Risk and position sizing: the part that decides whether you're still here in a year

Survival first

Trading edge compounds only if you survive to use it. A strategy with a 70% win rate and a 20% chance of ruin per year is a strategy that ruins you within five years with near certainty. Everything in this chapter is about pushing that ruin probability toward zero while keeping the edge intact.

Size by worst plausible loss, not by margin

Margin (Chapter 3) is the exchange's estimate of a bad day or two. It is not your loss limit. For every position, compute the loss at a move the market has actually produced — NIFTY has gapped 4–5% overnight several times in the last decade; individual stocks 20% on results. For defined-risk structures that is simply the max loss the builder shows. For undefined ones (short strangles, naked options, futures) it is the loss at a 4–5% adverse gap, which is often five to ten times the margin.

Put numbers on it with today's chain. Below is a live iron condor: its max loss is fixed and known before entry, so the 2% rule turns directly into a lot count. The short strangle in Chapter 10 has the same strikes without the wings — same credit give or take, and a max loss the model calls unlimited; there, the number you size by is what a 4–5% gap would do.

Stops that work, and ones that don't

• Price stops on the underlying work: “if NIFTY trades 23,400 I'm out.” They express your thesis being wrong.
• Premium-based stops on short options work if wide: “buy back if the option doubles.” Rule of thumb for premium sellers: exit at 2× the credit, take profit at 50% of it. The asymmetry is deliberate.
• Tight premium stops on long options do not work: theta and bid-ask noise trigger them constantly. A long option's stop is its size — buy only what you can afford to lose entirely.
• No stop protects against a gap. Overnight risk is sized, not stopped. Defined-risk structures are the only real overnight protection.

Liquidity and slippage

NIFTY and BANKNIFTY weeklies are the most liquid options on earth; most stock options are not. Look at the bid-ask spread before you trade: a ₹2 spread on a ₹10 option is 20% round-trip cost before anything happens. Rules: trade strikes with visible OI and volume, use limit orders, avoid the first and last ten minutes of the day for anything but the index, and never let a position grow to a size you could not exit in one order.

Events

Know every scheduled event inside your expiry: RBI policy, the Union Budget, US Fed decisions, index-heavyweight results, and for stocks their own results dates. Either your position is designed for the event (a vega-neutral spread, a small defined-risk straddle) or it should not be open through it. Unscheduled events — geopolitical shocks — are what the 1–2% rule is for.

A pre-trade checklist

1. What is my view in four parts (direction, magnitude, timing, vol)? If I can't state it, no trade.
2. Is IV high or low for this underlying (percentile)? Does the structure I've chosen want that?
3. Where are the breakevens versus the implied move (the live straddle)? Am I inside or outside what the market expects?
4. What is the worst plausible loss in rupees, and is it under 2% of capital?
5. What is my exit for right, for wrong, and for “nothing happened”?
6. Any events inside the expiry? Is this a stock option that could be physically settled?
7. Can I exit this size in one order at a fair price?

Seven questions, sixty seconds. They would have prevented most of the losses in SEBI's 93%.

Finally: the mistakes everyone makes once, and the glossary.