETH show what happens when short options are left uncovered.
Every spread in the course so far has been vertical: same expiry, different strikes (5-6 Put SpreadPut SpreadLong put + short put. Cheaper protection than a plain put, but with limited coverage.Read the lesson →, 5-11 CollarCollarAn options strategy: long put + short call + the underlying asset held. Caps both potential loss and potential gain.Read the lesson →). A calendar spread flips the axis — same strike, different expiries. You buy a long-dated option and sell a short-dated one against it. You earn not from the direction of the price, but from the fact that the short-dated option loses time value faster than the long-dated one.
You hold a long-term lease on an apartment (105 days) and sublet the same apartment month by month. The monthly sublet, recalculated per day, costs more than your long lease — you collect that difference. But you can only sublet as many rooms as you rent yourself. One room more — and that is no longer a sublet, it is a promise you have nothing to back with.
What it is
| Term | Meaning |
|---|---|
| Calendar spread | Same strike, different expiries: long the far one + short the near one |
| Diagonal spread | Both the strike and the expiry differ |
| Base | The far-dated long option — your "long lease" |
| Layer | The near-dated short option at the same strike — the "sublet" |
There is one engine: time value does not decay in a straight line. The closer to expiry, the more time value an option loses in a single day (theta grows as the deadline approaches — 5-3). Numbers from a real book: a 14-day layer burns roughly 2.7 times faster per day than a 105-day base, a 49-day one about 1.5 times, a one-day one about 10 times. You sell what burns fast and hold what burns slowly.
Two roles, no third one
| Role | What it is | When it is opened |
|---|---|---|
| Base (long) | A quarterly pure strangle: long put + long call, both ~0.30–0.35 delta from spot | Only when DVOL is below 48–50 (vol is cheap) |
| Layers (short) | Monthly (and one-day) shorts at the same strike as the base | Only when DVOL is above 45 (the premium is worth it) |
- The base = the engine + the licence (the right to sell layers). The layers = the base's financing.
- Vol direction: buy cheap, sell expensive. Never the other way round.
- The base is a pure strangle with no same-expiry cap — there is no third "protective short" category. Why — see the section on the real campaign, where it cost money.
A book like this is run on an options exchange with margin — in the course that is DeribitDeribitA crypto options exchange. In our strategy this is where we buy put protection and sell covered calls.Read the lesson →, which does not hold an EU MiFID II licence: a Lithuanian trades through a third-country platform with no EU supervision. Details in lesson 5-4. ETH options there are coin-settled: 1 contract = 1 ETH, the premium and P&LP&LProfit & Loss — used to evaluate strategy results.Read the lesson → are counted in ETH (SOL_USDCSOLSolana's native token. Used to pay gas and as collateral.Read the lesson → is linear, contract of 10 SOLSOLSolana's native token. Used to pay gas and as collateral.Read the lesson →, premium in USDCUSDCA USD stablecoin issued by Circle. Regulated, audited monthly. Our primary stablecoin and loan asset in the Kamino/Aave strategies.Read the lesson →).
The coverage invariant — you check it, you don't feel it
Before any order, all four points must be green:
| # | Condition | What it protects |
|---|---|---|
| 1 | short puts ≤ base long puts | Every short put has a long underneath it |
| 2 | short calls ≤ base long calls | Every short call has a long above it |
| 3 | short call strike ≥ base call strike | In a rally the long call covers the short fully |
| 4 | short put strike ≤ base put strike | In a drop the long put covers the short fully |
At least one violated → no new orders. This is the difference between "covered" and "naked": every short has a long that covers it both above and below. The course does not teach naked options (5-4, 5-7).
Selling layers on margin is risky in itself. Without the coverage invariant and the capital rule (below) this is not a calendar spread — it is a naked short position with a pretty name.
