Risk Management 101

There is an order of operations in the market that almost everyone learns backwards. Return is what you hope for. Risk is what you decide. Only one of those is under your control, and it is the one most beginners never consciously set.

Think of a kirana shop owner: he does not bet his whole stock on one festival. He keeps enough back to open the shutter next morning. That is what being 'in the game' means. Every method, however good, produces losing trades. A method that is right 55% of the time will still hand you six losses in a row eventually — that is not bad luck, it is ordinary probability. Risk management is what makes those six losses an inconvenience rather than an ending.

There is also a compounding argument. A participant who loses a small, fixed amount when wrong keeps most of their capital through a bad stretch, and therefore has most of their capital available when conditions improve. A participant who loses large amounts arrives at the good stretch with too little capital left for it to matter.

So the sequence for every position is: decide what being wrong costs, then work out the size, then place the order. Never the other way around. The moment you ask 'how many shares can I afford' before 'how much am I willing to lose', the arithmetic is already broken.

Losses and the gains needed to undo them are not symmetric, and the asymmetry gets vicious quickly. The reason is that a percentage loss is taken on a larger base than the percentage gain that must repair it.

Work it through with real numbers. Start with ₹1,00,000 and lose 50%. You now have ₹50,000. To get back to ₹1,00,000 you must gain ₹50,000 on a base of ₹50,000 — that is a 100% gain. The loss was 50%; the repair is 100%.

The formula is short: required gain = loss ÷ (1 − loss). At a 20% loss that gives 0.20 ÷ 0.80 = 25%. At a 60% loss it gives 0.60 ÷ 0.40 = 150%. The curve is flat and forgiving in the single digits and becomes almost vertical past 50%.

The practical consequence is a change in what you optimise for. Beginners try to maximise the size of their wins. The arithmetic says something different: keep every individual loss inside the flat part of this curve, and the recovery cost never becomes the problem.

Notice also what the table implies about time. A 70% drawdown needs a 233% gain to repair. Even if that is achievable, the years spent getting back to where you started are years the capital produced nothing.

Risk of ruin is the chance that a normal run of losses reduces your capital to a point from which recovery is no longer realistic. It depends on three things: how much you risk per trade, how often you are wrong, and how long the streak runs.

You do not need probability theory to feel the force of it. Just compound the losses. If you risk a fixed percentage of your remaining capital each time, then after n consecutive losses you hold (1 − risk) raised to the power n of what you started with.

Read the table below slowly, because it contains the entire argument for small position sizes. At 1% risk per trade, twenty losses in a row — a genuinely miserable stretch — leaves you with about 82% of your capital and needs a 22% gain to repair. At 20% risk per trade, the same twenty losses leave you with roughly 1% of your capital, and the repair is arithmetically out of reach.

Also notice something subtle. Doubling risk from 1% to 2% does not double the damage of a streak — it roughly doubles it early and worsens faster later. Risk compounds against you in exactly the way returns are supposed to compound for you.

A streak of ten or twelve losses is not a sign that your method is broken. It is a thing that happens. The only question that matters is whether your position size made it a bruise or a burial.

The 1% rule states that the maximum you lose on any single position, if your stop is hit, is 1% of your total trading capital. Not 1% of the position — 1% of everything.

Be precise about what is being measured. On ₹5,00,000 of capital, 1% is ₹5,000. That ₹5,000 is your maximum loss on the trade, not the amount you invest. The amount invested might be ₹1,30,000 or ₹40,000 depending entirely on where your stop sits.

This distinction is where most beginners go wrong. 'I only put ₹20,000 into it' says nothing about risk. If that position has no stop, the risk is ₹20,000. If it has a stop 4% below entry, the risk is ₹800. The rupee amount deployed and the rupee amount at risk are different numbers, and only the second one is governed by this rule.

Some people run 2% instead of 1%, and some run 0.5% while learning. The specific figure matters less than the fact that it is fixed in advance and applied to every position without exception. A rule that bends for the trade you feel strongly about is not a rule — and the trade you feel strongly about is exactly the one that needs it.

Practical note for small accounts. On ₹25,000 of capital, 1% is ₹250, which will often be too small to take a sensible position after costs. The honest response is to trade fewer, smaller positions and treat the constraint as information about account size, rather than to quietly raise the risk percentage until the trade fits.

Position sizing is the arithmetic that turns your risk rule into a share quantity. It needs four inputs and one division. There is no judgement in it once the stop is chosen.

The four inputs are: total capital, risk per trade as a percentage, the entry price, and the stop price. From those you compute the risk in rupees, then the stop distance per share, then the quantity.

Worked example one, illustrative throughout. Capital ₹5,00,000. Risk 1%, so ₹5,000. A share you are considering trades at ₹640, and the support zone that would invalidate the idea sits just under ₹620, so the stop goes at ₹616. Stop distance = ₹640 − ₹616 = ₹24 per share. Quantity = ₹5,000 ÷ ₹24 = 208.3, rounded down to 208 shares. Actual risk = 208 × ₹24 = ₹4,992. Position value = 208 × ₹640 = ₹1,33,120, which is about 27% of capital — affordable.

