Options Risk Engineering

Everything below is a consequence of that sentence. A trader who only thinks "I'm bullish, buy a call" is holding four exposures while managing one. A trader who thinks "I want +delta, I want the move soon, and I want IV to hold up" has stated a precise position that can be sized, hedged, and evaluated. The second trader loses less often for reasons that have nothing to do with being right about direction.

The learning path runs: pricing intuition → implied volatility → Delta → Gamma → Theta → Vega → how they interact → IV Rank and Percentile → skew and term structure → earnings and IV crush → Greeks of spreads → position Greeks → strategy selection → management → practical rules.

Five levels of mastery

Options knowledge stacks. Each level is useless without the one beneath it and each one changes what the level above means.

Pricing mechanics

Intrinsic and extrinsic value

Every option price is two numbers added together. Intrinsic value is what the option is worth if exercised right now: for a call, max(S − K, 0); for a put, max(K − S, 0). Extrinsic value (time value) is everything else, and it is the part that time and volatility act on. At expiration extrinsic value is exactly zero, which is why theta exists.

Notice that the ATM option carries the most extrinsic value. That is where uncertainty is greatest, so that is where time decay and volatility sensitivity are greatest. Deep ITM options behave like stock; deep OTM options behave like lottery tickets; ATM options are where the Greeks live.

Why two options with the same strike can have different prices

Same strike, same expiry, different day: the market's expectation of movement changed. Same strike, different expiry: more time means more possible paths, so more extrinsic value. Same strike, same expiry, call versus put: put-call parity ties them together (C − P = S − K·e^(−rT)), so any difference beyond that is carry, dividends, or hard-to-borrow cost, not "mispricing".

American versus European exercise

US equity options are American: exercisable any day. Index options (SPX, NDX) are European: exercised only at expiry and cash-settled. This matters for early-assignment risk on short calls ahead of dividends and short puts that go deep ITM, and for the fact that European index structures never have assignment surprises.

Implied volatility

Forget the formula for a moment. IV is the market's expectation of how much the stock might move. Precisely, it is the volatility number that, plugged into a pricing model, reproduces the option's current market price. The model does not produce IV; the market does, and the model just decodes it.

The stock hasn't changed. The strike hasn't changed. The expiration hasn't changed. The only difference is how much movement the market is pricing in.

IV does not tell you direction

XYZ at $100 with 80% IV does not mean the market expects the stock to rise. It means the market is pricing a lot of potential movement: $100 → $130, or $100 → $70, or even $100 → $105 if the feared event fizzles. Direction comes from positioning and from your thesis, never from IV itself.

Why options get expensive when IV rises

Think of insurance. A house in a safe neighborhood is cheap to insure; a house next to a wildfire zone is expensive. When the market expects large moves, risk rises and premium rises with it. The first rule to memorize:

The trap that catches new option buyers

XYZ = $100. You buy the 30-day $105 call for $3.00. The stock doesn't move. You think you've lost nothing, but if IV fell substantially the option might now be $1.80. Nothing happened to the stock; something happened to the price of uncertainty. That is vega at work, and it is why the Greeks are not optional.

Historical, realized, and implied

There are two different volatility questions. How volatile has the stock actually been? That is realized (historical) volatility, measured from past returns. How much volatility are options pricing right now? That is implied volatility. The gap between them is where volatility trading lives: if options price 60% but you believe the stock will move like 30%, selling volatility may have an edge; if options price 30% but an event could produce 60% realized, buying volatility may. This is far more sophisticated than "buy calls when bullish".

Expected one-standard-deviation move over a horizon: S × IV × √(days/365). A $600 stock at 24% IV over 30 days implies roughly ±$41. Compare that to what your thesis says and you have a magnitude test for the trade.

IV Rank and IV Percentile

"IV is 60%, so options are expensive" is incomplete. Expensive relative to what? 60% could be extreme for a utility and cheap for a biotech. So we normalize.

IV Rank places today's IV between its 52-week low and high: (IV − low) / (high − low). If IV ranged 20%–80% over the year and sits at 68% today, IV Rank ≈ 80. IV Percentile asks a different question: on what fraction of days in the past year was IV below today's level? A single spike can inflate the range and make IV Rank look low while IV Percentile is still high; reading both prevents being fooled by one outlier day.