The layer cycle
Entry conditions — all of them together
| Condition | Limit |
|---|---|
| Nearest layer | ≤ 7 days to expiry or already expired |
| MM after the new layer | below 65% |
| DVOL | above 45 |
| Coverage check | all 4 points green |
| Moratorium | no longer in force |
| Size | shorts ≤ base longs (puts and calls counted separately) |
| Spot near the strike (delta above 0.40) | the layer is skipped |
What happens at expiry
| Scenario | Action | Result |
|---|---|---|
| Short OTM (the normal beat) | Nothing | The short disappears, the premium stays, the longs are untouched → a window for a new layer. The default state, ~8–9 months a year |
| Short ITM | Wait until expiry (extrinsic burns in your favour), on settlement day — the pair | Sell N same-strike base longs + (while the old cap exists) buy back N cap shorts. Settlement is paid with realised long profit, the book shrinks proportionally |
| Short has burned ≥ 85% or ≤ 0.004 ETH remains | Buy back early | The remaining premium no longer pays for the risk; capacity is freed |
An ITMITMIn The Money — an option that has intrinsic value.Read the lesson → short is not closed early and not left open — it is paid off with the pair on expiry day. Selling a long on its own leaves naked shorts; buying back a short on its own throws away extrinsic that would still have burned in your favour.
The roll — the base is never held to expiry
With ~70 days left the whole book is moved to the next quarter. The order matters:
- New longs are bought first — coverage never breaks, not even for a day
- Strikes are reset to the new spot (0.30–0.35 delta), not copied
- Old longs: the winning side can ride to the pin, the dead one is closed for what is left
- Only then — new layers
If vol is high at roll time — a cash pause. The calendar does not pull.
The real campaign — with a liquidation and a minus
This is the author's real position, not a recommendation. The course executes no trades and holds no funds. All results, losses included, are published openly: dashboard.defiriskos.com/eth.html. State as of 2026-09-11, ETH $2,457.
The book (abridged)
| Instrument | Side | Qty | Days | Role |
|---|---|---|---|---|
| ETH-25DEC26-1500-P | LONG | 7 | 105 | base (put side) |
| ETH-25DEC26-2500-C | LONG | 5 | 105 | base (call side) |
| ETH-25DEC26-3000-C | SHORT | 5 | 105 | old cap — a mistake, carried until closed in pairs |
| ETH-12SEP26-2500-C | SHORT | 1 | 1 | layer (one-day) |
| ETH-25SEP26-2500-C | SHORT | 2 | 14 | layer |
| ETH-30OCT26-2500-C | SHORT | 2 | 49 | layer |
- Book net: delta −1.383 ETH · theta +$18.67/day · vega +1.41
- MarginMarginThe collateral deposit required to open a leveraged position or a sold option position.Read the lesson →: MM 28.4% ($1,317 / $4,637) · free $2,991
- PremiumPremiumThe price of an option. Paid by the buyer, received by the seller.Read the lesson → collected from layers: 0.0530 ETH ($130). One-day layers: 7 of them, 0.0384 ETH, 5/5 expired OTMOTMOut of The Money — an option with no intrinsic value, only time value.Read the lesson →
- ETH balance −0.110 (target +0.3–0.5) — the book is financed with USDCUSDCA USD stablecoin issued by Circle. Regulated, audited monthly. Our primary stablecoin and loan asset in the Kamino/Aave strategies.Read the lesson → collateral, interest is being paid
Results — openly
| Metric | Value (2026-09-11) |
|---|---|
| Net ETH accumulated (metric #1) | +0.0462 ETH ($113). Start 2026-08-21: −0.188 ETH → change +0.234 ETH |
| Realised since 08-01 | −0.0417 ETH (options −0.0361 · settlement −0.0045 · negative-balance interest −0.0011) |
| Liquidation 2026-08-19 21:07 | 5× 3000C forced buy-back, −0.2245 ETH. Cause: too little capital (~$2,500 against 10 shorts), not the strategy itself. Consequence: 14-day moratorium 08-19 → 09-02 |
| SEP/OCT layers | Sold when ETH was well below 2500 (avg. 0.006 / 0.0175 ETH) — now ATMATMAt The Money — an option whose strike sits at the current price.Read the lesson →, MTM deep in the red; only a pin saves them, if ETH ≤ 2500 on settlement day |
The base long 2500C was capped with a short 3000C — that turned the base into a vertical 2500/3000 spread, which stops earning above 3000. But the layers (short 2500C) keep losing in a rally. Above 3000 the short calls add up: 5 layers + 5 caps = 10 vs 5 longs — 5 contracts naked. The coverage invariant (point 2) was violated, and ~$2,500 of capital was not enough for ten shorts. The result — a forced buy-back and −0.2245 ETH.