Always round the quantity down, never up. Rounding up puts you above your own rule, and the habit of rounding up is how 1% becomes 1.4% becomes 3%.

Worked example two, where the sanity check bites. Capital ₹1,00,000, risk 1% = ₹1,000. Share at ₹1,850 with a tight stop at ₹1,844 — a distance of only ₹6. Quantity = ₹1,000 ÷ ₹6 = 166 shares. But 166 × ₹1,850 = ₹3,07,100, which is three times the entire account. The capital constraint binds before the risk constraint. The correct response is to size to what capital allows — at most 54 shares — which puts actual risk at 54 × ₹6 = ₹324, or 0.32%. Under your limit, which is fine. What is not fine is borrowing to reach 166 shares because the arithmetic 'said so'.

Worked example three, the wide stop. Capital ₹1,00,000, risk ₹1,000, entry ₹500, and the structure puts the stop at ₹440 — a distance of ₹60. Quantity = ₹1,000 ÷ ₹60 = 16 shares, a position of ₹8,000. It feels absurdly small. It is correct. A wider stop must be paid for with a smaller position, and the fact that this feels wrong is precisely why so many people skip the calculation.

A stop-loss belongs at the price where your reason for the position stops being true. That is a structural question, not a percentage one.

Percentage stops sound disciplined and are usually arbitrary. 'I always use a 5% stop' means your exit is determined by a number you chose at home rather than by anything happening in the market. On one share, 5% sits far below a level that has held four times. On another, 5% sits inside the ordinary daily range and will be hit by noise alone.

The structural approach is different. If your reason for the position is that a support zone from ₹618 to ₹625 has held repeatedly, then your idea is wrong if price closes clearly below that zone. The stop goes just below it — say ₹616 — because that is the price that falsifies the reason.

Two refinements are worth adopting early. First, place the stop a little beyond the level rather than exactly on it, because levels are zones and price routinely pokes through by a rupee or two. Second, consider using a closing price rather than any touch as your trigger, so that a brief intraday spike does not remove you from an idea that is still intact.

Once the stop is set by structure, the position size adjusts to fit — not the other way around. If the structurally correct stop is wide and the resulting position feels too small to be interesting, that is the market telling you this particular setup is expensive in risk terms. Moving the stop closer to make the position larger is the same as deciding to be wrong more often, on purpose.

A mental stop is a price you have decided you will exit at, without placing any order. The intention is real. The protection is not.

The problem is that a mental stop asks the version of you under maximum emotional pressure to execute the decision made by the version of you who was calm. Those are not the same person. At the moment the level breaks, you will have reasons — the market is weak today, it will bounce, let me see the close, the news is temporary.

There are also mechanical failures that have nothing to do with willpower. You may be in a meeting, travelling, or asleep before a market open. Your internet may fail. A resting stop order does not require you to be present, alert or calm.

The honest counterargument deserves a hearing. A resting stop order can be triggered by a brief spike that immediately reverses. That is a real cost. The answer is to place the stop beyond the noise — a little past the structural level rather than on it — not to remove the order entirely.

There is also a limit worth stating plainly: no stop order guarantees an exit price. If a share gaps below your stop level at the open, the exit happens at the market price then available, which can be well below your stop. That is why position size assumes the worst case, and why a stop is a risk-control tool rather than a risk-removal tool.

The risk-reward ratio compares what you lose if wrong to what you gain if right. Risking ₹24 per share to make ₹48 is 1:2. It is usually written in units of R, where 1R is your risk — so a 1:2 trade targets 2R.

Here is the point almost everyone misses. A win rate on its own tells you nothing, and a risk-reward ratio on its own tells you nothing. They are two halves of one equation, and only together do they say whether an approach makes money.

The breakeven win rate for any ratio is 1 ÷ (1 + R). At 1:1 you need to win 50% of the time to break even. At 1:2 you need 33.3%. At 1:3 you need 25%. At 1:0.5 — risking two rupees to make one — you need to be right 66.7% of the time just to stand still.

Worked example. Twenty trades at 1:2, risking ₹5,000 each, winning 8 of them. Wins: 8 × ₹10,000 = ₹80,000. Losses: 12 × ₹5,000 = ₹60,000. Net +₹20,000, while being wrong 60% of the time. Change nothing except the ratio — make it 1:1 — and the same 8 wins give ₹40,000 against ₹60,000 of losses, a net loss of ₹20,000.

Two honest caveats. First, all of this is before costs; brokerage, exchange charges, STT, stamp duty and GST come out of the wins and add to the losses, so the real breakeven win rate is a little higher than the table shows. Second, the reward figure must be a level price can plausibly reach, not a number chosen to make the ratio look acceptable. Inventing a target to justify a trade is the most common way this arithmetic gets quietly falsified.

