Part XII · Strategy atlas · Chapter 12E

Expert strategies

Multi-leg structures for specialists

Where we are

Chapter 12D left one side of each position open. The expert tier is about equivalence instead: synthetic positions that rebuild stock or an option from other legs, and the box spread, a loan made out of options. Their Greeks mostly cancel, so what decides them is financing, put-call parity and the cost of trading four legs.

What you need first

Put-call parity and the synthetic positions it implies (Chapter 5F).
Present value and compounding (Chapter 1A).
Calendars, diagonals, straddles and strangles (Chapter 12C).
Early exercise of American options (Chapters 5F and 12H), and the Section 1256 tax treatment (Chapter 2E).

1 · How to read this tier

The cards and the notation are those of Chapter 12C: strikes run from lowest to highest, $K_1 < K_2$ , with a net credit $C$ or a net debit $D$ , and $S_T$ is the price at expiry. Parity appears often, so it is written here in words, with $S$ the underlying's price, $r$ the risk-free rate and $T$ the time to expiry in years:

\text{call} - \text{put} = S - K e^{-rT}

(12E.1)

Here the call and the put share the strike $K$ and the expiry, and $K e^{-rT}$ is the strike's present value (Chapter 5F). The formula assumes European options and no dividends.

2 · Synthetic structures

Long and short synthetic future

Synthetic future	At a glance
Legs	Long: buy 1 call and sell 1 put at the same strike $K$ . Short: sell the call and buy the put
P&L at expiry	Long: $S_T - K - D$ , where $D$ is the call's price less the put's
Maximum profit and loss	Those of the underlying: unbounded above; for the long, $K + D$ lost if the underlying goes to zero
Breakeven	$K + D$
Greeks at entry	Delta about ±1.0, gamma about 0, theta about 0, vega about 0: the option Greeks cancel by put-call parity

A synthetic future moves one for one with the underlying. By (12E.1), its net premium is $S - K e^{-rT}$ . With the strike at the price, that is just the interest on the strike: the synthetic holds the underlying without paying for it, and the premium is the financing cost. With SPX and the strike at 6,000, a 4% rate and three months to expiry, that is $6{,}000 \times (1 - e^{-0.01})$ , about 59.7 points.

When borrowing the underlying is hard or expensive, as in a short squeeze on a single stock, the synthetic can be cheaper than the literal stock. Chapter 5F works the arithmetic for a borrow fee.

Synthetic put

A synthetic put shorts 100 shares and buys 1 call at the same strike. By parity it replicates a long put, and traders use it when listed puts are illiquid or expensive. With the shares sold at $S_0$ and the call bought for $P$ , the P&L at expiry is $(S_0 - S_T) + \max(S_T - K, 0) - P$ per share. The most it can lose is $(K - S_0) + P$ , if the underlying finishes at or above $K$ , and it breaks even at $S_0 - P$ . Short shares also carry a borrow cost (Chapter 5F).

Long and short combo

A combo is the synthetic long or short stock by another name. It is one half of a put-call parity arbitrage: the conversion and the reversal below pair it with the opposite position in the stock.

Strip and strap

A strip buys 1 call and 2 puts at the same strike $K$ , which gives it a bearish bias. A strap buys 2 calls and 1 put, a bullish bias. Both are asymmetric straddle ratios: long convexity in both directions, with more weight on one side. For a debit $D$ , each loses at most $D$ , at $S_T = K$ .

Strip: the P&L falls with slope −2 below $K$ and rises with slope +1 above it. Its breakevens are $K - D/2$ and $K + D$ .
Strap: slope −1 below $K$ and +2 above. Its breakevens are $K - D$ and $K + D/2$ .

Priced with Chapter 12C's $40 call and $40 put at 6,000, either costs 120 points. The strip breaks even at 5,940 and 6,120, and the strap at 5,880 and 6,060.

Long and short guts

A long guts buys an in-the-money (ITM) call at $K_1$ and an ITM put at $K_2$ , with $K_1 < S < K_2$ . It has the volatility exposure of a straddle, or more exactly of the strangle at its two strikes, but is more capital-intensive. Between the strikes both options are in the money and together pay $K_2 - K_1$ , so for a debit $D$ the most it can lose is $D - (K_2 - K_1)$ . It breaks even at $K_2 - D$ and $K_1 + D$ .

By parity, a long guts is a long strangle at the same two strikes plus a bond that pays $K_2 - K_1$ at expiry. It has the strangle's Greeks and ties up the extra capital in the bond. The short guts is the mirror image.

Double diagonal

A double diagonal combines two diagonal calendars, one on the call side and one on the put side, at different strikes. It earns income from the near-dated short options and holds forward volatility through the far-dated long ones. Like every calendar, it has no closed-form breakeven (Chapter 12C). Suppose the long options sit further out than the short ones. The most it can then lose at the near expiry is about the net debit plus the larger gap between a short strike and the long strike behind it.

