Risk Management on a Frontier Market: Position Sizing and Stop-Losses That Survive the PSX

Every risk-management guide repeats the same two rules: never risk more than 1–2% of your portfolio on a single trade, and always use a stop-loss. The first rule is genuinely good advice. The second was written for markets where you can actually sell — deep order books, no daily price limits, stop orders that fill somewhere near their trigger. The PSX is not that market, and pretending otherwise is how careful-sounding traders take careless-sized losses.

This guide keeps what travels well, and rebuilds what does not.

Risk per trade: the arithmetic almost nobody does

The 1–2% convention says: on any single trade, the amount you stand to lose if your exit is hit should be 1–2% of your total portfolio. Not the position size — the loss.

Most retail investors have never run the calculation. It takes three steps:

1. Decide the rupees at risk. Portfolio of Rs 1,000,000, risking 1%, means Rs 10,000 on this idea.
2. Decide the price that proves you wrong. You buy at Rs 200 because you expect the uptrend to hold; a close below Rs 180 would mean it has not. That is Rs 20 of risk per share.
3. Divide. Rs 10,000 ÷ Rs 20 = 500 shares. Position size: Rs 100,000.

Notice what happened: the position size was an output. You did not decide to put Rs 100,000 in — the distance to your exit decided it for you. Tighten the exit to Rs 190 and the same Rs 10,000 of risk buys 1,000 shares. Widen it to Rs 160 and you may only buy 250. A wide stop is not reckless and a tight stop is not prudent; what is reckless is choosing the position size first and discovering the risk later, which is how nearly everyone actually does it.

Why 1%? Because you will have losing streaks, and the question is whether you survive them. Ten consecutive 1% losses leave you at 90% of your capital — annoying. Ten consecutive 10% losses leave you at 35% — finished, for reasons the drawdown table below makes brutal.

A 1:3 risk-to-reward label does not make the 3 likely

The other staple of trading content is the risk-to-reward ratio: risk Rs 20 to make Rs 60, and you "only need to be right a third of the time".

True as arithmetic, useless as stated. Anyone can draw a target three times further away than their stop — the chart does not object, and the market was not consulted. The probability of reaching that target is not printed anywhere, and it is usually far lower for distant targets, which is precisely why they are distant. A 1:3 trade that works 20% of the time loses money steadily: 0.2 × 3R − 0.8 × 1R is a negative number wearing a professional-looking ratio.

The honest use of risk-to-reward is as a filter, not a forecast. It ensures that when you are right, you are paid enough to fund the times you are wrong. It says nothing about how often you will be right — and no published win rate, including our own model's record on the model performance page, is stable enough to make that expected-value arithmetic precise. Assume your hit rate is worse than your backtest, because it will be.

Why stops are different on the PSX

Here is the part the imported textbooks skip. A stop-loss is an instruction to your broker, not a contract with the market. Three features of the PSX regularly break it.

Circuit breakers: sometimes there is no exit at any price

PSX scrips trade within daily price limits — circuit breakers that cap how far a stock may move in one session. For most of the market's history the band was 5% either side (or Rs 1, whichever is higher); the exchange has adjusted the bands over the years, so verify the current limits in PSX regulations rather than assuming a number. There is also a market-wide halt mechanism triggered by large index moves.

Now run the scenario. Bad news lands overnight — a shock result, a regulatory action, a gas-tariff decision against a fertiliser name. The stock opens with sellers stacked at the lower limit and no buyers at all. It is "locked at lower circuit": a price is displayed, but no trades are happening, because a trade requires someone on the other side. Your stop at Rs 180 is a theoretical object. The stock sits limit-down at Rs 176 all day, opens limit-down again tomorrow, and perhaps the day after. Mid-cap stocks going through multi-day lower-circuit sequences after bad news is not a tail event on this market; it is routine. By the time a buyer appears, the price may be 15% below the level your risk calculation assumed.

Gaps: a stop is a trigger, not a floor

Even without a locked circuit, prices jump. Your stock closes at Rs 185; overnight the rupee slides or a policy decision lands; it opens at Rs 168. A stop at Rs 180 does not sell at Rs 180 — it becomes an order when the price is already through the level, and fills at whatever the market offers next. You budgeted a 1% loss and took 2%.

