From a loss you can take to a quantity
Position size has one correct derivation and it runs backwards: decide the rupees you can lose, divide by the loss per unit, and the quantity falls out. Margin available is not an input.
Most traders size positions by answering the wrong question.
The wrong question is "how much can I buy?" It has an easy answer — the broker shows it, right there on the order ticket — and it is the answer that ends accounts.
The right question is "if I am wrong, what do I lose?" It is harder because you have to decide something, but once decided, the quantity is arithmetic.
The only equation you need
Quantity = risk budget ÷ risk per unit
Two inputs, in this order.
Risk budget is the rupees you are willing to lose on this trade. Not a percentage you feel good about, a number of rupees. Convert it from a percentage of capital by all means, but write down the rupees, because rupees are what you will actually watch disappear.
Risk per unit is what one unit costs you if the trade goes wrong. For a directional position it is the distance from entry to stop. For a long option held to expiry with no stop, it is the entire premium — there is no other honest answer.
Choosing the risk budget
As a fraction of capital, the conventional range is 0.5% to 2% per trade, and the conventional range is roughly right for reasons the chapter on risk of ruin makes precise.
The intuition: you need to survive a bad run long enough for your edge to show up. At 1% per trade, eight consecutive losses costs you about 8% — unpleasant, recoverable, and not a threat to your ability to keep trading. At 10% per trade, those same eight losses take 57% of your account, and recovering from that requires more than doubling what is left.
Eight consecutive losses is not a disaster scenario. At a 55% hit rate it happens roughly once in every five hundred trades, which is a year or two of ordinary activity. Your sizing has to assume it, because it will arrive.
One refinement that matters more than the exact percentage: the budget should be a fraction of current capital, not starting capital. Recalculated after losses, it shrinks your size when you are losing, which is precisely when you want it smaller.
The worked case
Capital ₹6,00,000. Risk budget 1% = ₹6,000.
You want to buy an option at ₹120 and you will exit if it trades at ₹102. Risk per unit is ₹18.
Quantity = 6,000 ÷ 18 = 333 units.
Lots come in fixed sizes, so you cannot buy 333. You buy 300 — always rounding down, because rounding up means choosing a number you already said was too large.
Actual risk: 300 × ₹18 = ₹5,400. Slightly under budget, which is the correct side to be on.
Now notice what the margin route would have given you. If the broker asks ₹640 a unit, your ₹4,80,000 of free margin permits 750 units — and 750 × ₹18 is a ₹13,500 loss, two and a quarter times the risk you had decided on, arrived at without deciding anything.
Four details that bite
Round down, always. The quantity you compute is a ceiling, not a target. Rounding 333 up to 375 because it is a neater number is a 12% increase in risk chosen for aesthetic reasons.
The stop has to sit on a valid tick. Instruments trade in fixed price increments, and a stop price that falls between them is not a tighter stop — it is a rejected order. Compute the level you want, then snap it to the instrument's tick size, and recompute the risk per unit from the snapped price rather than the ideal one. The platform's order ticket does the snapping for you; a spreadsheet will not.
A quantity cap is a seatbelt, not a limit. Setting a maximum order quantity for yourself is worth doing, and the point of it is to ask rather than to refuse — the one time you genuinely mean to place something large, you should be able to, after confirming you meant it. A cap that silently blocks you teaches you to raise the cap; a cap that asks teaches you to notice.
For a long option with no stop, risk per unit is the whole premium. It is tempting to reason that it will not go to zero. Weekly options expire worthless all the time. If you have no exit rule, size as though the premium is gone, because sometimes it is.
Where this sizing rule does not apply
The equation above assumes you can state a loss per unit. Three cases where you cannot, and what to do instead:
Short options. The loss per unit has no upper bound, so there is no number to divide by. You size these on margin and stress tests rather than on a stop: how much does this position lose if the underlying gaps 3% against me overnight? That figure, not a stop distance, is your risk per unit.
Multi-leg structures. The basket's maximum loss is a property of the whole structure, and it assumes every leg is in place. Sizing on the basket's theoretical maximum is right only if the structure can be relied on to exist in full — which is the subject of the fourth chapter in this module.
Positions with a gap-risk tail. A stop at ₹102 does not guarantee an exit at ₹102. If the instrument can open below your stop — and almost all of them can — your real risk per unit is larger than the stop distance by an amount nobody can know in advance. The usual response is to size on the stop and accept that the tail exists, while never letting a single position be large enough that the tail alone would be serious.
What to take from this chapter
- Quantity = risk budget ÷ risk per unit. Both inputs are decided before the quantity exists.
- Margin available answers a question about the broker's risk appetite, not yours. It is not a sizing input.
- 0.5% to 2% of capital per trade, recalculated on current capital so your size shrinks when you are losing.
- Size so that eight consecutive losses is survivable, because at a 55% hit rate it arrives about once in five hundred trades.
- Round the quantity down. Snap the stop to a valid tick and recompute the risk from the snapped price.
- A long option with no exit rule risks the entire premium. Size it that way.
- Short options and multi-leg structures need stress tests, not stop distances.