Swing trading rules: the risk governance manual

1.Risk a fixed fraction of capital on every trade

Risk the same small fraction of current equity on each trade, sized from the distance to the stop. Not a fixed rupee amount, not a feeling about conviction, a fixed fraction.

State it as a number and hold it: risk a small fraction of the account, commonly in the region of half a percent to one percent, on every position. Because the fraction is fixed to current equity, the rupee amount at risk falls automatically as the account draws down and rises as it grows. You never decide size in the moment; the stop distance and the fraction decide it for you. A wider stop buys fewer shares for the same risk, a tighter stop buys more, and the loss if the stop triggers is the same fixed fraction either way.

The mechanism is risk of ruin and the arithmetic of expectancy. A positive-expectancy method still delivers losing streaks, and a streak of fixed-rupee bets is a straight line towards zero, while a streak of fixed-fraction bets is a decaying curve that mathematically cannot reach it. Fixing the fraction converts an existential risk into a survivable drawdown, which is the entire game: you cannot compound an edge from an account that has been wiped out. The formal version of this trade-off, how the probability of ruin collapses as the fraction shrinks, is worked through in the risk-of-ruin calculator.

2.Cap total open risk across correlated positions

Sum the risk of every open position as if all stops trigger at once, and hold that sum under a fixed ceiling. Count correlated positions as one.

Per-trade discipline is necessary and not sufficient. If every trade risks one percent but twelve are open, the book is risking twelve percent on a single ugly session, which no per-trade rule prevents. The second rule caps the sum: total open risk, the loss if every stop fills at once, stays under a fixed ceiling, commonly around four to six percent of equity. When the ceiling is reached, the next setup waits, however good it looks, until an existing position closes and frees risk budget. The portfolio, not the individual trade, is the unit that has to survive.

The mechanism is correlation, and it is where most sizing discipline quietly leaks. Three long positions in the same sector are not three independent bets: they share a driver, so one sector shock moves all three the same way at the same time, and they hit their stops together. Counted naively that looks like three one-percent trades; in reality it is close to one three-percent trade. Correlation collapses diversification exactly when you need it, in a shock, so correlated exposures must be summed as a single risk and charged once against the cap. The worked correlated-exposure example further down makes the true combined risk explicit.

3.Size every position as if it can gap through the stop

Do not size off the stop distance alone. Size off a plausible adverse gap that opens past the stop, and take the smaller share count.

A stop is a trigger, not a guaranteed fill. On a swing position carried overnight, news can gap the price straight past your level, and the fill comes at the open, not at the stop. So sizing purely on the distance from entry to stop understates the risk of exactly the trades that hurt. The rule is to size on the loss you would take if the price gapped through the stop to a plausible adverse open, and to keep that figure inside your per-trade limit. The share count that satisfies the gap scenario is smaller than the one the stop distance alone would allow, and the smaller number is the one you use.

The mechanism is the difference between the risk you planned and the risk you carry. Work an example. An idea near ₹500 has a technical stop at ₹480, so the stop distance is ₹20; on a ₹3,000 per-trade limit that naive sizing buys 150 shares. But a bad-news gap could open the position near ₹450, a ₹50 loss per share. At 150 shares that gap is a ₹7,500 loss, more than double the limit you agreed to. Size on the ₹50 gap instead and the position is 60 shares, whose gap loss is the ₹3,000 you actually budgeted. You give up reward on the quiet majority of days to make the loud day survivable, which is the whole point of governance.

4.Take no fresh position into a scheduled event

Do not open a new position into a known binary or scheduled event. The defined list: company results, monetary-policy decision days, and expiry-day idiosyncrasies.

Before any entry, check the calendar. If a scheduled binary event falls inside the intended holding window, the position is not taken, or an existing one is trimmed ahead of it. The list is specific and non-negotiable: company results, which can gap a single name violently; monetary-policy decision days, which move the whole market at once; and expiry-day idiosyncrasies, where derivative settlement distorts the cash price in ways a swing thesis never priced in. An event that lands after you are already in, with a stop respected and size correct, is a managed risk. Opening fresh risk into a known event is a choice to gamble.

The mechanism is the absence of edge and the presence of gap risk together. Your method has a structural edge on the alternation of impulse and pause; it has no edge on the direction of an earnings surprise or a rate decision, which is close to a coin flip. Worse, the event is exactly the kind of catalyst that gaps the price straight through a stop, so the one time you most need the stop to work is the one time it is most likely to be jumped. Trading an event mixes a zero-edge bet with the highest slippage conditions of the calendar, which is the opposite of what the rest of this manual is built to do.

5.Let the market regime govern net exposure

The book's net exposure obeys the market filter. When the regime is not aligned with your direction, you carry less, or nothing, regardless of how good the individual setups look.

Governance sets not just per-trade size but the total the book is allowed to carry, and that total is gated by the regime. When the market filter is aligned, up for a long-only swing method, the book may run toward its normal open-risk cap. When the filter turns, the response is mechanical: new exposure is reduced or halted, and the book is allowed to run down as positions close. You do not argue with the gate with a chart of a single attractive name. The regime rule is the top-level switch that decides whether the rest of the machine is even allowed to run.

The mechanism is that a directional edge is conditional on the environment. A continuation method has positive expectancy only while the broader market trends with it; run the same setups into a downtrend or a high-volatility chop and pauses stop resolving your way, breakouts fail, and the edge inverts from positive to negative. Cutting net exposure when the regime is unfavourable is not a market call, it is refusing to deploy a tool outside the conditions it works in. How to define that filter so it reads the same way every week, an index against a long moving average with volatility contained, is set out in the regime-filters guide.

