Backspread
A put backspread sells one put and buys two at a lower strike on one expiry. Every written option is covered, so the loss is limited: at most the gap between the strikes plus the net cost, lost at the lower strike itself. What it gives up is the middle: a modest fall into the zone between the strikes is its worst outcome. In return it needs no uncovered permission, its margin does not move, and below the lower breakeven it gains £10 a point on the FTSE 100 contract, with rising volatility helping on the way. It is designed for a large, fast fall, bought cheaply because the written put pays for most of the two bought ones.
The worked trade is the ratio spread page's trade with every side reversed: the same strikes, the same expiry, the same fills. Figures that are simply that page's with the signs changed are quoted from it rather than worked again; this page covers what is different. The index appears as a model input, not as a view on where it goes.
Sell one, buy two: the ratio page's trade turned round
Skew makes this side of the trade the dearer one. The two bought 10,300 puts carry 15.71% against 14.56% on the sold put, which is why the ratio page can sell this shape for a credit and this page has to pay for it. The position's vega at entry is £100.74 a volatility point: like its delta, the ratio page's figure with the sign reversed.
The worked example: bought for a 16.5-point debit on Monday 17 August 2026
The maximum loss is 18.6 times the £170.10 that leaves the account on day one. Read off the skew surface (model probabilities: risk-neutral, lognormal), a settlement above 10,600, where only the debit is lost, has 65.3%; between the two strikes, where the loss is largest, 14.4%; below the 9,983.5 breakeven, a 7.1% fall, 10.8%.
Open this worked example in the strategy builder (the builder takes the two fills and solves each leg's volatility from them).
The trough at 10,300, and the open side below 9,983.5
At settlement the line is flat at the debit above 10,600, drops £10 a point into the trough at 10,300, then climbs £10 a point for every point lower. Before expiry the curves sit well above the trough: with 30 days left the lowest the position shows is −£440, near 10,640, against the £3,170.10 loss at 10,300 on the last day. The trough is a settlement-day shape that the mark hardly shows until the final fortnight, so the screen understates where the loss will be.
Debit or credit, puts or calls: four backspreads on one chain
Where the bought strikes sit decides whether a backspread costs money or pays it, and on the FTSE 100 the side of the chain matters as much as the distance. The chart overlays the worked example with the same trade using 10,200 puts, which opens for a credit; the table adds a narrow version and the call-side backspread.
Probabilities are model probabilities (risk-neutral, lognormal, from the skew surface), not forecasts. The requirement is the short spread's width on the strategy-based schedule; the second bought option is paid in full. Surface volatilities: 10,200 put 16.10%, 10,500 put 14.94%, 10,900 call 13.45%, 11,200 call 12.36%.
Three things come out of the menu. First, the credit put version is £370.00 better anywhere above 10,600 (a 65.3% model probability), but £630 worse at its trough, and it needs a fall 163 points larger before it pays; below 10,200 the debit version stays £1,630 ahead. Being paid to hold it is bought with the move itself. Second, narrowing the bought strike to 10,500 cuts the worst result to £2,150.10 but raises the debit sevenfold, to 114.5 points: the loss above 10,600 becomes £1,150.10, so the narrow version needs the fall to happen, not just to be possible. Third, the call backspread opens for a 39.0-point credit, because the two far calls are the cheapest options on this surface. It is the same 39 points the ratio page's call mirror pays as a debit. Its worst result, £2,615.10 at 11,200, needs a rally of about 4%, and it pays on a rally beyond 11,461.0. Open the call backspread in the builder.
The structure is long vega in every version, so the level of volatility on the day matters. Over the year to 30 June 2026 FTSE Russell's 60-day FTSE 100 IVI ran from 13.54 to 25.59, averaging 17.24, so the example's 14% at the money sits near the bottom of that range. With every strike 3.24 points higher the same structure would cost 52.22 points, £522.16, which is the ratio page's credit at that level with the sign reversed; two points lower it would cost 2.22 points. The implied volatility page explains how to place a reading in its 12-month range.
What the mark does in a fall: sticky-strike against sticky-delta
A backspread's mark in a sell-off depends on an assumption the payoff diagram never shows: what happens to each strike's volatility when the index moves. Every FTSE page in this library uses sticky-strike, where each strike keeps its own volatility and a stated shift is added. The alternative, sticky-delta (or sticky-moneyness), moves the whole skew with the index, so a strike that becomes closer to the money takes the lower at-the-money level. The implied volatility page explains both. The table prices the backspread under each, with the same parallel shift.
Under sticky-strike the bought puts keep their high skew volatility as the index falls towards them, so the mark straight after an ordinary fall is £423.03 higher than under sticky-delta, where they are re-marked at the new at-the-money level. Neither assumption changes the settlement column: at 10,156.8 on 16 October the structure loses £1,738.10, whatever the mark said a month earlier. The ratio page, whose position is this one reversed, uses the same sticky-strike assumption, so the two pages' stress figures mirror each other exactly.
A requirement that stays at £3,000, and the row that misleads
For this family the stress test comes first. The rows move the index at once on 17 August with a stated parallel volatility shift, and read the mark, the settlement result if the index then stays there, the requirement and the delta. The method is on the Level 3 page; its gap history lists one-day falls as large as the first row.
The requirement column never moves, because a short spread is charged its maximum potential loss at the strikes, here the 300-point width (FINRA Rule 4210(f)(2)(H)(i), FINRA Rule 4210, checked 27 September 2026), and the width does not care where the index is. That is the practical difference from the ratio page's stress table, where the same moves push the requirement from £13,400 to £31,070. It also means the account can be sized once: funded with £6,170.10, the requirement plus the whole maximum loss, no mark can take equity below the requirement. Across index levels from 9,800 to 11,000, every day to expiry and volatility shifts of −5, 0, +10 and +20 points, the worst mark the model produces is −£2,704.01, on the day before expiry at 10,300: it never reaches the settlement maximum. The margin spiral that a ratio can meet does not apply.
