PM Margin Worked Examples

Every requirement on this page was computed by the Hypercall margin engine, the same code that checks orders and runs the liquidation watcher, loaded with the production policy (version 3, see PM Risk Parameters). The scenario rows were recomputed independently and agree with the engine to the cent. Only the market inputs are made up, so the arithmetic is easy to follow.

Inputs used throughout

Input	Value
Valuation time	2026-10-07 08:00 UTC
Spot	BTC $100,000, ETH $4,000, AAPL $250
Implied volatility	Flat per underlying: BTC 50%, ETH 60%, AAPL 30% (real marks vary by strike)
"30-day" options	Expire 2026-11-06 (BTC and ETH at 08:00 UTC, AAPL at 20:00 UTC)
"24-hour" options	Expire 2026-10-08 08:00 UTC
Rate	5%, the PM pricing rate
Hyperliquid collateral	Large enough that no perp leg is liquidated in any scenario, except in example 8

One option contract is one unit of the underlying. Scanning risk is the worst weighted scenario P&L, shown as a positive number. IM = max(scanning risk, option floor) + gamma overlay, and MM = 0.85 × IM, because none of these books has open orders.

To get the same breakdown for your own book on live marks, use POST /risk/simulate/portfolio.

Summary

#	Book	Scanning risk	Floor	Overlay	IM	MM	Binding scenario
1	Short 1 BTC 110k call	10,689.13	1,650.00	0	10,689.13	9,085.76	T4
2	Short BTC 105k/110k call spread	1,508.33	75.00	0	1,508.33	1,282.08	10
3a	Short BTC 100k straddle	10,889.29	3,000.00	0	10,889.29	9,255.89	T4
3b	Long BTC 100k straddle	1,706.91	0	0	1,706.91	1,450.87	6
4	BTC iron condor, 24 hours	657.74	15.00	2,000.00	2,657.74	2,259.08	11
5	Long 1 BTC-PERP, short 110k call	13,912.16	1,650.00	0	13,912.16	11,825.34	T1
6	BTC 95k/90k put spread, short 0.10 BTC-PERP	958.39	75.00	0	958.39	814.63	T4
7	BTC call spread, ETH short puts, AAPL long straddle	3,920.20	345.00	0	3,920.20	3,332.17	per underlying
8	Long BTC 100k call, short 0.5 BTC-PERP	depends on HL collateral	0	0	853.45 or 14,469.63		6 or venue liquidation

All amounts in USDC.

1. Naked short call

Book: short 1 BTC 110,000 call, 30 days. Value today $2,379.74.

Scenario	Spot	Vol	P&L	Weight	Weighted
1	+12%	+35%	-7,437.03	1.00	-7,437.03
10	+12%	0%	-5,248.05	1.00	-5,248.05
T2	+25%	+55%	-17,483.16	0.60	-10,489.90
T4	+40%	+70%	-30,540.36	0.35	-10,689.13

$$\text{Option Floor} = 0.015 \times 110{,}000 = \$1{,}650$$ $$\text{IM} = \max(10{,}689.13,\ 1{,}650) = \$10{,}689.13 \qquad \text{MM} = \$9{,}085.76$$

The tail sets the margin: even at 35% weight, a 40% rally with vol up 70% beats every core scenario. The floor does not bind on a 30-day call; it matters for far out-of-the-money shorts whose scenario losses are small.

2. Short call spread

Book: short 1 BTC 105,000 call ($3,838.70), long 1 BTC 110,000 call ($2,379.74), 30 days. $1,458.96 credit, $5,000 wide.

Scenario	Spot	Vol	P&L	Weighted
10	+12%	0%	-1,508.33	-1,508.33
1	+12%	+35%	-1,301.36	-1,301.36
T2	+25%	+55%	-2,131.27	-1,278.76
2	+8%	0%	-1,011.06	-1,011.06

Floor. The calls are not net short (one short, one long), so the floor basis is the spread's largest terminal loss, $5,000, not the $105,000 strike: 1.5% × 5,000 = $75.

$$\text{IM} = \max(1{,}508.33,\ 75) = \$1{,}508.33 \qquad \text{MM} = \$1{,}282.08$$

The spread needs about 30% of its width, against $10,689 for the naked 110k call in example 1. Note that the binding scenario moved from the tail to a core +12% move: the long wing caps what the tail can do.