Skew, term structure, and IV crush

Volatility skew

Options at different strikes on the same expiry carry different IVs. In equities, downside puts almost always carry higher IV than equidistant upside calls, because crashes are fast and violent while rallies grind, and because institutions structurally buy put protection. Plotted across strikes this is the skew (or the "smile" when both wings are elevated, as in currencies). Practical consequences: put credit spreads collect more premium per unit of distance than call credit spreads; risk reversals and jade lizards exist to harvest that asymmetry; and a steepening skew is itself a fear signal.

Term structure

IV also differs by expiry. Normally the curve is upward sloping (contango): longer-dated options carry slightly higher IV because more can go wrong. Around events it inverts (backwardation): the expiry that contains earnings carries far more IV than the ones beyond it. Calendar and diagonal spreads are term-structure trades, and they only make sense once you can read this curve.

IV crush

Before earnings the front-month IV might be 80%. The morning after it is 35%. The uncertainty resolved, so the price of uncertainty collapsed. If you bought a call at 80% IV, the stock can move in your direction and the option can still lose money. The expected move implied by the straddle price (roughly straddle ÷ S) is the market's estimate of the earnings jump; your trade needs the stock to exceed it, not merely to move.

The Greeks: four forces on an option

Delta

A call with delta 0.50: the stock goes $100 → $101, the option gains about $0.50, so a $2.00 option becomes roughly $2.50. Stock falls $1, option falls to roughly $1.50. Delta also converts options to share-equivalents: ten 0.30-delta calls behave, locally, like 300 shares. But delta is not constant, and that is where gamma comes in.

Gamma

Delta 0.50, gamma 0.05. Stock up $1: delta becomes 0.55. Another dollar: 0.60. Another: 0.65. Gamma is the rate at which delta changes, and it is why an option can suddenly become extremely powerful as the underlying approaches the strike. Compare three calls with the stock at $100: the $50 call (delta ≈ 0.99) has nowhere to go; the $150 call (delta ≈ 0.05) stays near zero unless something dramatic happens; the $100 call (delta ≈ 0.50) can run 0.50 → 0.70 → 0.90 → 1.00. ATM options have the highest gamma, and short-dated ATM options have extreme gamma, which is why 0DTE contracts behave violently.

Theta

Theta of −0.08 means about $0.08 per share, or $8 per contract, leaves the option every day, all else equal. Decay is not linear: it accelerates into expiry for ATM options and is front-loaded for far OTM ones. Buyers can get direction right and still lose because they were too slow; sellers collect this rent in exchange for taking on gamma risk. Theta and gamma are the same trade seen from opposite sides.

Vega

Vega of 0.15 means a one-point rise in IV (say 24% → 25%) adds about $0.15 to the option. Vega is largest for ATM options and grows with time to expiry, which is why long-dated options are volatility instruments first and directional instruments second. Earnings trades are vega trades whether the trader knows it or not.

Greeks scenario lab

The chat proposed a concrete drill: QQQ at $600, a 30-day $610 call at 24% IV. Move the stock, let days pass, shock IV, and watch what the option does. The numbers here are computed live with Black-Scholes (risk-free 4%, no dividends), so they replace the chat's illustrative placeholders with real values. The option's entry price is frozen from the starting inputs so the P&L attribution below stays honest.

Try the four drills from the chat: stock $600 → $605 (delta plus a little gamma); IV 24% → 30% with nothing else changed (pure vega); five days pass (pure theta); stock $600 → $590. Then combine them and notice that the attribution lines rarely add up exactly to the actual change. The residual is the cross-effects: gamma, and the way vega and theta themselves shift as the stock moves. That residual is what "the Greeks interact" means.

The exposure table

Don't think "I'm buying a call." Think "what exposures am I buying?" A long call is simultaneously long direction, long acceleration, short time, and long volatility. This table is worth memorizing; every spread in the library below is built by adding rows of it together.

Compare a long call with delta +0.55, gamma +0.06, theta −0.08, vega +0.15 against a short put with delta +0.30, gamma −0.04, theta +0.07, vega −0.14. Both are "bullish". The first needs the stock to move up quickly while IV stays firm. The second needs the stock merely not to fall, and gets paid by the calendar and by falling IV. These are completely different bets that a direction-only trader would call the same thing.

Spreads as Greek arithmetic

A bull call spread (buy the $100 call, sell the $110 call) isn't a "bullish strategy"; it is two rows of the table added together. The long call contributes +delta +gamma −theta +vega. The short call contributes −delta −gamma +theta −vega. The net is still +delta, +gamma, −theta, +vega, but every one of those is smaller than owning the $100 call alone. You cancelled part of each exposure and gave up part of the upside to do it. That is what a spread is: partial cancellation of Greeks in exchange for a capped payoff.