You can follow every rule so far and still take one enormous, undiversified bet without noticing. It happens through correlation — the tendency of certain shares to move together because they respond to the same underlying facts.

Suppose you hold five positions, each sized at exactly 1% risk. On paper that is five separate ideas totalling 5% of capital at risk. Now suppose all five are lenders. A single policy announcement, a change in interest rates, or a sector-wide concern moves all five in the same direction on the same morning. Your five 1% risks behave like one 5% risk.

The same thing happens with themes rather than sectors: five different companies that all depend on the same export market, or on one commodity price, or on one government programme. The company names differ. The underlying bet is identical.

There is a further layer in a market-wide fall. In a sharp decline, correlations across almost everything rise toward one — shares that normally behave independently all fall together. Diversification is weakest exactly when you need it most, which is an argument for keeping total exposure modest rather than for abandoning diversification.

The practical fix is to count exposure by theme, not by ticker. Before adding a position, ask what would have to be true in the world for this position and my existing ones to lose money on the same day. If the answer is 'one thing', you are adding to an existing bet, not making a new one.

Portfolio heat is the sum of the risk on every open position — the total you would lose if every stop were hit. It is the portfolio-level version of the 1% rule, and it is the number most beginners have never calculated.

Compute it directly. Five open positions, each risking ₹5,000 against ₹5,00,000 of capital, gives ₹25,000 of open risk, or 5% heat. If all five stops hit in a bad week, that is your worst case before slippage and gaps.

Setting a maximum matters because opportunities cluster. When the market is moving, four or five setups appear at once, each perfectly reasonable in isolation. Without a heat cap, an enthusiastic week quietly produces 10% of capital at risk simultaneously, all of it in the same market conditions.

A common structure looks like this: a per-position cap of 1%, a total heat cap of somewhere around 5% to 6%, and a rule that a new position can only be opened if it fits under the cap. Once you are at the cap, the next setup — however good — waits until an existing position is closed or its stop is moved to breakeven.

That last mechanic is useful. When a position moves in your favour and you move its stop up to your entry price, its risk becomes roughly zero and its heat is released, freeing capacity for the next idea. Heat management turns risk into a budget you spend and recover, rather than a number you discover afterwards.

Exchanges halt trading in a share when it moves too far too fast, because extreme moves and clear thinking do not coexist. The same logic applies to a person. A drawdown limit is a circuit breaker you write for yourself, in advance, in numbers.

It has two parts. A loss limit — a level of drawdown at which you stop for a defined period. And a re-entry condition — what has to happen before you resume at full size.

An example structure, described as a framework rather than a prescription. Three losing positions in a single day ends trading for that day. A 6% drawdown from your peak capital halves your position size until you recover half of it. A 10% drawdown stops new positions entirely until you have reviewed the last twenty trades in your journal and identified whether the losses came from following your rules or breaking them.

That distinction is the real purpose. Losses from following your rules are variance, and the response is to keep going at reduced size. Losses from breaking your rules are a process failure, and the response is to fix the process before risking more capital. Without a written limit, you will never pause long enough to tell the difference.

The numbers are yours to choose. What is not optional is that they are chosen before the drawdown, written down, and expressed as prices or percentages rather than feelings. A limit invented during a bad week will always be set just below wherever you currently are.

Averaging down means buying more of a position that has moved against you, in order to reduce your average purchase price. It is the single most common account-ending habit, and it is dangerous precisely because it can be described in the language of discipline.

Look at what actually happens to the arithmetic. Suppose you buy 200 shares at ₹640 with a stop at ₹616, risking ₹4,800 — within your 1% limit on ₹5,00,000. Price falls to ₹600, your stop level has already been breached, and you buy 200 more. You now hold 400 shares at an average of ₹620. Your original invalidation level is gone, and there is no new one, because the reason you bought has already been proven wrong.

The risk has not been averaged. It has been doubled, in a position that has already demonstrated your read was incorrect, and the loss required to reach it was accepted rather than acted on. Every rule earlier in this lesson has now been broken in a single order.

There is a legitimate cousin that gets confused with this, and the difference is worth stating clearly. A pre-planned staggered entry — deciding before you enter that you will build a position in two or three tranches at levels you specified in advance, with one stop below all of them and a total risk still inside your limit — is a sizing method. Adding to a loser after your stop has been breached, without a plan, is not. The test is simple: was the additional purchase written down before the position was opened?

The opposite habit is the one worth building. When a position is wrong, it is closed at the level you decided when you were calm. When a position is right, size can be added only in the direction of the move and only if total risk remains inside the rules.

Risk management is the whole reason the rest of your analysis has a chance to work. It does not make you right more often. It makes being wrong survivable, repeatedly, for as long as it takes. Markets carry real risk and capital can be lost; nothing in this lesson is advice, a recommendation, or a promise about any outcome — it is education about method, and the arithmetic is yours to apply.