A stock trade ends when you sell it. A short option is different: it ends on its own, and most of its money arrives before the end. So every open cash-secured put or covered call quietly asks the same question all day long — is the premium still left in this contract worth the risk still left in it?

The number that frames the question is percent of premium captured:

captured % = (credit received − cost to close today) ÷ credit received

Sell a put for $1.40 and it can now be bought back for $0.35 — you have captured 75% of everything the trade will ever pay. The last $0.35 is all that remains, and the obligation runs until expiry either way.

This page is about the trade that is paying you. When the stock has gone through your strike and the row is red, the questions change entirely — that half lives in The Underwater Put.

The popular rule — buy back a short option once half the credit is captured — works because of an asymmetry in how these trades pay. Premium comes in fastest early, while the risk you are being paid for shrinks much more slowly. By the time 80% of the credit is in, the remaining 20% pays a fraction of the original rate for nearly the same exposure to a sudden move.

The more precise version of the rule is remaining yield on remaining time: what does the premium still in the contract pay, per day, on the collateral still locked up?

remaining annualized = (time value remaining ÷ collateral) × (365 ÷ days to expiry)

A $65 put sold at 21 days for $1.30 opened at roughly 2% on collateral — about 35% annualized. Ten days later it marks $0.40: the remaining $0.40 over the remaining 11 days works out to about 20% annualized. Still respectable. But at $0.15 with a week left it has fallen under 10% — the capital is now earning a fraction of what a fresh trade at your own entry standards would pay. That comparison, against the rate you demand at entry, is what the 50% shorthand is approximating.

Notice what this catches that a flat 50% rule misses in both directions: a trade only 40% captured whose remaining yield has gone dead, and a trade 60% captured that is still paying handsomely for its remaining days.

The 50% rule was popularized for 30–45 day trades, and it is tenor-dependent. On a monthly, time decay is slow in the first weeks and the early exit gives up little. On a 5-day weekly the shape inverts: theta is steepest in the final two days, which means a weekly sold Monday earns its best-paid days on Thursday and Friday. Closing it Wednesday at 50% hands back exactly the part of the trade that pays the most per day of risk.

This is why the two Opty AI books manage winners differently on purpose. The monthly book takes profits early on a sliding target. The weekly book never closes early for profit — its early exits are defensive only — because at that tenor the end of the trade is the point of the trade.

Near expiry, the remaining-yield math develops a blind spot: dividing by one or two days makes tiny numbers look large. A nickel of remaining premium on $5,000 of collateral with two days left annualizes to about 18% — which reads as “hold” — while what is actually on the table is five dollars against the full weight of a bad headline landing on the strike.

So the yield test carries a companion condition: below some absolute remaining premium — a nickel, a dime — the trade has paid everything it meaningfully can, whatever the annualized figure says. The industry has effectively institutionalized this: most major brokers waive the contract fee entirely on buybacks at $0.05 or less. A closing order that costs nothing to place is the market's way of saying the trade is over.

Every option price splits into two parts. Intrinsic value is what the contract would be worth exercised right now — for a put, the strike minus the stock price when the stock is below the strike, otherwise zero. Extrinsic value is everything above that: the part of the price that exists only because time remains.

Extrinsic is the number that times both early assignment and a roll, and going in-the-money is not the trigger — extrinsic running out is. Consider a $50 put with the stock at $44.10, so $5.90 of intrinsic value:

• Marking $7.70, the contract holds $1.80 of extrinsic. The market is still paying $1.80 for the time remaining, and an early exercise would throw that away — so assignment before expiry is unlikely, and a roll here spends value the position still owns.
• Marking $6.20, extrinsic has shrunk to $0.30. There is almost nothing left for the other side to forfeit by exercising, so early assignment becomes live — and a roll now gives up almost nothing. This is the window where deep in-the-money positions get repositioned.

Covered calls add one calendar item: dividends. When a short call's extrinsic value falls below the amount of an upcoming dividend, exercising the day before the ex-dividend date becomes the profitable move for the call's owner — capturing the dividend costs them less than the time value they give up. Short calls on dividend payers deserve a glance at the ex-date whenever extrinsic is thin.

Rolling — buying back the current contract and selling another at a later expiry, often a different strike — feels like one continuous position. It is not. It is a close, with a real realized result, followed by a brand-new trade that deserves the same scrutiny as any other entry: would this strike, at this expiry, on this stock, pass your entry rules if you were looking at it fresh today?

The honest arithmetic is the net credit: new premium collected minus the cost of the buyback. A roll that nets a credit extends the trade while continuing to pay for the risk. A roll done mainly so a losing number never appears in the journal — collecting pennies to defer a loss for months — is the most common self-deception in the wheel. The extrinsic test above is the antidote: a roll timed when extrinsic is nearly gone forfeits almost nothing, whereas a roll made because being in-the-money feels uncomfortable pays for comfort with value the position still holds.