Thin books: the quoted price is for someone smaller than you

On many PSX listings the visible order book holds a few hundred shares per price level. A market order of any size walks down through those levels, and your average fill can land rupees below the quote — the same exit-risk problem covered in how to read a stock page. Illiquidity converts your stop from a line into a region.

The conclusion is the most important sentence in this guide: on the PSX, position size is the real risk control, and the stop-loss is only an intention. The stop defines the loss you planned; the position size caps the loss you actually take when the plan meets a locked circuit. Size every position as if the exit will slip badly — because on the trades where it matters most, it will.

Mental stops, hard stops, and the thin-market caveat

A hard stop is an order resting with your broker; a mental stop is a level you promise yourself you will act on. Textbooks say hard stops, because mental stops fail in a predictable way: the level arrives, and you renegotiate. It's oversold. It's manipulation. I'll average down. That renegotiation is catalogued in the mistakes retail investors keep making, and it turns a planned 10% loss into an unplanned 45% one.

But the PSX adds two caveats. First, the mechanics: not every retail trading terminal here supports true resting stop orders, so a "hard stop" may in practice mean a price alert plus a manual order — ask your broker what actually happens rather than assuming. Second, the thin-market problem: in an illiquid stock, a resting stop is an offer to buy liquidity at the worst possible moment. One modest sell order can print through your level in a shallow book, fill your stop at the bottom of a spike, and reverse within the hour. In liquid names — OGDC, HBL, LUCK — a hard stop behaves roughly as advertised. In thin ones, the honest hierarchy is: first own less of it, then use a written mental stop with a rule you have committed to in advance, and treat a hard stop as the tool most likely to be harvested.

The arithmetic of losing

Losses and gains are not symmetric, and the asymmetry gets worse the deeper you go.

A 10% loss is a bad month. A 50% loss requires a doubling — something the broad market can take years to deliver. The KSE-100 peaked near 53,000 in May 2017 and did not reclaim that level until late 2023: over six years for buy-and-hold to break even in nominal terms, and far longer in real terms once you account for inflation and the rupee. Every sizing rule above exists to keep your individual mistakes in the top rows of this table, where recovery is a task rather than a miracle.

You own fewer bets than you think

Risk per trade is half the job. The other half is noticing that separate positions can be the same position.

Own HBL, UBL and MEBL and you hold three tickers and one bet: the policy rate, which peaked at 22% in 2023–24 and rewarded bank margins on the way up exactly as cuts squeeze them on the way down. LUCK plus MLCF is one bet on the construction cycle, coal prices and the rupee. FFC plus EFERT is substantially one bet on gas pricing decisions. The sectors page makes these groupings visible, and comparing two same-sector holdings usually reveals how little diversification they add to each other.

Then there is the layer above sectors. On the days that define your year — a rupee slide, an IMF programme stalling or resuming, a political shock — correlations across the whole exchange converge towards one, and everything falls together. The macro cycle that drives this is worth understanding in its own right (market cycles and the IMF), but the risk lesson is blunt: diversification within the PSX has a ceiling, because the index itself is a concentrated position on Pakistan's macro story. Spreading across twenty tickers diversifies your company-specific risk and almost none of that.

Cash is a position

The final tool is the one retail culture treats as failure. Holding cash is not "being out of the market" — it is an allocation to the one asset that cannot lock at lower circuit.

For long stretches of recent history it even paid well: with the policy rate at 22%, money-market funds and T-bills offered returns most equity portfolios failed to beat that year. Rates have come down since; what cash currently yields is a question for current fund sheets, not an article. But the deeper argument is not the yield. Cash is the only asset that lets you act during the week everyone else is a forced seller — when good companies are locked limit-down beside bad ones and the order books are one-sided. The investors who buy those weeks are not braver than you. They simply still had something to buy with, because they sized every earlier position as if the exit might fail.

That is the whole discipline, compressed: assume the stop will slip, size so it does not matter, group your bets honestly, and keep enough cash that other people's forced selling becomes your opportunity instead of your company.