Expiry Day Mechanics for Systematic Options Traders
What actually changes in the final hours: gamma, theta, pin risk and the settlement rules that turn a profitable position into an unexpected liability.
Expiry day is where option strategies that backtest beautifully go to die, and the reason is rarely the direction of the underlying. It is that three separate things — gamma, liquidity and settlement — all change character in the last few hours, and a strategy calibrated on normal days is operating outside its assumptions.
Gamma is the whole story
Trend chart
Convexity risk rises as remaining time collapses
A relative index visualises the coupled expiry-day trade-off: less time value remains while near-the-money gamma risk accelerates.
On an illustrative 0 to 100 scale, near-the-money gamma risk rises from 20 at 09:30 to 35 at noon, 70 at 14:00 and 100 near 15:00. Remaining time value falls from 100 to 78, 45 and 5 over the same checkpoints. The two series show why accelerated theta cannot be separated from accelerated gamma.
Relative conceptual indices, not calibrated Greeks. Actual gamma and theta depend on spot, strike, implied volatility and the exact time remaining.
Gamma measures how fast delta changes. As time to expiry approaches zero, an at-the-money option’s delta has to converge to either 0 or 1 — the option either finishes worthless or finishes as a unit of the underlying. Compressing that transition into a few hours means gamma near the money grows without bound.
The consequence for a short-gamma strategy, which describes almost every premium-selling structure: the position’s directional exposure grows the more the underlying moves toward your strike, and it grows faster the closer expiry is. You are short an exposure whose size increases exactly when it is hurting you.
Time
Gamma near ATM
Delta after a 0.3% move
Practical effect
09:30
Elevated
Modest
Adjustable with a hedge
12:00
High
Noticeable
Adjustment cost rising
14:00
Very high
Large
Hedging becomes reactive, not preventive
15:00
Extreme
Near-binary
Effectively a directional bet you did not choose
How the same short straddle behaves at different times on expiry day
Theta is not a constant drip
The convenient mental model — an option loses a fixed amount of value per day — is wrong in the way that matters most. Time value decays proportionally to the square root of remaining time, which means the decay rate accelerates as expiry approaches. Most of an ATM weekly option’s remaining time value disappears on the final day, and a large fraction of that on the final afternoon.
This is what makes expiry-day premium selling superficially attractive: the theta capture per unit of time is at its maximum. It is also the same fact that makes gamma extreme, because gamma and theta are two views of the same convexity. There is no version of this trade where you collect accelerated theta without carrying accelerated gamma. Backtests that show otherwise have a fill model problem.
Accelerated theta and extreme gamma are the same number seen from two directions. Any strategy claiming the first without the second is mispricing its fills.
Pin risk
Decision tree
The expiry decision near a short strike
Settlement type changes the consequence of carrying ambiguity through the close.
If spot is outside the predefined pin band, follow the normal exit rule. If spot is inside the band, identify settlement type. Index options settle in cash but still carry settlement-price uncertainty, so close or explicitly accept it. Stock options settle physically, creating a delivery or receipt obligation and associated margin, so the default systematic action is to close before the cutoff.
The pin band and cutoff time belong in configuration and should trigger independently of the strategy signal.
When the underlying settles very close to a strike you are short, you do not know whether that option will be exercised until settlement is determined. For index options this is a cash-settlement question and resolves cleanly against the settlement price. For stock options it is a delivery question, and the difference matters enormously.
Handling pin risk systematically
Define a pin band around each short strike — a percentage of the underlying — and close positions inside it before the close rather than carrying the ambiguity.
Do this on a time trigger, not a discretionary one. The whole point of automation is that this decision is not being made at 15:25 under pressure.
Budget the cost of that exit into the strategy. Closing a near-worthless option at the worst spreads of the day is a real cost, and it is the price of not carrying settlement uncertainty.
Physical settlement on stock options
Timeline
Physical-settlement risk starts before expiry day
Margin and operational checks must escalate while there is still enough liquidity and time to close or roll.
Several sessions before expiry, classify every stock derivative as physically settled and identify open obligations. As expiry approaches, broker margin requirements ramp. Before the final session, decide to close, roll or intentionally fund delivery. On expiry day enforce the cutoff and close any position lacking an explicit delivery decision. After settlement reconcile cash, shares and charges.