The row that misleads is −1 SD. An ordinary 5.5% fall marks the backspread up £1,824.46 at once, while settling at that level would lose £1,738.10: a £3,562.56 gap between screen and settlement. The mark is paid for by the bought puts' volatility and time value, and both have gone by 16 October.
Four endings, with the worked plan's conventions firing first
The worked plan uses three teaching conventions (methods page): it takes profit when the mark reaches half the maximum loss, £1,585.05; it closes at 21 days to expiry (Friday 25 September) if the index is between 10,250 and 10,650, around the trough; otherwise it closes at 7 days to expiry (Friday 9 October). None is a rule. Closing results include commission and half the quoted spread both ways.
Adjusting instead of closing is possible. Rolling the sold put lower narrows the trough, but on a fairly priced chain the debit it costs is the model value of the loss it removes, so it changes the shape of the risk rather than its value; the rolling page sets out the arithmetic of close-or-roll decisions in general.
Greeks: long volatility, and what the trough costs a day
Reverse every sign in the first column and it matches the ratio page's entry Greeks. How position Greeks are stated in pounds is on the Greeks page.
The trough columns show what sitting at 10,300 costs. With 30 days left the position is still long volatility (+£139.93 a point) and long gamma, but it pays £38.00 a day for them; with 7 days left the daily cost is £105.64 and the mark has fallen to −£1,450.37. Gamma there is highest (+£0.284 per 10 points), so a sharp move either way helps, but only a fall of more than 300 points helps enough. In the −2 SD column the structure has become what it was built to be: delta −£6.59 a point, gaining on further falls and on rising volatility.
On BP shares: the short put assigned early, the long puts still open
On ICE single-stock options every leg is American and physically settled, and that changes the backspread's ending. BP's model-sheet inputs apply: a 530p model level for 17 August 2026 (the actual close was 519.6p), one 26% volatility for all strikes, no dividend until the 12 November ex-date, binomial-tree pricing for the American puts, 1,000 shares per contract and £1.40 per contract in commission. BP appears only as a model input, not as a view on the company.
Selling one October 510 put at 12.25p (model 12.27) and buying two 480 puts at 4.50p (model 4.46) opens for a credit of 3.25p, £32.50: without the index's skew, the far puts are cheap enough to pay for themselves. Now suppose BP falls to 455p by Friday 9 October, a week before expiry. The tree values the 510 put at exactly its 55.00p of intrinsic value, 0.00p of time value, while interest on the 510p strike for the week is 0.37p a share, so its holder gains by exercising early and assignment is likely (early assignment of puts). The assignment buys 1,000 BP shares at 510p: £5,100.00, plus £25.50 of SDRT, which the assigned writer pays as the buyer (who pays SDRT), plus £1.40, £5,126.90 in all. The two 480 puts are still open, worth 25.32p each. Exercising one of them delivers the shares at 480p for £4,798.60 after commission; the buyer of the shares then pays the SDRT on that side. What remains is one bought 480 put, the part of the structure that pays on a further fall. The assignment page covers one leg of a spread assigned while the others stay open.
UK tax: a same-year loss, and a February version with a taxable first year
The written 10,600 put is a grant, a disposal on 17 August (TCGA 1992 s144(1)); the bought puts are disposed of when they lapse or settle. Ending on 16 October, all three results fall in 2026/27 and offset each other; other gains are assumed to use the £3,000 exempt amount.
The trap is a February opening. On Monday 15 February 2027, on the April expiry (60 days, same inputs and premiums; a February start, not the usual March, because 60 days from 15 February reach the 16 April expiry), a settlement above 10,600 leaves the written put's £1,773.30 gain in 2026/27, costing £319.19 at 18% or £425.59 at 24%, while the bought puts' £1,943.40 loss falls on 16 April, in 2027/28. A trade that lost £170.10 shows a taxable first year; relief is deferred, and lost only if never used. Settled below 10,600, everything is dated 16 April, in 2027/28. HMRC gives no worked example; the reading follows s144A(2) (tax page: across 5 April, written and bought options). Results go in the SA108's other property section; ISAs cannot hold options and no SIPP administrator allowing them was found (checked 26 September 2026; wrappers).
Costs, the contract, and a spread permission rather than an uncovered one
Costs. £5.10 of commission opens three contracts, none is due at cash settlement, and another £5.10 closes them early. Half an illustrative 2-point quoted spread is £10 a contract each way, £60 round trip, 36% of the £165.00 debit: on a cheap structure the spread is a large part of the cost. There is no SDRT on the cash-settled index contract; on the BP version it is £25.50 at assignment.
Contract and access. The worked example uses ICE FTSE 100 options (ESX): £10 a point, European, cash-settled on the EDSP with the expiring series stopping after about 10:15 on the third Friday (FTSE 100 contracts). The Mini FTSE 100 daily options (8LX, £1 a point) have front five daily expiries plus the third Friday, too short for a 60-day structure, and broker access has to be checked (contract sizes). A bearish index view can also be held through a spread bet or CFD, with different tax (three ways to hold it).
Permission. Every written option here is covered, so no uncovered permission is involved. Taken apart, the structure is a short put spread, which IBKR places at Options Level 3, plus a bought put, Level 2 (IBKR options trading permissions, checked 26 September 2026). At IBKR a spread needs a margin account; its cash accounts take covered calls and cash-secured puts only (accounts and permissions; spread margin).
Other ways to hold a view on a large fall
The ratio page's family table sets the backspread beside the ratio, the bear put spread and two butterflies on the same strikes. For a large fall in particular: