Systems and process

OpeningOne view, three outcomes

You think XYZ (stock at $200) will rise to around 215 within a month. It reports earnings in 10 days, and implied volatility has been bid up to 55% ahead of the report. Three traders, A, B and C, see it exactly the way you do, and each places a trade.

The stock then does just what you all expected: 203 on day 7, a jump to 206 after earnings on day 10 with IV falling back to 32%, and 212 by day 30, up 6% for the month. Here is how the three traders did.

A bought the cheapest option. It expired before earnings, which is why its IV was only 40%. But the move took until day 30 to play out: when the option expired on day 7 the stock was only at 203, and the option went to zero. B gave the trade enough time but bought a single call at 55% IV. After earnings IV fell 23 points, and the Vega loss ate the gain from the 6% rise. C used a spread: the short 215 call offset most of the Vega, and the short leg sat right on the target price. The same move made 57%.

All three made exactly the same call on direction, yet the results ran from −100% to +57%. The difference came from two things other than direction: volatility and time. This chapter turns C's thinking into a fixed process you can follow every time.

14.1The nine-step process

The process makes every trade answer the same nine questions. Only when the questions are the same can you compare results across trades and trace a mistake to a specific step.

Options trading is an environment of incomplete answers: a correct decision can lose money, and a wrong one can make money. From the P&L of a single trade you can't tell good judgment from good luck. The process breaks a trade into nine stages, each with a clear output, so in review you can ask “which step went wrong?” instead of just saying “this one lost.”

Beginners spend most of their time on steps 1 and 8: finding opportunities, watching the screen, deciding when to sell. Long-run results depend more on steps 3, 5 and 7. The structure decides in which scenarios you make money, the size decides whether you survive the worst scenario, and the management rules decide whether losses get out of hand. All three are done before the order goes in and take almost no screen time.

14.2A view in three dimensions

An options view needs three sentences: direction, volatility and time. Write only the first, and you leave the other two variables to luck.

Buying stock takes only a direction. An option's price depends on the underlying price, implied volatility and time remaining all at once, which correspond to three Greeks: Delta, Vega and Theta. If you don't take a position on the last two, the market takes one for you: buying a single option automatically bets that “IV won't fall and the move will come fast enough.” That is how A and B lost in the opening.

How to write the three sentences

• Direction sentence: “The underlying reaches ___ (target price) within N days; a break below ___ means I'm wrong (invalidation price).” A neutral view is also a direction sentence: “Stays between 95 and 105 for 30 days.”
• Volatility sentence: “IV is ___% now, IV Rank ___, 20-day realized volatility ___%. I think IV will rise, fall or stay flat from here, because ___.”
• Time sentence: “The catalyst is ___. It arrives on day ___ at the earliest and day ___ at the latest.”

Write the direction sentence so it can be compared with the market

Only when you write down a target price can you compare it with what the market is quoting. With the stock at $100, IV at 30% and 30 days to go, the market's 1-standard-deviation expected move is:

Your target is 105 in 30 days, or 0.58 standard deviations. Working back from option prices, the market puts the probability of finishing above 105 at 28.4% (the N(d₂) of the 105 call). If you think the probability is closer to one half, you disagree with the market, and buying calls makes sense. If you think it's only about 30%, your “bullish” view is no different from the market's pricing, and that isn't a trade.

The volatility sentence decides whether you buy or sell

The same at-the-money straddle is worth 6.85 at 30% IV and only 4.57 at 20% IV. If you have reason to think realized volatility over the next 30 days will be only 20%, then buying options priced at 30% means paying 6.85 for something worth 4.57. Even if you get the direction right, you are running into a headwind. In that case, favor selling structures or spreads, so Vega isn't working against you.

Three common failures, and the sentence each one left out

14.3The strategy selection matrix

Pick one value for each of three dimensions—direction, IV environment and whether there is an event—and the structure is mostly settled. The matrix rules out structures with the wrong shape before you place the order.

How to set the three dimensions

• Direction: bullish, bearish or neutral, taken from the direction sentence.
• IV environment: split IV Rank into three tiers: below 25 is low, 25 to 50 is mid, above 50 is high. IV Rank = (current IV − 52-week low IV) ÷ (52-week high − 52-week low) × 100. Then cross-check with IV ÷ 20-day realized volatility (IV/HV): a high IV Rank with IV/HV below 1 means volatility itself has picked up, and options aren't actually expensive.
• Event: whether there is an earnings report, FOMC meeting, CPI release, product launch or ex-dividend date during the period you plan to hold.

