$330 to buy back a 10C. $220 received for a 12C written immediately after. The old leg realized −$231; the new premium covered only 67% of the buy-back. Including the original $99, the whole cycle netted −$11.
That is a roll. It did not get me even. It got me a later accounting date.
1. A Roll Does One Thing
A roll is two actions in one ticket:
- Buy back the old leg (pay money; realize a loss or lock a gain).
- Sell a new leg (collect money; change expiry or strike).
Of the four variables you can only change two: expiry (roll out) and strike (roll up / roll down). Contract count usually stays.
So a roll is never a strategy. It is a repricing: you use the new leg’s cash to buy back the time and space the old leg lost.
2. Three Payments
A roll looks free. It isn’t. You pay three times:
① Spread. Buy back at the ask, sell at the bid — two crossings. On a book like TRON’s $0.05 / $0.50, closing 23 contracts costs $1,035 in spread alone, 11.6% of the position (the illiquidity premium).
② Time. The new leg expires later, so your capital is locked longer. RGTI traded 2 days for 93 — those extra 91 days are the price of not being able to touch that risk.
③ Position. The most ignored payment. Rolling down moves the strike closer to spot and raises assignment probability: RGTI went from Δ −0.98 to Δ −0.58. It sounds like a downgrade; −0.58 still means roughly a 60% chance you eventually take delivery.
Three payments buy you a later deadline and possibly a bigger obligation.
3. What a Qualified Roll Looks Like
Two real cases side by side:
| Item | NOK 10C→12C | RGTI 14.5P→14P |
|---|---|---|
| Buy-back cost | $330 | $468 |
| New-leg proceeds | $220 | $552 |
| Coverage | 67% (short $110) | 141% |
| Time change | ~5 weeks | 2 days → 93 days |
| Verdict | Failed | Qualified, with a cost |
The test is one line: new-leg proceeds ≥ buy-back cost (coverage ≥ 100%). NOK’s 67% means I paid $110 to avoid being called — that isn’t a roll, it’s buying peace of mind.
RGTI’s 141% and $84 net look good; but it pushed assignment to January, and the new leg was still ~15% in the money. Qualified does not mean cheap.
4. Four Conditions, In Order
- Money: new premium ≥ buy-back cost. If not, you are net paying — ask what you bought.
- Calendar: the new expiry must not cross earnings. NOK nearly missed the earnings calendar; that was the lesson from −231 out, +220 in — order matters.
- Worst case: after the roll, is the worst case larger or smaller? Larger is not a roll, it is adding.
- Exit: what is the buy-back trigger on the new leg? A roll without a trigger hands the decision to the expiry date.
All four pass, it’s a roll. Fail the first, and it’s paying to postpone a loss.
5. Four Ways to Die
One: rolling bigger every time. Each roll pushes the strike further down, the obligation gets heavier, until you are short a deep in-the-money contract with two choices: take delivery or realize a bigger loss.
Two: missing your own hard deadline. I set a hard deadline on RGTI, missed it by a day, the stock fell another 5%, and the buy-back was about $114 worse than planned. The expensive part of a roll isn’t the spread, it’s hesitation.
Three: forgetting earnings. A new leg that crosses an earnings date trades “more time” for “bigger gap risk.” Check the calendar before the price; it matters more.
Four: rolling illiquid names. No bid to buy back, no buyer for the new leg — you eat the spread. Being able to fill matters more than which strike you roll to.
6. How to Book a Roll
This is where people fool themselves: a roll hides the loss inside the new leg, and the screen looks like you only lost time.
Book it in three parts:
- Old leg: buy-back price − entry price = realized P&L (NOK: −$231).
- New leg: premium received, booked separately; not profit until closed.
- Whole cycle: include the original premium (NOK: −$11).
Book only the first and you feel robbed; book only the second and you feel rich. Both together is the truth.
7. When Not to Roll
- The thesis is broken. A roll defers; it does not repair. Sell the stock.
- Assignment was the plan. Let it happen; don’t pay net money to roll down and avoid shares.
- The new leg is less liquid than the old one. You are trading one unsellable contract for a worse one.
- You just don’t want to face the loss. That is psychology, not strategy — and the market charges the most for psychology.
8. Seven Questions Before You Roll
- Is new premium ≥ buy-back cost? What is the coverage ratio?
- Does the new expiry cross earnings or an event?
- After the roll, is the worst case bigger or smaller?
- What does assignment probability become at the new strike?
- Have I written the buy-back trigger for the new leg?
- Is the new leg more or less liquid than the old one?
- If it drops another 10% tomorrow, do I roll again or take the loss?
9. One Line
A roll doesn’t remove risk. It stretches the installment plan.
Every roll should answer: how much did I pay, how much time did I buy, and how big is the obligation I took on.
Related: Selling puts: win rate vs payout · The illiquidity premium · Selling into earnings
Data & Sources
- Updated 2026-11-13 | First published 2026-11-13
- NOK 10C→12C: buy-back $330 (realized −$231), new premium $220, original $99, see the roll review
- RGTI 14.5P→14P: bought back 2 at $2.34, sold 2 JAN 2027 at $2.76, net +$84, old leg −$390.74, see the deep roll
- TRON spread: $0.05 / $0.50, 23 contracts → $1,035 ≈ 11.6% of the position, see the illiquidity premium
- Delta-to-probability mapping is a rough approximation, not an identity
Risk Boundaries
Structure and costs only; not investment advice. Rolling defers and can enlarge risk: longer lock-up, higher assignment probability, worse liquidity. A roll can buy time; it cannot buy direction.