3. Straddles

3a. Short straddle. Short 1 BTC 100,000 call ($5,907.37) and short 1 BTC 100,000 put ($5,497.53), 30 days.

Scenario	Spot	Vol	P&L	Weight	Weighted
T4	+40%	+70%	-31,112.25	0.35	-10,889.29
T2	+25%	+55%	-17,939.97	0.60	-10,763.98
T3	-40%	+90%	-28,696.03	0.35	-10,043.61
T1	-25%	+70%	-15,714.40	0.60	-9,428.64
1	+12%	+35%	-7,801.24	1.00	-7,801.24

Floor: the call is net short, so the full strike basis applies to both legs, 1.5% × (100,000 + 100,000) = $3,000. IM $10,889.29, MM $9,255.89.

Scanning risk is the worst single direction, not the sum of the two legs. The put loses on the way down, the call on the way up, and no scenario does both.

3b. Long straddle. The same two options bought, $11,404.90 premium paid.

Scenario	Spot	Vol	P&L
6	0%	-15%	-1,706.91
7	-4%	-15%	-1,478.97
3	+4%	-15%	-900.04

IM $1,706.91, MM $1,450.87. A long option is not margin-free under PM: if vol falls 15% with spot unchanged, the straddle loses $1,707 of value, and that loss is charged. The premium itself is paid from cash, not held as margin.

4. Iron condor, 24 hours to expiry

Book: short 98,000 put ($323.86), long 97,000 put ($153.83), short 102,000 call ($345.53), long 103,000 call ($173.49), all expiring in 24 hours. $342.07 credit, $1,000 wide on each side.

Scanning risk is $657.74 (scenario 11, spot -12%), close to the $1,000 width less the credit. Scenario 10, spot +12%, is $0.34 behind.

Floor. Neither side is net short, so the basis is the condor's largest terminal loss, $1,000 (only one side can finish in the money): 1.5% × 1,000 = $15.

Gamma overlay. Both shorts are inside 48 hours. Each type is its own group:

Group	Gross short	Net short	Matched	Largest terminal loss	Overlay
Puts 98k/97k	1	0	1	1,000	min(1% × 100,000 × 1, 1,000) = 1,000
Calls 102k/103k	1	0	1	1,000	min(1% × 100,000 × 1, 1,000) = 1,000

$$\text{IM} = \max(657.74,\ 15) + 2{,}000 = \$2{,}657.74 \qquad \text{MM} = \$2{,}259.08$$

The same condor at 30 days needs $62.95, with no overlay. Crossing the 48-hour line adds $2,000, and IM ends up at about 2.7 times what the condor can lose. That is the cost of the overlay sitting on top of scanning risk.

Under the planned defined-risk cap (V3). This rule is not in production. As designed, on the same book it would charge IM $1,000.00 and MM $1,000.00: the condor's maximum loss, with no 15% maintenance discount on it. See Upcoming changes.

5. Covered call with a perp

Book: long 1 BTC-PERP on Hyperliquid, short 1 BTC 110,000 call, 30 days.

Book	Scanning risk	Binding scenario
Long 1 BTC-PERP alone	15,000.00	T1: $-25{,}000 \times 0.60$
Short 110k call alone (example 1)	10,689.13	T4
Covered call	13,912.16	T1: $-23{,}186.93 \times 0.60$

Floor $1,650 (the call, as in example 1). IM $13,912.16, MM $11,825.34.

The covered call needs more than the naked call. The short call caps the perp's upside, but the risk of a covered call is the downside, and on the way down the call only gives back its $2,380 of value. PM charges the long perp's selloff, softened slightly by the call. If you want a covered call to margin like a spread, buy a put.

6. Put spread hedged with a perp

Book: short 1 BTC 95,000 put ($3,272.66, delta -0.32), long 1 BTC 90,000 put ($1,734.47, delta -0.20), 30 days. The spread is long about 0.12 delta. Add a short BTC-PERP against it:

Short perp	Scanning risk	Binding scenario	IM	MM
0	1,703.19	11 (-12%, 0 vol)	1,703.19	1,447.71
0.05	1,103.19	11 (-12%, 0 vol)	1,103.19	937.71
0.10	958.39	T4 (+40%, +70% vol)	958.39	814.63
0.15	1,658.39	T4 (+40%, +70% vol)	1,658.39	1,409.63

Floor $75 in every row: perps do not change the option floor.