Three rules behind the matrix

1. IV sets the sign of Vega. Use buying structures when IV is low (net Vega positive), selling structures when IV is high (net Vega negative) and spreads when IV is in the middle (net Vega near zero), so that direction becomes the main source of profit and loss.
2. An event means risk must be capped. IV gets bid up before an event and collapses after it. “Buy before the event, sell before the event” and “sell before the event, buy back after it” are two completely different trades. If you hold through an event, the structure must have a known maximum loss.
3. Neutral views are the most sensitive to IV. A neutral structure earns nothing from Delta; it can only make money from Theta or Vega. That is why neutral views are hardest when IV is low: sellers collect too little, and the most common right answer is not to trade.

14.4Strikes, expiration and size

Expiration follows the time sentence, strikes follow the direction sentence, and the number of contracts follows the worst case.

Where “expiration ≥ 1.5×” comes from

This rule of thumb rests on the “cliff” at expiration: if the catalyst comes even one day after expiration, the option goes to zero. Your sense of timing is usually less accurate than your sense of direction, so leave a margin for time. The chart below draws the trade-off: the stock sits at 100 until the catalyst arrives, then rises in a single step.

When the catalyst arrives on time, short-dated options have the most leverage; when it arrives late, they go straight to zero. Longer-dated options make a bit less when the catalyst is on time and still hold value when it is late. If you say “about a month,” it could really be three weeks or six, so choose an expiration of roughly 1.5 times six weeks. Events with fixed dates are the exception: once the earnings date has been announced there is little timing uncertainty, and the first expiration after the event will do. What you have to worry about then is IV crush.

Where “exit with 21 days left” comes from

At 21 days left, each day's decay as a share of the price is more than twice what it was at 45 days, and it accelerates even faster after that. For buyers, this is the point where the cost of holding on gets steep. For sellers, little premium is left while Gamma keeps growing, so the same price move causes a bigger swing in Delta and the risk/reward gets worse. The rule doesn't apply to deep ITM options or to LEAPS with more than a year left; their Theta is small to begin with.

The sellers' “short leg at Delta 0.15–0.30” can be read with the probability interpretation from Chapter 4: a put with a 0.20 Delta has about a 20% chance of finishing in the money, but close to a 40% chance that the stock touches the strike at some point before expiration. So design management rules around a “touch,” not just around expiration.

Size: work back from the worst case to the number of contracts

For example, with an $80,000 account, a 2% per-trade cap is $1,600. If one spread's maximum loss is $489, you can trade ⌊1600 ÷ 489⌋ = 3 spreads. Note that the number of contracts is the last thing you calculate. Don't decide on “10 contracts” first and then go looking for a structure.

14.5Executing the entry

Execution won't turn a bad trade into a good one, but it can leak away a good trade's edge bit by bit. Spread orders, limit prices, timing and scaling in are four ways to plug the leaks.

The example below uses a 30-day 100/105 call spread. With the stock at $100 and IV at 30%, the two legs have theoretical values of 3.59 and 1.66. Suppose the quotes are as follows:

1. Use a combo order, not separate legs

Placing the two legs as separate orders is called “legging in.” Suppose you first buy the 100 call at 3.70, and before you can sell the 105 call the stock drops $0.50: the 105 call's mid falls from 1.66 to 1.51, and you can only sell at about 1.45. Your total cost becomes 3.70 − 1.45 = 2.25, 0.31 more than the combo order's mid, or $31 extra per spread. The stock could just as well move in your favor, but that amounts to placing an extra directional bet with no edge at entry. A combo order also guarantees that both legs fill together, never just one.

2. Start the limit at the mid

The natural price of 2.10 is the “fill right now” price; the mid of 1.94 is the average of the two sides. Start your order at the mid and give up $0.01–0.05 every minute or two; you will usually fill near the mid. The difference looks small: 0.16 per spread, $160 for 10 spreads. But this spread's maximum profit is only 5 − 1.94 = 3.06. Filling at the natural price both in and out gives up 0.32 on the round trip, or 10.5% of the maximum profit. Selling strategies make less per trade, so for them the percentage is even higher.

The bid–ask spread as a percentage of the mid is a direct measure of liquidity. This course's threshold: no more than 10% near the money, and within 5% is better. On contracts above 15%, almost no strategy can cover the cost of execution.

3. Avoid illiquid times of day

• The first 15–30 minutes after the open: market makers are still adjusting quotes for overnight news, spreads are wide, and IV often runs high. Options bought in this window tend to be overpriced.
• The last few minutes before the close: quotes widen, especially on contracts close to expiration.
• Around data releases: just before and after CPI, FOMC statements and similar events, quotes can briefly disappear or widen sharply.
• Outside regular trading hours: single-stock options trade only during regular hours. Some index options have extended sessions, but liquidity there is clearly worse.