6.Log every trade with its reason, and review weekly

Every trade is logged with its reason at entry: the setup, the level, the stop, the size, the event check. The book is reviewed on a fixed weekly cadence, judged in aggregate.

The rule has two halves and both are mandatory. At entry, before the outcome is known, you record why: the setup being traded, the level that makes it valid, the stop, the intended size, and the event calendar you checked. At exit, you record what actually happened and whether the plan was followed. Then, once a week on a fixed cadence, you read the aggregate rather than the last trade: expectancy by setup, whether stops were respected, whether size followed the rule, whether open risk stayed inside the cap. The review grades the process against its own rules, not the mood of the most recent result.

The mechanism is that a single trade carries almost no information and a batch carries a lot. Any one outcome is dominated by variance: a good decision can lose and a bad decision can win, so reacting to the last trade optimises for noise. The reason logged before the outcome is what lets you separate a sound decision from a lucky one later, which is impossible to reconstruct from memory after the fact. The weekly aggregate is where a rule that slipped, a stop quietly widened, a size crept up, gets caught while it is still one incident and not yet a habit. What a rigorous log actually contains, and how to grade it, is the subject of the trade-journal grader.

7.Run a drawdown circuit-breaker

Cut position size at defined equity-curve thresholds, and stop for the month at a hard maximum loss. The de-risking is pre-committed, not decided in the drawdown.

Set the thresholds in advance and let the equity curve trip them. A common shape: at a first drawdown threshold, position size is cut, say, to half; at a deeper threshold, new risk goes to zero and the book is flat; and a hard monthly maximum loss stops trading until the next period. None of this is decided while you are down, when judgement is worst; it is decided now, in calm conditions, and executed mechanically when the curve crosses a line. The circuit-breaker takes the size decision out of your hands at exactly the moment you are least able to make it well.

The mechanism is the geometry of losses. Drawdowns compound against you: a book down 20 percent needs a 25 percent gain to recover, and a book down 50 percent needs to double just to get back to even. De-risking into a drawdown is not fear, it is arithmetic that protects the capital base you need to compound back from. By shrinking size as equity falls, the circuit-breaker slows the descent through the region where recovery maths turns brutal, and it guarantees that a bad run ends with an account still large enough for a later, smaller edge to rebuild. Continuing at full size into a losing streak is how a survivable drawdown becomes a terminal one.

8.Never widen a losing stop

Once set, a stop moves only in your favour. It is never widened to give a losing trade more room, and you never add to a loser. This rule has no exceptions.

The stop is fixed at entry by where the idea is proven wrong, and from that moment it may ratchet toward profit but never away from it. When price approaches the stop, the temptation is to move it down and give the trade room, or to average down and lower the effective cost. Both are the same error wearing two hats: each converts a loss you defined in advance into one you no longer control. The rule is stated absolutely on purpose, because it is the rule most often broken and the one whose breach does the most damage: never widen a losing stop, and never add to a losing position.

The mechanism is expectancy. A positive-expectancy system depends on losses being bounded and consistent, so that winners, worth more in rupees, can dominate the sum over many trades. Widening a stop or averaging down deliberately increases size and risk on the trades with the worst realised outcomes, which is precisely the wrong tail to add exposure to; it fattens the losses that the whole distribution was engineered to keep small. One trade with an uncapped loss can erase the gains of many disciplined ones and undo the arithmetic the other seven rules exist to protect. You scale into strength, never into a position the market is already telling you is wrong.

Rule 2 stands or falls on counting correlation honestly, so here it is made concrete. Three long positions are open, each sized by rule 1 to risk one percent of equity, which looks on the blotter like three percent of independent, diversified risk. But all three are midcaps in the same sector, driven by the same thing: a policy or demand shock to that sector moves them together, and their stops are the kind of levels that get taken out on the same red candle. The table sets the naive count beside the true combined risk.

The lesson is not that the arithmetic changes: three one-percent risks still sum to three percent. It is that the three percent is correlated, so it behaves like a single three-percent position rather than three independent one-percent ones, and it must be charged against the open-risk cap as one block. Genuine diversification would spread the same risk budget across unrelated drivers, so that no single shock can hit the whole allocation at once. The blotter's comfort of three separate line items is exactly the illusion rule 2 exists to strip away.

That upstream judgement, deciding which exposures share a driver and sizing the book so no one shock can breach the cap, is the part that separates a rule set on paper from one that holds under stress, and it is exactly what the method we teach is built around. The rules are the easy half to state; applying them when a fourth attractive name in the same sector appears is the discipline.

A governance manual can read as excessive caution until you look at what happens without one. The clearest current evidence in the Indian market is from the regulator itself. In its study of the equity derivatives segment published in July 2025, SEBI found that 91 percent of individual traders lost money in the equity derivatives segment in FY25, with aggregate net losses of ₹1,05,603 crore, a figure that had widened sharply on the year before. It is not a statement about any one strategy; it is what a large population of undergoverned, often over-leveraged risk-taking produces in aggregate. The rules on this page are the opposite posture: bound the size, bound the total, bound the gap, refuse the coin flip, obey the regime, and preserve the base.

Read as a whole, the eight rules are a single idea applied at every level: decide the loss before the trade, and let nothing in the heat of the moment enlarge it. Rule 1 bounds the single trade, rule 2 bounds the portfolio, rule 3 bounds the overnight gap, rule 4 refuses the un-edged bet, rule 5 bounds the whole book to the environment, rule 6 makes the process visible, rule 7 bounds the drawdown, and rule 8 forbids the one action that unbinds all the others. Plug the swing framework into this rule set and the method has somewhere safe to operate. The setups will keep changing; the governance is the part that keeps the account alive to trade them.