Exact broker ramp schedules and cutoffs vary. Configure them from current broker and exchange rules rather than hard-coding the illustrative day labels.
This one produces the largest single unpleasant surprises in Indian derivatives, because the failure is silent until it is expensive.
Stock futures and stock options on NSE settle by physical delivery. An in-the-money stock option held to expiry does not settle in cash — it becomes an obligation to deliver or receive shares, with the corresponding margin requirement in the days around expiry and the full delivery value at settlement. A short in-the-money call on a stock, left open through expiry, obliges you to deliver shares you may not hold.
Rules worth hard-coding
Never carry a stock option into expiry without an explicit decision
Default to closing. Delivery should be a choice you made, not a state you arrived in.
Set the close trigger by days-to-expiry, not by expiry day
The margin ramp begins before expiry, so an expiry-day exit is already late.
Treat index and stock options as different instruments in your risk config
Cash settlement and physical settlement are different products that happen to share a payoff diagram.
Alert on any open physically-settled position inside the ramp window
Independently of the strategy. This is a portfolio-level check, not a strategy-level one.
Liquidity in the final hour
ATM strikes on index weeklies stay liquid essentially until the close. Everything else does not. Strikes several increments away, which had a workable market at midday, can be quoted a rupee wide on a two-rupee premium by 15:00 — see slippage and impact cost for why percentage-of-premium is the metric that matters here.
For a strategy that plans to exit its wings, this is the binding constraint. The exit price assumed in the backtest, drawn from a dataset that recorded whatever trades happened, is not the price available to an order arriving at 15:05.
ATM
Stays tradeable
Through to the close on index weeklies
±2 strikes
Degrades noticeably
Spread widens as a share of premium
Far wings
Effectively untradeable
Spread can exceed the premium itself
What this means for strategy design
None of the above says expiry-day strategies do not work. It says the assumptions that hold on a normal Tuesday do not hold, and a strategy has to be designed for the regime it trades in.
1
Treat expiry day as a separate strategy
Different parameters, different position limits, different exit rules. Not the same strategy with the same config running on a different date.
2
Cap short gamma explicitly
Not just position size. A gamma limit is what stops a position from becoming directional without a trade being placed.
3
Use time-based exits, not only price-based ones
A hard "flat by 15:00" rule removes both pin risk and the worst of the liquidity problem, at a known cost.
4
Backtest expiry days separately
Aggregate statistics across all days hide the expiry-day distribution completely, and it is a different distribution.
5
Assume the worst fills of the week
If the strategy survives full-spread fills at 15:00 prices, it is robust. If it needs mid, it is not a strategy.
Frequently asked questions
Why does gamma increase so much on options expiry day?
Because an at-the-money option's delta must converge to either 0 or 1 by expiry — the option finishes worthless or finishes as a unit of the underlying. Compressing that transition into a few hours means the rate of change of delta, which is gamma, grows very large near the money. A short-gamma position therefore acquires directional exposure that grows fastest exactly when the underlying is moving against it.
Is theta decay linear as expiry approaches?
No. Time value decays roughly with the square root of remaining time, so the decay rate accelerates as expiry approaches. Most of an at-the-money weekly option's remaining time value disappears on the final day. That accelerated theta is the same convexity as the extreme gamma — you cannot collect one without carrying the other.
What is pin risk in options?
Pin risk is the uncertainty when the underlying settles very close to a strike you are short: you do not know whether the option will be exercised until settlement is determined. For index options it resolves in cash against the settlement price. For stock options it becomes a delivery question, which is a materially larger problem. Systematic strategies usually define a band around each short strike and close inside it on a time trigger.
Do Indian stock options settle in cash?
No. Stock futures and stock options on NSE settle by physical delivery. An in-the-money stock option held to expiry becomes an obligation to deliver or receive shares, with margin requirements that ramp up in the sessions before expiry. This is the most common source of large unpleasant surprises for automated strategies, because the margin escalation begins before expiry day itself.
Should I backtest expiry days separately?
Yes. Gamma, theta, liquidity and settlement all behave differently on expiry day, so aggregate statistics across all sessions average away a distribution that is genuinely distinct. Treat expiry as a separate strategy with its own parameters, position limits and exit rules rather than the same configuration running on a different date.