4. Planned scaling in

Splitting a position into 2–3 tranches that enter on time or price conditions reduces timing error. For example, you plan 3 spreads: 2 today, and 1 more if the underlying pulls back to a support level within 5 trading days without breaking the invalidation price. The scaling conditions must be written down before the first tranche. Deciding after a loss to “buy a little more to bring down the cost” is not scaling in; it is the averaging down to rescue a losing trade that Chapter 15 covers.

14.6Write management rules before you enter

Management rules must be written down before you enter. Once you hold the position, an open profit makes you want to lock it in early and an open loss makes you want to wait a little longer. The position has already changed your judgment.

Write four rules for every trade: take-profit, stop-loss, time and adjustment. The table below gives starting points by structure; adjust the numbers based on your own journal statistics.

What these rules actually change

Let's test three common credit-spread rules with a simulation. Sell a 45-day 92.5/87.5 put spread: stock at $100, IV 30%, short-leg Delta −0.20, credit 0.81, maximum loss 4.19. Assume realized volatility of 25% (below implied, so the seller has an edge), reprice at IV 30% at each day's close, and simulate 20,000 paths.

• A take-profit on its own lowers expectancy. It cuts the winners short and leaves the losers as they were; expectancy per 30 days held falls from 11.4 to 6.0.
• Take-profit plus stop-loss: expectancy about the same, tail cut in half. Expectancy per 30 days is 11.2, almost the same as holding to expiration; the average loss in the worst 5% drops from −419 to −214, and the standard deviation of single-trade results drops from 149 to 89.
• Rules mainly change the shape of the distribution. With a smaller tail, the same risk budget can carry more trades, and the account is less likely to be badly hurt by a run of bad luck.

14.7The options trading journal

An ordinary trading journal records buys, sells and profit and loss. An options journal also has to record the volatility, Greeks, expected move and events at entry. Otherwise, when you review, you can't tell whether you made money from direction, from volatility or from time.

Six fields an ordinary journal doesn't have

1. IV, IV Rank, IV/HV: shows whether you bought when options were expensive or sold when they were cheap. After three months, grouping results by IV environment often reveals one environment in which you keep losing.
2. Greeks snapshot: the whole position's Delta (shares), Gamma, Theta ($/day) and Vega ($/IV point). Record it at entry, at exit and after each event.
3. Expected move: one standard deviation to expiration and to the event. Compare it with the actual move at exit, and you'll know whether you made money because “I was right” or because “the move was bigger than the market expected.”
4. Event: type and date. Keep separate statistics for event trades.
5. Reason for the structure: why this one, and why not the other two. A few months later, this column is often the most valuable.
6. R-multiple and attribution: P&L ÷ planned maximum loss, so trades of different sizes can be compared. For attribution, pick exactly one of four categories: judgment error, execution error, risk-management error, or normal variance.

OPTIONS TRADING JOURNAL  No.: ____  Date: ____  Account equity: ____
Underlying: ____  Underlying price: ____

[View: three sentences]
Direction: reaches ____ (target price) within ____ days; a break below ____ means I'm wrong (invalidation price)
Volatility: IV ____%  IV Rank ____  IV/HV20 ____  My view: ____
Time: catalyst ____, earliest day ____, latest day ____

[What the market is pricing]
Expected move (1σ): to expiration ±____  to the event ±____
Event: ____ (earnings / FOMC / CPI / ex-dividend)  Date ____

[Structure]
Structure: ____  Legs: ____ (strike / expiration / contracts / buy or sell)
Why this one: ____
Why not the other two: ____
Greeks snapshot (whole position): Δ ____ shares  Γ ____  Θ $____/day  Vega $____/IV point
Max profit ____  Max loss ____  % of account ____%  Breakeven ____

[Entry]
Mid ____  Fill ____  Slippage ____  Time of day ____  Scaling conditions ____

[Management rules: written before entry]
Take-profit: ____
Stop-loss: ____ (as an underlying price)
Time: ____
Adjustment: ____ (at most ____ times)

[Exit: fill in afterward]
Date ____  Fill ____  P&L ____  R-multiple ____  Held ____ days
Exit snapshot: Δ ____  IV ____  Actual move in underlying ____ (vs. expected move)
Trigger: take-profit / stop-loss / time / outside the rules
Rule broken: yes / no  Which one: ____
Attribution: judgment error / execution error / risk-management error / normal variance
If I did it again: ____

Spreadsheet header row:
Date,Underlying,Structure,DTE,Contracts,IV,IVRank,IV/HV,Expected move,Delta,Gamma,Theta,Vega,Event,Max loss,% of account,Fill,Exit price,P&L,R-multiple,Days held,Trigger,Violation,Attribution

Be especially careful to log the trades that “made money but were done wrong”: you broke a rule, went over your size limit or chased an entry at a time you shouldn't have, and it happened to pay off. Losses force you to reflect; profits reward the wrong behavior. The more of these trades your journal holds, the more your profits depend on luck.

14.8Thinking in expectancy

Whether a strategy is good depends on its expectancy over dozens of trades. That number can only come from your journal, not from your gut.

The breakeven win rate turns “is the win rate high enough?” into a question with a benchmark. A credit spread that makes $100 and loses $300 needs a win rate above 75% just to break even; an OTM buying strategy that makes $300 and loses $100 only needs 25%. Comparing these two strategies on win rate alone is meaningless.

Expectancy has error too

The expectancy you compute from a journal is a sample estimate: the smaller the sample, the larger the error. Call the standard deviation of individual results σ. The standard error of the average of N trades is σ ÷ √N, and the 95% interval is roughly expectancy ± 1.96 standard errors. If you only know the win rate and the average win and loss, you can approximate with a two-point distribution: σ = (avg win + avg loss) × √(p(1 − p)).

Example: win rate 80%, average win $120, average loss $300, 30 trades. Expectancy is 0.8 × 120 − 0.2 × 300 = $36, σ = 420 × 0.4 = $168, the standard error is 168 ÷ √30 = $30.7, and the 95% interval is [−$24, +$96], which includes zero. For the lower bound of the interval to turn positive you need about (1.96 × 168 ÷ 36)² ≈ 84 trades. The win rate itself has an interval of 80% ± 14.3%, or 65.7% to 94.3%, and the breakeven win rate of 71.4% falls inside it.

Four easy statistical mistakes

• The tail hasn't shown up yet. If a strategy has a 5% chance per trade of hitting its maximum loss, the probability of not hitting it once in 28 trades is 0.95²⁸ = 23.8%. A loss you don't see in the sample still exists.
• Mixing different structures in one set of statistics. Credit spreads and long single options have completely different payoff shapes, and an expectancy averaged across them tells you nothing. At least group by structure, and preferably by IV environment as well.
• Using dollars instead of R-multiples. As your size grows, dollar statistics get dominated by the last few big trades. Convert to R-multiples (P&L ÷ planned maximum loss) and recompute.
• Ignoring holding time. An expectancy of $20 over 40 days versus $15 over 10 days: the second uses capital far more efficiently per day (see Table 14-8).

14.9The weekly scan

Spend an hour each week running your underlyings through filters for liquidity, IV and the event calendar. This does more than watching the screen every day for opportunities to make sure you only trade where the matrix has an answer.

1. Fixed candidate pool: 20–40 stocks and ETFs you know well, with liquid options. If an underlying isn't in the pool, you don't trade it this week.
2. Liquidity filter: ATM bid–ask spread no more than 10% of the mid; open interest of at least 500–1,000 contracts at the main strikes.
3. Event calendar: mark earnings, FOMC, CPI and ex-dividend dates over the next two to six weeks.
4. IV metrics: IV Rank and IV/HV20. Besides IV Rank, look at IV percentile: the share of days in the past year with IV below today's. One extreme episode pushes the 52-week high up and keeps IV Rank low for a long time; the percentile is much less affected by it.
5. Sort into buckets: seller candidates, buyer candidates, event candidates, skip.
6. Write the three sentences: write a view for each candidate; if you can't, move it to “skip.”

• AAA: IV Rank 58.8 and IV/HV 1.33, so the options really are expensive. Earnings are 26 days out; choosing contracts that expire within 21 days avoids them, so use a selling structure.
• BBB: IV Rank is only 15.2 ahead of earnings, which looks cheap. But the 52-week high of 70% probably came from a single event last year, which drags IV Rank down. First compare this report's expected move with the actual moves on the last few reports, then decide whether to buy or sell.
• FFF: the IV Rank of 87 is propped up by earnings and will very likely collapse in 3 days. It suits an event calendar spread or a small event iron condor, not buying a single option.
• DDD: IV is low and earnings are far off. If you have a directional view, it suits buying options or a debit spread; with no view, leave it alone.

The output of the scan is a short list of no more than three to five underlyings. For the rest of the week you do only two things: manage existing positions, and wait for names on the short list to meet their entry conditions.

CasesTwo cases

ToolsChecklist and self-test

Your checkmarks are saved only in your own browser.

Self-test: five questions