How to Roll an Option
To roll an option, close the existing contract and open a new one with a different strike or expiry, usually as a single order. Treat it as opening a new position: it needs its own view, its own strike and its own deadline, not the previous trade's.
Rolling means closing one contract and opening another at the same moment, usually as one order. The mechanics are simple. The difficulty is that it lets a position continue without the decision that opening a position normally requires.
Before you start
A fresh view, because a roll is a new position and needs its own reason. The previous view either still holds, in which case say so, or it does not, in which case the roll is not the answer.
The net cost or credit of the roll, calculated before it is placed. Rolling out often costs money; rolling out and further away costs more.
An honest answer to whether you would open this new position from flat. This is the whole test, and it takes ten seconds.
The steps
1. Apply the from-flat test first
No existing position, no premium already spent, just the chart and the contract. If you would not open it, the roll is being driven by the old trade rather than by the market.
2. Write the new view down
Same requirement as any option purchase: how far and by when. “It still might work” is not a view, it is the absence of one.
3. Price the roll before placing it
The order shows a net figure. A roll that costs a debit is adding money to a position that has not worked, which may be right and should at least be a conscious decision.
4. Change one thing at a time
Rolling out buys time. Rolling the strike changes what has to happen. Doing both at once usually means the new position needs a larger move over a longer window and costs more.
5. Place it as a single order
Two separate orders leave you briefly with both positions or neither, at whatever prices the market offers in between. The combined order removes that.
6. Count the roll as a closed trade in the record
The first contract was closed at a profit or a loss, and that number belongs in the journal. A rolled position that stays open for months hides several completed trades inside one entry.
7. Set a limit on how many times you will roll
Two is a reasonable ceiling. Without one, the position can continue indefinitely, and the only thing being managed is whether the loss has been recorded yet.
How to tell it worked
The from-flat test was applied and passed, before the order was placed.
The net debit or credit was known before submission, not discovered afterwards.
The closed leg was recorded as 1 completed trade in the journal.
And this position has been rolled at most 2 times in total.
What each roll costs
Two spreads: one closing, one opening. On an illiquid contract that pair can be a large share of the remaining premium, and three rolls means six spreads paid on a position that has yet to work.
Liquidity in the new expiry is a separate question from the old one. Rolling into a quiet expiry solves a timing problem and creates an exit problem.
Rolling out against rolling down
Rolling out extends the deadline and keeps the view. It is the honest version, and its cost is usually a debit — you are buying more time for the same idea.
Rolling down or away changes what has to happen. The new strike is easier to reach, which is another way of saying the original view has been abandoned and replaced by a smaller one.
Doing both is where the position stops being reviewable. A different strike, a different date and a different net cost mean nothing about the new trade can be compared to the reasoning that started it.
Rolling a short position is a different decision
A short option rolled for a credit is being paid to extend the obligation. That is structurally different from a long position rolled for a debit, and it is why the same word covers two trades with opposite cash flows.
The credit is not evidence the roll was right. It is compensation for continuing to carry a risk you have already found uncomfortable, and taking it means agreeing to that risk for longer.
The from-flat test still applies, in the same form. Would you sell this contract today, at this strike, with this cash set aside, if you had no position? If not, the credit is a payment for staying in something you would not enter.
The original data
Of the 24,971 unique videos in research/search-study-corpus.jsonl, 0 mention rolling an option in
the title. Rolling in any sense appears in 9 at a median of 4,207, and options generally in 889 at
10,399. The counts come from site/corpus_count.py, which deduplicates by video id.
0 videos in 24,971 on the adjustment that keeps a losing options position alive. The entries have substantial coverage and the mechanism that converts one bad entry into six months of them has none — which is a reasonable description of where instructional attention goes generally.
The answer to the question on that chart is that an intact view is exactly when a roll is defensible — and it still has to pass the from-flat test at the new strike and the new price. If the answer to that is yes, roll it. If the answer is that you would not open it today, the view is not as intact as it feels.
When it fails
The failure is the perpetual roll, and it never presents itself as a decision. The contract approaches expiry, the view still feels right, and rolling costs a small debit. That happens again six weeks later, and again. A year on, the position has consumed several times its original premium in debits and spreads, the underlying has done nothing the original view described, and the loss has still not been recorded — which was the actual function the rolls were performing.
The second failure is skipping the from-flat test. The old trade then drives the new one.
A third is rolling out and down together. Nothing about it is comparable afterwards.
A fourth is two separate orders. The gap between them is uncontrolled.
A fifth is not recording the closed leg. Several trades hide in one journal entry.
And a sixth is no ceiling on rolls. Without one, the position outlives the reasoning.
Related
Rolling options covers the combined order and its legs. Options expiry is the deadline that prompts most rolls. And strike price is the other thing a roll can change.
The test that stopped me rolling badly was asking whether I would put this exact position on from flat, today, with no existing trade. If the answer is no, the only thing the roll achieves is that the loss does not appear on the statement yet — and it is larger by two spreads than it was before I moved it.
— Michael Whitman
This page is educational, not financial advice. Test every idea on your own charts before risking money.