The conclusion, written into the doctrine on 09-07: the base must be a pure strangle. The cap is now carried until closed in pairs; ~22% probability it will be ITMITMIn The Money — an option that has intrinsic value.Read the lesson → by December.
Risk rules — "paid for in blood"
| Rule | Meaning |
|---|---|
| Moratorium | Any liquidation → 14 days of no new shorts |
| Size is a constant | After a loss the size is not increased; growth only follows equity → base → layers |
| The MM ladder | below 65% for new orders · 80% → buy back the nearest layer's calls (not the longs!) · 90% → deposit capital NOW |
| Capital | MarginMarginThe collateral deposit required to open a leveraged position or a sold option position.Read the lesson → balance ≥ 2× MM when spot = strike; ≥ ~$300 for every short contract; deposit when the distance to liquidation is less than 12% or MM is above 65% |
| Collateral currency | = the book's currency. Target +0.3–0.5 ETH standing balance. A negative ETH balance costs ~9.6% APRAPRAnnual Percentage Rate — the annual interest rate WITHOUT compounding. 10% APR = +10% per year.Read the lesson →. Do not convert settlement proceeds into USD |
| Vol direction | Buy with DVOL below 48, sell above 45. Never the other way round |
| Monetising the insurance | Long put ITMITMIn The Money — an option that has intrinsic value.Read the lesson → → sell half by levels, not all of it |
Closing by the speed of the move
| Move | Action |
|---|---|
| Sharp ↑ (above 8% in 1–3 days) and MM above 80% | Buy back the nearest layer's calls. Do not touch the base longs — they are earning |
| Sharp ↓ | Close nothing — this is the best side (the long puts). Take put profit in parts by levels |
| Slow ↑, spot crosses the layer strike × 0.97 | Pre-settlement: compare "settlement + pair" vs "buy-back". Wait — extrinsic burns in your favour; the pair only on expiry day, if ITMITMIn The Money — an option that has intrinsic value.Read the lesson → |
The campaign runs in the same cross-margin account that also holds the SOLSOLSolana's native token. Used to pay gas and as collateral.Read the lesson → strategy's insurance (the SOLSOLSolana's native token. Used to pay gas and as collateral.Read the lesson → puts). An ETH margin crisis would liquidate the SOL insurance too. Separate strategies in the same account are not separate.
The campaign's goal is to accumulate ETH. Metric #1: net ETH accumulated = ETH balance + value of the longs − obligations of the shorts. Every decision is checked with one question: is that number larger afterwards?
An honest expectation
- Today's theta (~17%/month on capital, 2026-09-11) is a pin bonus — the shorts sit exactly ATMATMAt The Money — an option whose strike sits at the current price.Read the lesson →. That is not a steady state.
- From zero, by the doctrine (e.g. a MAR27 2300P/3300C base, 6 cycles), the pin scenario ≈ +0.11 ETH per unit ≈ ~4%/month — a scenario from the campaign analysis, not an expectation for the reader.
- The strategy = a range bet with a limited loss (net debit). Realised P&LP&LProfit & Loss — used to evaluate strategy results.Read the lesson → since 08-01 is negative.
- One-day layers do not recover losses — they are the same bet, only shorter.
Summary
| Aspect | Meaning |
|---|---|
| Structure | Long quarterly strangle (base) + short monthlies at the same strike (layers) |
| Engine | The near option burns ~2.7× faster than the 105-day base (14-day layer) |
| Vol rule | Buy with DVOL below 48, sell above 45 |
| Coverage | The 4-point invariant — violated → no orders |
| ITMITMIn The Money — an option that has intrinsic value.Read the lesson → expiry | The pair on settlement day, never separately |
| RollRollRolling an option — closing the old one and opening a new one with a later expiry.Read the lesson → | With ~70 days left, new longs first |
| Capital | ≥ 2× MM at the pin, ≥ ~$300/short, 14-day moratorium after a liquidation |
| Real campaign | +0.0462 ETH net · −0.0417 ETH realised · −0.2245 ETH liquidation (2026-09-11) |
The task — run the real book through the coverage invariant and the capital rule by hand (~20 min). All on paper: no buying, no selling.
This is a PAPER check — we are learning to read a book, NOT to trade. Selling layers on margin without coverage and the capital rule is a naked position. DO NOT BUY OR SELL ANYTHING.
This is learning, not investing — use only small amounts you treat as tuition.