A put credit spread (sell the $100 put, buy the $90 put) nets +theta, −vega, −gamma, with modest +delta. The bet: the stock won't move too far against me, time will pass, and volatility won't explode. Premium selling works beautifully in calm markets and turns ugly when volatility expands, because −vega and −gamma are the two exposures that hurt most in a fast sell-off, and you are short both.

Once you see spreads this way, every strategy in the library becomes readable at a glance: look at the sign of each net Greek and you know what has to happen for it to make money and what will hurt it even if you are right about direction.

Strategy library

Twenty-four structures, from the basic four to the ones that only make sense once you read skew and term structure. Every diagram is generated from the legs: the model stock is at $100 with 30% IV and 45 days to expiry, each leg is priced with Black-Scholes, and max profit, max loss, breakevens and net Greeks are computed from the resulting position rather than typed in. The solid line is P&L at the front expiry; the dashed line is P&L today, which is what you actually see on your screen the day after you open the trade. Use Underlying IV to see how the same structure looks when volatility is cheap or expensive.

The decision framework

Before any trade, answer six questions in order. The strategy is the output of the answers, never the starting point.

Worked examples from the chat: bullish + big move + soon + low IV → long call. Bullish + moderate move + high IV → call debit spread, or better, a put credit spread that gets paid for the high IV. Neutral + high IV → premium-selling structures (iron condor, strangle). Bullish + high IV + willing to own shares → cash-secured put. Bullish + already own shares + high IV → covered call.

Selection matrix: direction × IV regime

The matrix is a starting point, not a rule. A low IV Rank on a stock whose realized volatility has collapsed even further is not cheap; a high IV Rank into a binary event is not necessarily expensive. Question 4 in the framework compares IV to your forecast of realized volatility, and that comparison overrides the rank.

Position and portfolio Greeks

Each option's Greeks are per share. A position's Greeks are Greek × contracts × 100 × sign, summed across legs. A portfolio's Greeks are summed across positions, ideally beta-weighted to one index so that +delta in AAPL and −delta in XOM can be netted honestly.

Instead of "I have five bullish trades", the professional statement is: "My book has +$85,000 of SPY-equivalent delta, negative gamma, +$2,400 of daily theta and −$18,000 of vega per one-point rise in IV." Every item in that sentence is a decision. Negative gamma with positive theta means the book earns rent and loses on a fast move; negative vega means a volatility spike hurts across every position at once, which is exactly when correlations go to one. Concentration risk in options is not just too much in one name; it is too much of one Greek.

Delta hedging and gamma scalping

A long straddle is delta-neutral at entry and picks up delta as the stock moves (positive gamma). Selling shares as the stock rises and buying them back as it falls locks in that gamma as realized P&L; the cost is the theta you pay every day. Gamma scalping is profitable when realized volatility exceeds the IV you paid. Sellers of premium run the mirror image: they collect theta and are forced to buy high and sell low when they hedge, which is fine as long as realized stays below implied. This is the mechanism behind the volatility risk premium: historically implied has exceeded realized more often than not, which is why systematic premium selling can work, and why it pays for that edge with occasional severe losses.

Stress testing

Shock the book on three axes at once: price ±5%, ±10%; IV ±10 points, +30 points; time +1 day, +5 days. The scenario that hurts is almost always "down 10%, IV +30" for a short-premium book, and "flat, IV −10, +5 days" for a long-premium book. Size positions so the worst plausible scenario is survivable, and remember margin requirements expand in exactly that scenario.

Trade management

Taking profits and cutting losses

Short premium: closing at 50% of max profit typically captures most of the theta while avoiding the gamma-heavy final weeks; many practitioners close or roll at 21 days to expiry regardless. Long premium: define the exit before entry, because a long option's edge decays daily. A useful rule is to exit when the thesis has been invalidated, not when the P&L reaches a round number.

Rolling

Rolling is closing one position and opening another; it is not a way to avoid a loss, only to change its shape. Roll out (later expiry) for more time and more credit; roll up or down (different strike) to follow the stock; roll only for a net credit when defending a short position, otherwise you are adding risk to a losing trade. Roll a covered call up-and-out when the stock runs through the strike and you still want the shares.

Adjusting spreads

An iron condor whose put side is tested can be defended by rolling the untested call side closer (collect more credit, reduce upside room) or by rolling the tested side out in time. A short strangle can be converted into an iron condor by buying wings when IV spikes. The general principle: adjustments should reduce the Greek that is hurting you, which usually means cutting delta first and gamma second.