3 · Box spread

Box spread	At a glance
Legs	Buy a call at $K_1$ , sell a call at $K_2$ , sell a put at $K_1$ , buy a put at $K_2$ , all on one underlying and one expiry
Payoff at expiry	$K_2 - K_1$ , the width, wherever the underlying settles
Greeks at entry	Delta about 0, gamma about 0, theta about 0, vega about 0: by put-call parity the box is deterministic, and every Greek cancels

The legs pair into two synthetics: a synthetic long at $K_1$ and a synthetic short at $K_2$ . Their slopes cancel at every price, so the box spread pays $(S_T - K_1) - (S_T - K_2) = K_2 - K_1$ whatever happens.

The box spread is a synthetic loan. If you sell the box, you collect the net credit up front and owe the width at expiry: you have borrowed money at an implied interest rate. If you buy the box, you pay the net debit up front and receive the width at expiry, which lends money at the same kind of rate. With a credit $C$ received, a width $W$ owed, both in dollars, and $n$ days to expiry, the simple annual rate $r$ solves $C(1 + rn/365) = W$ :

r = \frac{W - C}{C} \times \frac{365}{n}

(12E.2)

For SPX, boxes receive the tax-advantaged Section 1256 60/40 treatment, which makes the box the cleanest available retail borrowing or lending instrument. For borrowing, the rate is often within 5 to 25 basis points (bps) of broker call rates. For lending, it is within a few bps of Treasury bill (T-bill) rates.

Sold SPX boxes give a retail account the institutional alternative to a broker margin loan for accessing leverage. Their implied rate is typically materially below the margin rates retail brokers charge, where the broker's spread over the federal funds rate is often 100 to 400 bps.

Worked example: A box on SPX

Step 1: the box. With SPX at 6,000 and 90 DTE, construct a box at 6,000 and 6,100. The width is 100 points, or $W = \$10{,}000$ per box.

Step 2: the loan. Suppose selling the box brings in a net credit of $C = \$9{,}940$ . You receive $9,940 today and owe $10,000 in 90 days.

Step 3: the rate. By (12E.2), $r = (60/9{,}940) \times (365/90)$ , which is about 2.45% as an annual percentage rate (APR).

Step 4: against the T-bill. If T-bills yield 4.2%, that is a very cheap loan. By the lending rates above, though, a box priced this far below the T-bill rate would not normally find a lender. At 4.2% the same box would bring in about $9,898, for about $102 of interest, so read the $9,940 as a number chosen to show the arithmetic.

Liquidity in box spreads on the listed exchanges varies widely. You generally need to trade them on tastytrade, Interactive Brokers, or a specialty broker that quotes box-spread liquidity. Some brokers do not execute boxes at all.

The box can fail in three ways:

Early assignment on the short legs. SPX boxes are European-style and safe from it. SPY boxes are American-style and carry early-assignment risk on the short put: a deep in-the-money put is exercised early when the interest on its strike exceeds its remaining time value (Chapter 12H). Stick to European-style index options for boxes.
Counterparty and clearing-firm risk. In a cautionary tale from 2019, a Robinhood user sold boxes over and over for what was described as "infinite leverage", because the broker's system did not margin the future obligation correctly. The leverage was real. When the market moved, the user was wiped out by margin calls on positions that had been booked as a riskless arbitrage. The box itself is riskless at expiry, but the path to expiry can produce margin calls if the broker does not compute the offsetting risk correctly.
Friction. Round-trip execution costs on a four-leg structure are typically 1 to 4 cents for each dollar of width, or $100 to $400 on a $10,000 box. That is meaningful against the example's $60 of implied interest.

The honest framing is that boxes are an interest-rate instrument that happens to be built out of options. The Greeks are not the interesting part: the implied rate is.

Drag the days to expiry from a month to a year. At the example's 2.45%, $100 of costs makes a 30-day box cost 14.8% a year, and a one-year box 3.5%.

rate before costs 2.45%

days to expiry 90

round-trip cost $100

credit C = $9,940 ($60 interest)

rate after costs = 6.60%

beats 5.2% margin from 137 days

beats 8.2% margin from 65 days

A loan that pays for its costs only with time. Selling a box 100 SPX points wide borrows its credit C today against the $10,000 owed in n days, at the rate r = (W − C)/C × 365/n. Set the rate before costs and the term, and the credit follows; the round-trip cost, which the chapter puts at $100 to $400 on this box, comes off the credit. The curve is the rate after costs at every term for the chosen rate before costs (dashed), against the example's 4.2% T-bill and a band of margin loans 100 to 400 bps above it, with the T-bill standing in for the federal funds rate. The figure opens on the worked example: 2.45% for $9,940 over 90 days, below the T-bill, where 4.2% would bring in about $9,898.

