15.6 Playbook 6: a defined-risk Nifty put spread into the RBI decision (7 October 2026)

15.6.1Prerequisites and what this playbook teaches#

Read first (essential) M01 · 1.7–1.8 (costs and tax), M03 · 3.1.1 (how many trades distinguish edge from noise) and 3.9 (sizing), M08 · 8.8–8.9 (Black-76 and the Greeks), M09A · 9A.2 (the chain from the bhavcopy), 9A.7 (the straddle-implied move) and 9A.10 (extracting an event's implied move, with Worked example 9A.5 on this same decision).

For depth: M11 · 11.6–11.7 (the policy rule and transmission behind an RBI decision) and M09A · 9A.9 (the term structure around an event).

What it teaches How to turn a scheduled macro event into an options position whose worst case is known before entry: choose the structure from what the chain says about the event, pick strikes from the implied move rather than from a hunch, size so that the whole debit is 1R, attribute the result to delta, gamma, theta and vega, and then ask whether eleven events are enough to say anything about the rule. They are not.

This is an educational reconstruction of a trade on published end-of-day data, not a recommendation. Fills are assumed at the bhavcopy close plus a stated slippage; nobody could have traded exactly these prices.

15.6.2The setup as testable rules#

A rule is testable if two people applying it to the same data get the same trade. The rules below were written so they can be run mechanically on every RBI decision since February 2025 (15.6.7).

Rule	Statement	Course link
E1 Event	A scheduled RBI MPC decision day $T$ . Entry at the close of the previous session $T-1$ .	M11 · 11.7, M04A · 4A.11.3
D1 Direction	If the Nifty close on $T-1$ is below its 50-day simple moving average, buy a bear put spread; otherwise buy a bull call spread.	M04 (trend filters)
X1 Expiry	The nearest Nifty expiry strictly after $T$ .	M09A · 9A.9
K1 Long strike	The listed multiple of 50 nearest the put–call-parity forward $F$ for that expiry.	M09A · 9A.2
K2 Short strike	$K_1$ minus (puts) or plus (calls) the ATM straddle price, rounded to the nearest 50.	M09A · 9A.7
S1 Size	Lots $=\lfloor 10{,}000 / (\text{debit}\times\text{lot size})\rfloor$ ; zero lots means no trade.	M03 (3.23)
O1 Exit	Close both legs at the close of $T$ . No stop: the debit is the stop.	M04A · 4A.10.1

A bear put spread buys a put at $K_1$ and sells a put at a lower strike $K_2$ . Its value at expiry is $\min(\max(K_1-S,0),\,K_1-K_2)$ , so the most it can lose is the debit paid and the most it can be worth is the width $K_1-K_2$ . The ATM straddle (call plus put at the same strike) is approximately the market's expected absolute move to expiry, $\mathbb E|S_T-K|$ (M09A · 9A.7), so rule K2 places the short strike about one expected move away: the spread pays its maximum if the move to expiry is at least what the market prices.

Why a debit spread and not a short straddle or a naked long put? Rule D1 expresses a direction. The choice between buying and selling premium comes from 9A.10: if the term structure shows a large event variance $e$ , the event is expensive and selling it has a case; if it shows none, buying is cheap relative to what the event could do. A spread rather than an outright put cuts the vega paid to about a quarter and the theta to about a fifth of an outright ATM put's (table in 15.6.4), and caps the loss without a stop order that could fill badly in a fast market (M03 · 3.9, option positions).

15.6.3The scan and the episode#

The event The MPC met on 5–7 October 2026 (RBI press release on the MPC meeting schedule for 2026-27). On Wednesday 7 October it raised the repo rate by 25 bp to 5.50% and moved the stance from "neutral" to "calibrated tightening" (M11 · 11.7). The decision is announced in the morning, inside the cash session.

Data

Nifty 50 daily OHLC, Yahoo ^NSEI, to compute the 50-day SMA.
NSE F&O UDiFF bhavcopies (BhavCopy_NSE_FO_0_0_0_YYYYMMDD_F_0000.csv.zip from nsearchives.nseindia.com) for 6, 7, 8, 9, 12 and 13 October 2026: ClsPric, UndrlygPric, NewBrdLotQty (65).
India VIX, Yahoo ^INDIAVIX.

The scan on Tuesday 6 October, after the close

Quantity	Value
Nifty close	22,776.10
20-day / 50-day SMA	23,146.96 / 23,823.99
D1 verdict	close below the 50-day SMA: bear put spread
Expiry (X1)	Tuesday 13 Oct 2026, 7 calendar days, 5 sessions
Parity forward $F$	22,747.10 (from the strike where call and put prices are closest, $r=5.25\%$ )
ATM strike (K1)	22,750: CE 151.15, PE 154.05
Straddle	305.20, i.e. 1.34% of $F$
ATM implied volatility	12.15% (call and put agree)
India VIX	13.61

The straddle of 305.20 points corresponds, under a normal approximation ( $\mathbb E|X| = \sigma\sqrt{2/\pi}$ ), to a standard deviation of $305.20/0.7979 = 382$ points, 1.68% over five sessions, or 0.75% per session.

Is the event priced as a jump? Worked example 9A.5 ran Proposition 9A.9 on these chains. From the 6 October chains the 13 Oct and 27 Oct ATM vols (12.15% and 12.55%) give $e=-2.8\times10^{-5}$ : the term structure slopes upward and there is no detectable event premium. The 5 October chains gave $\sqrt e = 0.64\%$ , but that curve was inverted from its 1-day point, the mean-reversion confound of 9A.10. The honest reading on 6 October: the market did not price the decision as a large separate event. So buying premium was not paying up for the event, which is the condition under which the playbook buys rather than sells.

K2 $= 22{,}750 - 50\cdot\text{round}(305.20/50) = 22{,}750-300 = 22{,}450$ . The 22,450 put closed at 55.40.

15.6.4The plan before entry#

Written on the evening of 6 October, before the decision.

Thesis. The index is in a downtrend (below both averages, 4.4% below the 50-day SMA). A hawkish surprise or a disappointing stance would extend it. The options market is not charging an event premium, so a debit structure is not overpaying for the event.
Trigger. Rule E1: buy 22,750 PE and sell 22,450 PE (13 Oct expiry) at the 6 October close.
Fill assumption. The bhavcopy has no bid and ask. Assume we pay 0.50 index points of slippage on each option, each way (an assumption: ATM weekly Nifty options usually quote far tighter, so this is deliberately conservative). Net debit $=154.05-55.40+2(0.50) = 99.65$ points.
Size (M03 (3.23)). Risk budget 1% of ₹10,00,000 $=$ ₹10,000. One lot risks $99.65\times65 = $ ₹6,477.25, two lots ₹12,954.50. $Q = 1$ lot. $1R = $ ₹6,477.25.
Stop and invalidation. No stop order: the maximum loss is the debit. The thesis is invalidated by an index close above the 20-day SMA (23,147), but the time stop of rule O1 comes first.
Exit. Both legs at the close of 7 October, whatever happens. The trade is about the event; holding longer is a different trade with a different rule.
Payoff at expiry (for reference, since we exit before it). Maximum value $300\times65 =$ ₹19,500; maximum profit $(300-99.65)\times65 =$ ₹13,022.75 $= 2.01R$ . Break-even index level at expiry $22{,}750-99.65 = 22{,}650.35$ .
Break-even win rate. If outcomes were only "full width" or "zero", $p^* = 99.65/300 = 33.2\%$ ; with round-trip costs of about ₹137 in both outcomes, $p^* = (6{,}477.25+137)/(13{,}022.75-137+6{,}477.25+137) = 33.9\%$ . With an event-day exit the R distribution is continuous, but the same logic holds: the spread is priced so that, under the risk-neutral measure, its expected value equals its cost.
Expected R. Under risk-neutral pricing the expected gross R is approximately zero. Costs and the assumed slippage subtract about $0.021R + 0.020R = 0.041R$ . Any positive expectancy has to come from rule D1, the trend filter, and 15.6.7 asks whether there is evidence for that.
Costs (M01 · 1.7: statutory and exchange charges as of 6 Nov 2026). Four orders: brokerage ₹80 at an illustrative ₹20 per order (brokerage varies by broker); STT 0.15% of premium on the two sell orders; exchange 0.03553% of premium turnover; SEBI ₹10 per crore; stamp 0.003% on the two buy orders; GST 18% on brokerage, exchange and SEBI fees.

Greeks at entry (Black-76 on the forward, $r=5.25\%$ , $\tau = 7/365$ ; implied vols 12.15% at 22,750 and 13.03% at 22,450: the put skew).

	Delta	Gamma	Vega (per vol point)	Theta (per calendar day)
Long 22,750 PE, per unit	−0.4992	0.00104	12.55	−11.30
Short 22,450 PE, per unit	+0.2301	−0.00074	−9.56	+9.04
Spread, per unit	−0.2691	0.00030	2.99	−2.26
Position (65 units), ₹	−17.49 per point	0.020	+194 per vol point	−147 per day

The position gains about ₹17.5 for each point the Nifty falls, loses about ₹147 a calendar day to time decay, and gains ₹194 for each point implied volatility rises. The vega is a quarter of what a single ATM put would carry, which is the point of the spread.

15.6.5Day-by-day management#

Date	Nifty (O / H / L / C)	22,750 PE	22,450 PE	Spread	India VIX	Action and note
Tue 6 Oct	22,603.25 / 22,776.10 / 22,561.60 / 22,776.10	154.05	55.40	98.65	13.61	Bought 1 lot of the spread at 99.65 including slippage. Plan filed.
Wed 7 Oct	22,690.45 / 22,717.65 / 22,546.30 / 22,603.05	227.60	81.70	145.90	13.89	Repo +25 bp to 5.50%, stance "calibrated tightening". Nifty −0.76%. Closed both legs at the close (rule O1): 144.90 after slippage.

The log has two rows because the plan had two days. What happened afterwards is recorded in 15.6.7, not here, because it played no part in the decision.

Two observations from the event day. First, the 13 October implied vols fell (22,750: 12.15% → 11.51%; 22,450: 13.03% → 12.26%) while India VIX rose (13.61 → 13.89). VIX is computed from the near and next monthly series, not the weekly that holds the event, so it is the wrong gauge of the post-event IV crush in a weekly. Second, the crush was small, consistent with 15.6.3: there was little event premium to crush.

15.6.6Exit and result#

Item	Value
Entry debit / exit credit (points)	99.65 / 144.90
Gross P&L	$45.25\times65 =$ ₹2,941.25 $=0.454R$
Costs	brokerage ₹80.00, STT ₹27.59, exchange ₹11.98, SEBI ₹0.03, stamp ₹0.46, GST ₹16.56: ₹136.63
Net P&L before tax	₹2,804.62 $= +0.433R$
Tax	F&O profit is non-speculative business income taxed at the slab rate (M01 · 1.8; see M00 · 0.7). At an assumed 30% slab, before surcharge and cess: ₹841.39
After tax	₹1,963.23, 0.20% of capital

Attribution A Taylor expansion with the entry Greeks, per unit of the spread, for the move from 6 to 7 October ( $\Delta F = 22{,}590.89 - 22{,}747.10 = -156.21$ ; long-strike IV −0.65 points, short-strike IV −0.77 points):

Term	Formula	Points
Delta	$\Delta\cdot\Delta F = -0.2691\times(-156.21)$	+42.03
Gamma	$\tfrac12\Gamma(\Delta F)^2$	+3.67
Theta	one calendar day	−2.26
Vega	$\nu_L\Delta\sigma_L - \nu_S\Delta\sigma_S$	−0.77
Sum		+42.68
Actual (close to close)	$145.90-98.65$	+47.25

The 4.6-point residual is higher-order. As the forward fell toward the middle of the strikes, the long put went in the money and the short put moved toward at-the-money, so the spread's net vega changed sign, from +2.99 to −0.29 per unit. Full revaluation shows it: repricing the 7 October spread at the entry vols gives 144.41 against the actual 145.90, so in full revaluation the IV decline added about 1.5 points. A first-order vega estimate made at entry had the sign wrong by the exit. This is the cross-Greek (vanna) effect that M14 · 14.14 says a delta–gamma–vega expansion misses.

The trade made money for the reason in the thesis, direction, and not because of anything specific to volatility: about 90% of the move is the delta term.

15.6.7Post-mortem#

Process Every rule was followed. The fill assumption is the weakest link: closing prices are the last trades, not prices you could necessarily have dealt at, and on a decision day the close of each leg can be minutes apart.

What happened after the exit Not part of the trade, but worth seeing:

Date	Nifty close	Spread (close)
Thu 8 Oct	22,231.80	256.70
Fri 9 Oct	22,520.45	159.20
Mon 12 Oct	22,363.35	272.05
Tue 13 Oct (expiry)	22,721.95	28.05 (intrinsic: the 22,750 put settles at 28.05, the 22,450 put at zero)

Holding one more day would have earned about $1.1R$ more (an exit at 255.70 after slippage is $+1.57R$ gross); holding to expiry would have turned a $+0.43R$ trade into roughly $-0.7R$ before costs. Neither is evidence that rule O1 is wrong. A rule judged on the path that followed one trade is a rule fitted to one observation. If you want a different exit, write it down and run it on all the events below.

Base rates The same rules, applied mechanically to every RBI MPC decision from February 2025 to October 2026 (decision days from the RBI's press releases on the MPC meeting schedules for 2025-26, issued 26 March 2025, and for 2026-27), with bhavcopy closes, 0.50-point slippage per leg and the November 2026 charge stack throughout (STT was lower before April 2026, so early trades are charged slightly too much):

Decision	D1	Expiry	Strikes	Debit	Lots	Nifty on $T$	Net R
7 Feb 2025	put	13 Feb	23,650/23,250	130.05	1	−0.18%	+0.118
9 Apr 2025	put	17 Apr	22,600/21,950	181.40	0	−0.61%	no trade
6 Jun 2025	call	12 Jun	24,800/25,150	122.60	1	+1.02%	+0.593
6 Aug 2025	put	7 Aug	24,650/24,450	63.30	2	−0.31%	+0.271
1 Oct 2025	put	7 Oct	24,700/24,450	87.45	1	+0.92%	−0.711
5 Dec 2025	call	9 Dec	26,050/26,250	79.00	1	+0.59%	+0.563
6 Feb 2026	put	10 Feb	25,650/25,400	81.60	1	+0.20%	−0.271
8 Apr 2026	put	13 Apr	23,100/22,450	196.90	0	+3.78%	no trade
5 Jun 2026	put	9 Jun	23,500/23,200	110.55	1	−0.21%	+0.146
5 Aug 2026	call	11 Aug	24,500/24,800	107.90	1	+0.04%	+0.231
7 Oct 2026	put	13 Oct	22,750/22,450	99.65	1	−0.76%	+0.433

Lot size was 75 for the 2025 expiries in the table and 65 for the 2026 ones (NewBrdLotQty). Twice the rule skipped the event because one lot of a wide spread exceeded the 1% budget; one of those skipped days, 8 April 2026, was a +3.78% Nifty day, a move whose size suggests news beyond the policy decision (we have not attributed it). Skipping it was the rule, not luck or judgment. The mean absolute Nifty move on the eleven decision days was 0.78%, close to the straddle-implied 0.75% per session on 6 October: on average these decisions were ordinary days.

On the nine trades: seven winners, mean $\bar R = +0.153$ , standard deviation $s = 0.417$ , $t = \bar R\sqrt9/s = 1.10$ (one-sided $p = 0.15$ with 8 degrees of freedom). By M03 (3.2a), if the true expectancy were 0.153R, reaching $t=2$ in expectation needs $N = (2s/\bar R)^2 \approx 30$ trades; for 80% power, (3.2b) gives $N \approx 60$ . RBI decides six times a year and the rule trades about nine times in eleven, so 60 trades take about 12 years. Before then the policy regime, the lot size, the expiry calendar and the STT will all have changed again (M01 · 1.7 lists two STT increases in eighteen months). The sample cannot tell a trend-filter edge from noise, and no practical sample of this one event type can. If the idea has an edge, it has to be established on a broader set of scheduled events (RBI, Budget, US CPI and FOMC nights, quarterly results for stocks) under one rule, with the M04 multiple-testing safeguards.

What to keep The structure: a debit spread sized so the debit is 1R removes gap risk and stop slippage from an event trade. The habit: read the term structure (9A.10) before choosing whether to buy or sell the event. The humility: a 7-for-9 record looks good and proves nothing.

15.6.8Do it in the simulator#

The simulator fills NIFTY options at the bhavcopy settlement price of the first session that closes after the order is placed, and charges the M01 · 1.7 cost stack (see the simulator's page for the exact rules).

Before the next RBI decision (the MPC meets 2–4 Dec 2026; decision on 4 December per the RBI calendar), run rules E1–S1 on the chain of the previous session from the option chain page. Record $F$ , $K_1$ , the straddle, $K_2$ , the debit and the lot count in your journal, with the term-structure reading of 9A.10.
On the session before the decision, open Simulator → New order: buy the $K_1$ option and, as a second order, sell the $K_2$ option, same expiry, NRML, market, qty in lots. Use Preview to check that the risk amount is at most 1% of equity and that the short leg's margin is covered (the simulator's simplified margin treats the short option on its own).
On the decision day, place market orders to close both legs. Because fills are at the settlement price of the next processed day, your exit is the decision day's settlement only if the order is in before that day's close; note the timing in your journal.
Compare the simulator's fills and costs with your plan. Compute the R multiple and do the Greek attribution of 15.6.6 with your own numbers. Add the trade to the base-rate table above; it is the twelfth observation, not a verdict.

See it live: India VIX and the Nifty 50 at https://oicharts.com/indices in the sessions before the decision; the event's implied move itself comes from the chain (M09A · 9A.10).