Assignment, early exercise, and pin risk

American-style short calls get assigned early when a dividend exceeds the remaining extrinsic value; short puts get assigned early when deep ITM with little extrinsic left. Assignment on one leg of a spread leaves you with stock plus the other leg; it is rarely a disaster if you have the buying power, but it can be if you don't. Pin risk: the stock closes right at your short strike on expiration day and you don't know until Monday whether you were assigned. Close or roll short options that are near the money on expiry day rather than letting the auto-exercise rules decide for you.

When you are right and still lose

Direction right, IV crushed: vega loss exceeded delta gain. Direction right, too slow: theta ate the extrinsic before the move came. Direction right, move too small: bought OTM when the expected move said ATM. Direction right, spread too tight: gave up the upside for a credit that didn't compensate. Each is a diagnosis you can only make if you looked at the Greeks before entry.

The quant layer

Black-Scholes and its cousins

Black-Scholes prices a European option from five inputs: spot, strike, time, rate, and volatility. Its assumptions (lognormal returns, constant volatility, continuous trading) are all false, and the skew exists precisely because the market corrects for them. Binomial trees handle American exercise and discrete dividends. Monte Carlo handles path-dependent payoffs. For a retail trader the model's value is not the price (the market gives you that) but the Greeks and the ability to run the scenario lab above.

Probability

Probability ITM ≈ N(d2) for a call, from the same formula. Probability of profit is probability that the stock finishes beyond your breakeven, not beyond your strike. Probability of touch is roughly double the probability of finishing ITM, which is why a 16-delta short strangle that "wins 68% of the time" will be tested far more often than that. All of these are risk-neutral probabilities: they assume the stock drifts at the risk-free rate and returns are lognormal. Real distributions have fat tails, which is why the far wings are systematically underpriced by a naive model and why the market prices them higher.

Expected value

EV = P(win) × avg win − P(loss) × avg loss. A 90% win rate with a 10:1 loss-to-win ratio has negative expectancy. Options make it easy to build high-win-rate, negative-EV strategies; the discipline is in refusing them. Measure expectancy, profit factor, maximum drawdown, and the size of the worst tail loss, not win rate.

Variance risk premium

Across most liquid indices and large caps, implied volatility has exceeded subsequently realized volatility on average. That gap is the compensation sellers earn for bearing crash risk. Harvesting it is a legitimate strategy; forgetting that the premium exists because the tail is real is how those strategies blow up.

Systematic pipeline

This is the layer that connects everything above to a scanner or signal engine. A stock signal is the first step, not the last, and options are the execution layer chosen after the volatility regime is known.

Backtesting the options layer

A testable hypothesis looks like: "When IV Rank > 70 and the stock carries a bullish signal, does a 20–30 delta put credit spread, 30–45 DTE, closed at 50% profit or 21 DTE, produce positive expectancy?" Measure win rate, average winner, average loser, expectancy, profit factor, maximum drawdown, Sharpe, Sortino, tail losses, and, critically, regime dependence: a strategy that only works in 2017-style calm is a strategy for 2017. At that point you are no longer learning options; you are engineering a repeatable edge.

Practical constraints

Liquidity first: trade names with tight bid/ask spreads, real open interest, and daily volume, or the slippage eats the edge. Prefer 30–60 DTE for premium selling (theta is meaningful, gamma is still tame) and 60–120 DTE for directional debit trades (time to be right). Risk per trade of 1–3% of capital for defined-risk structures; size undefined-risk trades by their stress-test loss, not their margin. Cap the portfolio's net vega and net beta-weighted delta at levels you set in advance.

Pre-trade checklist

• Thesis: direction, magnitude, and timing stated in one sentence, with what would prove it wrong.
• Volatility: IV Rank and Percentile checked; IV compared to my realized forecast; event calendar checked for the chosen expiry.
• Structure: chosen from the framework, not from habit. Net delta, gamma, theta, vega written down with their signs.
• Numbers: max profit, max loss, breakevens, probability of profit, expected value computed. Expected move compared to the strikes.
• Liquidity: bid/ask under 5–10% of the mid; open interest present at every strike I'm using.
• Size: loss at the stress scenario is survivable and inside my per-trade and portfolio limits.
• Plan: profit target, loss limit, time exit, and adjustment rule set before entry.
• Assignment: dividend dates and early-assignment exposure known; buying power sufficient if assigned.