Each 0.05 of perp takes $600 off the -12% loss. Past about 0.10 the perp's own loss in the +40% tail becomes the worst case, and IM rises again. The best hedge balances the two sides of the grid; it is not the full delta.

The same logic works harder on a naked short. Short 1 BTC 95,000 put alone needs $11,107.61; with 0.33 BTC-PERP short against it, $6,487.61.

7. A book across three underlyings

Book:

• BTC: the call spread from example 2
• ETH: short 5 ETH 3,600 puts, 30 days ($103.24 each)
• AAPL: long 100 AAPL 250 calls ($9.16) and long 100 AAPL 250 puts ($8.11), 30 days

Each underlying is charged its own worst scenario, from its own grid:

Underlying	Scanning risk	Binding scenario	Floor
BTC	1,508.33	10 (+12%, 0 vol)	75.00
ETH	1,981.45	T3 (-40%, +90% vol, weight 0.35)	270.00
AAPL	430.43	equity-vol-down-25 (0%, -25% vol)	0
Account	3,920.20		345.00

$$\text{IM} = \max(3{,}920.20,\ 345) = \$3{,}920.20 \qquad \text{MM} = \$3{,}332.17$$

Two things to notice:

• AAPL is set by one of its extra scenarios. On the 17-scenario grid alone, the long straddle's worst case would be scenario 6 (vol -15%), $258.28. The AAPL override adds a -25% vol shock, which costs $430.43.
• No netting across underlyings. If all three were moved together in the same scenario and netted, the worst shared scenario would be scenario 3 (+4%, vol -15%) at a loss of $140.70. PM charges $3,920.20, because BTC, ETH and AAPL do not have to move together.

8. Long call hedged with a short perp, and the Hyperliquid check

The engine harness used for examples 1 to 7 gives every book unlimited Hyperliquid collateral, so this one is computed by hand from the same rules (see Hyperliquid liquidation check).

Book: long 1 BTC 100,000 call, 30 days ($5,907.37, delta 0.54), short 0.5 BTC-PERP on Hyperliquid in cross margin.

Scenario	Spot	Vol	Call P&L	Perp P&L	Total
3	+4%	-15%	1,549.98	-2,000.00	-450.02
6	0%	-15%	-853.45	0.00	-853.45
7	-4%	-15%	-2,739.48	2,000.00	-739.48
11	-12%	0%	-4,469.63	6,000.00	1,530.37

The call alone would need $4,469.63 (scenario 11) and the perp alone $7,500 (T2). Together, what is left is vega: $853.45, if the perp survives.

Does the perp survive on Hyperliquid? Assume BTC's Hyperliquid maintenance is a single tier at 40x max leverage, so 1.25%. Check Hyperliquid's current tiers for real numbers. The worst move for a short perp is T4, spot +40% to $140,000:

$$\text{P\&L} = -0.5 \times 40{,}000 = -20{,}000 \qquad \text{Maintenance} = 0.5 \times 140{,}000 \times 0.0125 = 875$$

The cross account is liquidated in that move if its equity $E$ satisfies $E - 20{,}000 \le 875$, so for any equity of $20,875 or less. Tail weights do not apply to this test.

• $25,000 on Hyperliquid. The perp survives. IM $853.45, MM $725.43.
• $10,000 on Hyperliquid. The perp is liquidated in the T4 move. Hypercall assumes the whole $10,000 is lost and the call is left on its own:

$$\text{Charge} = \max(\underbrace{853.45}_{\text{hedge intact}},\ \underbrace{4{,}469.63}_{\text{call alone}} + \underbrace{10{,}000}_{\text{lost collateral}}) = \$14{,}469.63$$

IM $14,469.63, MM $12,299.19. Raising the Hyperliquid balance above $20,875, or cutting the perp, brings IM back to $853.45. This is the rule the perp order form describes as the worse of PM and Hyperliquid margin. See Perp Leverage and Margin Modes.

What these examples leave out

• Open orders. Each example is executed positions only. Resting orders add their own reservation; see Open orders.
• Real volatility surfaces. Production uses each option's own mark IV. A skewed surface changes the numbers, not the method.
• Equity. These are requirements. Whether you can trade depends on equity against IM, which includes your HyperCore account value, Hypercall cash, option marks and any PM loans.

Related

• Portfolio Margin
• PM Risk Parameters
• Margin Floor
• Simulating PM Risk