4 · Conversion and reversal

A conversion buys the stock, sells a call and buys a put at the same strike: actual long stock plus a synthetic short. The position is delta-neutral and locks in a small profit if the parity gap is mispriced in the right direction. A reversal is the mirror image, short stock with a long call and a short put, used when parity is mispriced the other way.

A conversion costs $S - \text{call} + \text{put}$ today and is worth exactly $K$ at expiry. It profits when its cost is below the strike's present value, which by (12E.1) happens when the call is rich against the put:

\text{call} - \text{put} > S - K e^{-rT}

(12E.3)

Chapter 5F's example is one: with the stock and the strike at 100, a 4% rate and three months, parity asks for a gap of 1.00. A call at $4.50 and a put at $3.00 give a gap of 1.50. The conversion costs $100 - 4.50 + 3.00 = 98.50$ against a present value of 99.00, a riskless $0.50.

Both are professional arbitrage trades, and retail traders almost never see economic versions of them, because bid-ask costs typically exceed the arbitrage spread. The point for a student is that put-call parity is enforced by the existence of these structures. Anyone who quotes prices that violate parity offers an immediate arbitrage to professional desks, and the prices snap back within milliseconds.

5 · Cards are starting points, not signals

A payoff geometry is necessary but not sufficient. Every strategy in Chapters 12B to 12E also requires five things:

A market thesis (Chapter 11B).
A regime check against the dashboard and the composite scorecard (Chapters 11A and 11G).
A conviction score (Chapter 13B).
An explicit sizing rule (Chapter 13C).
A kill condition committed in advance (Chapter 13C).

Choosing a structure because its payoff diagram looks attractive is the most common retail failure mode. The structure is the vehicle for expressing a forecast, not the forecast itself. The work that produces edge is in steps 1, 2 and 3 above, not in selecting an exotic combination of legs.

In practice

Use a synthetic when it is cheaper than the direct position, which usually means an expensive borrow or an illiquid put. Before borrowing with a box, compute its rate by (12E.2) after costs, and compare it with the T-bill rate and with your broker's margin rate. Trade boxes only on European-style index options such as SPX, at a broker that quotes them and margins them correctly. Hold them long enough for the interest to outweigh the friction.

When it fails

Equivalence holds at expiry and in theory. Before expiry, three things can break it: early assignment on an American-style leg, a margin call from a broker that miscomputes a box, and trading costs on four legs that exceed the edge they capture. An arbitrage that looks riskless on paper, like the 2019 boxes, can still end an account along the way.

Decision rule

Borrow with a sold SPX box only when its after-cost rate beats your margin rate over the period you plan to hold it, and only at a broker that handles boxes.
Never build a box from American-style options such as SPY.
Leave conversions and reversals to professional desks. Use parity to check prices, not as a strategy.
Choose any expert structure only after the thesis, regime check, conviction score, size and kill condition exist.

Common mistake: treating a box as free leverage

A box is a loan, and a loan has to be repaid. Borrowing through boxes adds leverage exactly as a margin loan does (Chapter 1E). The Robinhood user of 2019 found that a riskless structure on paper can still bring margin calls when the broker books it wrongly.

What to carry forward

A synthetic future, a call bought and a put sold at one strike, moves one for one with the underlying. Struck at the price, its net premium is the interest on the strike.
Strips, straps and guts are straddle variants: weighted to one side, or rebuilt from in-the-money options at a higher capital cost.
A box pays its width whatever happens, so selling one borrows at the rate $(W - C)/C \times 365/n$ . SPX boxes are European and taxed under Section 1256.
Conversions and reversals enforce put-call parity, and their spread rarely survives retail costs.
A payoff diagram is not a thesis: the thesis, regime, conviction, size and kill rule come first.

6 · Check yourself

Question 1

SPX is at 5,000, a 5,000 call costs 110 and the 5,000 put 60, with three months to expiry and a 4% rate. Does parity hold, and what does the synthetic long cost?

Answer

Parity asks for $5{,}000 \times (1 - e^{-0.01})$ , about 49.75 points. The quotes give $110 - 60 = 50$ , within a quarter of a point, so parity holds to that tolerance. The synthetic long costs 50 points, the interest on the strike for three months.

Question 2

You sell a 1,000-point-wide SPX box, 180 days to expiry, for $97,500. What rate are you borrowing at?

Answer

The width is $1{,}000 \times \$100$ , which is $100,000. By (12E.2), $r = (2{,}500/97{,}500) \times (365/180)$ , about 5.2% a year.

Question 3

Why is an SPX box safe from early assignment, while a SPY box is not?

Answer

SPX options are European-style and can be exercised only at expiry. SPY options are American-style, so a deep in-the-money short put in a SPY box can be exercised early once the interest on its strike exceeds its time value (Chapter 12H).

Question 4

A strip costs 90 points at the 6,000 strike. Where does it break even?

Answer

It breaks even at $6{,}000 - 90/2 = 5{,}955$ below and at $6{,}000 + 90 = 6{,}090$ above, with its maximum loss of 90 points at 6,000.