Position sizing and the risk of ruin

Suppose you have made peace with everything the last twenty-four modules said and you still intend to use leverage — for a hedge, or with clear eyes and a small stake. Fair enough. This module is the one thing you must get right if you do, because it is the only control that is genuinely in your hands.

You cannot control whether any single trade works; the market decides that. You cannot control the news, the gap, the volatility. There is exactly one dial you turn yourself, before the outcome is known and while you are still calm: how much you put on the line. Position size is not one factor among many. For a leveraged trader it is the master control — the difference between a bad run that stings and a bad run that ends you.

And "ends you" is meant literally, because of one hard fact this shelf keeps returning to: some losses cannot be recovered from. This module is about that fact — the risk of ruin — and about the only defence against it, which is size.

Most losses are survivable. Down 10%, down 20%, even down 40% — painful, but you are still in the game, still able to earn it back. There is a level, though, past which recovery stops being arithmetic and starts being impossible, and that level has a name.

The risk of ruin is the probability that your losses reach a point from which you cannot come back — usually a wiped-out or margin-called account. What makes ruin different from an ordinary loss is that it is an absorbing state: a condition that, once entered, you can never leave. A ball rolling on a table can go anywhere; a ball that falls off the edge is simply gone. Zero capital is the edge of the table. Every other loss is a position on the table you can still play from; ruin is off it.

Two things follow, and they are the spine of this entire module. First, because ruin is permanent, its cost is not "a large loss" — it is every gain you would ever have made afterwards, forever. Second, and this is the part that undoes people: you do not need to be wrong on average to reach it. A trader with a genuine, positive edge — one who would make money over a thousand trades — can still be wiped out in the first twenty if each bet is too large, because a normal run of bad luck arrives before the average has time to show up. The edge only pays those still standing to collect it. Survival is not one goal among several. It is the precondition for every other goal.

Here is the machinery in plain numbers. Losing streaks are not rare; they are guaranteed. Flip a fair coin and a run of five tails will show up often across a few hundred flips. Trading is no kinder. So the real question is never "will I hit a losing streak?" — you will — but "when the ordinary streak arrives, does it end me?"

That answer is set entirely by how much you risk per bet. Risk a small slice and even a long cold run only dents you; risk a large slice and a routine streak reaches the edge of the table. The table below makes the difference concrete: the same run of five straight losses, at three different bet sizes. illustrative

Read the last column slowly, because it is the trap inside the trap. Losses and the gains needed to undo them are not symmetric. Lose 50% and you need +100% to break even; lose 83% and you need nearly +490%. The deeper the hole, the steeper the climb out — so large bet sizes do not just risk a bigger loss, they push you into a region where recovery is mathematically brutal even if you never quite hit zero. The 30% bettor who survives the streak is not really back in the game; he is facing a near-impossible climb.

Now put risk-of-ruin on a curve against bet size, and the shape tells the whole story: below some level ruin is remote; past it, ruin becomes near-certain given enough time.

The lesson of the curve is not "find the exact threshold." Nobody knows their true edge precisely enough for that. The lesson is which side of it to stand on, with room to spare — because the cost of being a little too cautious is a slightly smaller gain, and the cost of being a little too bold is the absorbing state.

Watch the two costs weigh against each other. illustrative

A trader has a ₹5,00,000 account and a genuinely favourable setup — say it works 55% of the time with even payoffs. A real, if slim, edge. The question is only: how much per trade? The answer is not "as much as I dare" and not "everything, I'm confident." It is the largest size that no realistic losing streak can end.

There is even a formula for the theoretical maximum, the Kelly criterion: it takes your edge and returns the bet size that grows capital fastest over the long run. For a 55%/45% even-money edge it suggests risking about 10% of capital per bet. And here is the part professionals actually live by: almost everyone bets a fraction of even that. Half-Kelly, quarter-Kelly. Why deliberately bet less than the growth-maximising size? Because the formula assumes you know your edge exactly, and you never do. Overestimate your edge even slightly and full Kelly tips you past the threshold into the steep part of the ruin curve. Betting a fraction buys a wide margin against the one error — overconfidence — that this whole shelf has shown you are wired to make.

So the sound read is: this trader, edge and all, risks perhaps 1–2% of the account per trade, not 10% and certainly not 80%. The size is set by the edge, discounted hard for the fact that the edge is a guess — never by how good the trade feels, never by the recent run, never by the size of the opportunity. The confident single trade that risks ₹4 lakh of a ₹5 lakh account is not brave; it is a coin-flip with your survival as the stake.

Position sizing is powerful, but it is a defence, not a cure, and it is honest to say what it cannot do.

It cannot turn a losing edge into a winning one. If your trades are negative-sum after costs — which, the SEBI data in the last part showed, is the norm for retail F&O — then careful sizing only decides how slowly you lose, not whether. Sizing keeps a real edge alive; it cannot manufacture one. Reducing bet size on a strategy with no edge just buys you a longer, gentler decline toward the same place.

It cannot protect against a gap through your stop. Sizing assumes your loss on a trade is roughly what you planned. But leverage can gap past your exit overnight, and a position sized "to risk 2%" can lose far more when the market opens 20% against you with no chance to get out. Real bet size must be figured on the worst plausible move, not the one you hoped to cap at.

It cannot survive being abandoned in the moment. A sizing rule only works if you keep it precisely when it is hardest — after a win, when you feel invincible, and after a loss, when you want it all back. The rule must be written down, fixed in advance, and obeyed against your own feelings. A sizing plan you override on your best or worst day is not a plan; it is a preference.

The rule is simple; keeping it is not. These are the classic ways it fails.

1. Sizing to the opportunity, not the account. "This is a once-in-a-year setup, so I'll go big." The rarer the chance feels, the larger the bet — and the larger the bet, the closer to the edge of the table. A great idea at a fatal size is still a fatal size.

2. Sizing to the recent streak. Adding after wins because you are "hot", or after losses to "win it back". Both point the biggest bet straight at a coin-flip. The account, not the streak, sets the size.

3. Measuring size by margin, not by loss. Leverage lets a small margin control a huge notional, so a position that "only cost ₹40,000 of margin" can lose several times that. Size is about how much you can lose, not how much you put up.

4. Ignoring correlation. Five "small" positions that all fall together in a crash are one large position in disguise. When trouble comes, things that seemed independent move as one, and your true bet size is the sum.

5. Confusing confidence with edge. Feeling sure is not an edge; it is the feeling the market is best at manufacturing just before it takes your money. Size by the odds you can defend on paper, not by the certainty you feel in your chest.