Risk Management First: Position Sizing, Planned Exits and Risk of Ruin

Almost all of the public conversation about markets is about selection: which asset, which narrative, which entry. Almost none of it is about the two decisions that actually determine whether an account is still functioning a year later, namely how much you commit to any single idea and what makes you leave. This article is about that second category. It is educational, it is not advice, and it deliberately contains no recommendations to buy or sell anything.

Why sizing does the work

Consider two people with identical views who buy the same thing at the same moment. One commits a small fraction of their capital. The other commits most of it. Their analysis was the same; their outcomes will not be. If the idea is wrong, the first person absorbs an annoyance and continues, while the second is now in a situation where every subsequent decision is made under pressure.

That is the whole argument for treating sizing as the primary variable. Selection determines whether you are right. Sizing determines whether being wrong matters. And since nobody gets to know in advance which category any particular idea falls into, the only part of the process you fully control is the size of the consequence.

The practical version is to define risk per idea as a small, fixed share of total capital, chosen so that a run of consecutive losses is unpleasant rather than structural. What counts as small depends on your circumstances, your timeframe and how much volatility you can genuinely tolerate rather than how much you imagine you can. Crypto markets tend to move considerably more than traditional ones, which means an equivalent percentage position carries a much wider range of outcomes here.

Working backwards from loss

The arithmetic runs in the opposite direction to how most people think about it. You do not decide how much you want to own and then wonder what could go wrong. You decide the maximum you are prepared to lose on the idea, identify the point at which you would accept the idea had failed, and let the distance between those two things determine the size of the position. A wider invalidation point means a smaller position, not a larger loss.

This has an unintuitive consequence worth sitting with: strong conviction should not increase position size, because conviction is not evidence. It is a feeling, and it is reliably at its most intense just before it becomes most expensive.

Deciding your exit before you enter

Before an entry, you are a reasonably objective observer of a situation. Afterwards, you are a participant with money at stake and an emotional interest in a particular outcome. Anyone who has held a losing position knows how quickly the reasoning changes: the timeframe extends, the thesis broadens, and the original invalidation point is quietly reinterpreted as noise. The purpose of deciding an exit in advance is not that your future self is stupid. It is that your future self is compromised.

A usable plan answers three questions before any capital moves. What specific observation would tell me this idea was wrong? What would tell me it has worked? And under what conditions would I leave for reasons that have nothing to do with price, such as a change in the underlying facts or in my own circumstances? Written down, these are commitments. Held in your head, they are intentions, and intentions are renegotiated under stress.

Exits get skipped because they force you to acknowledge the position might fail, which is uncomfortable at exactly the moment you feel optimistic. That discomfort is the point. An idea you cannot describe a failure condition for is not an idea, it is a hope.

Risk of ruin, and why it beats win rate

Risk of ruin is the probability that a sequence of losses reduces your capital to a level from which recovery is no longer realistic. The important word is sequence. Averages describe what happens across many outcomes; ruin is about the order they arrive in, and an outcome you cannot come back from ends the process no matter how favourable the long-run average would have been.

The asymmetry between losses and recoveries is what makes this so unforgiving. Losing a given percentage always requires a larger percentage gain to get back to level, and the gap widens sharply as the loss deepens. A severe drawdown requires a gain most people would never expect from a single idea, which is why deep drawdowns tend to end participation rather than being traded out of.

This is also why win rate is such a poor summary of skill. A high win rate tells you how often outcomes were positive, not how large they were. Being right frequently while occasionally taking a very large loss can leave you worse off than being wrong frequently with small, contained losses and a few substantial gains. Worse, win rate is trivially inflated by refusing to close losing positions, which improves the statistic while increasing risk of ruin.

Things that quietly increase risk

• Correlation you have not counted. Holding several crypto assets often means holding one position expressed several ways, since they tend to move together when it matters most.
• Leverage. Borrowed exposure does not just amplify outcomes; it introduces a mechanism by which a temporary move can close your position permanently, at the worst available price.
• Adding to losers. Increasing size as a position moves against you raises exposure precisely as the evidence for the idea weakens.
• Risking money with a job. Capital needed for rent, tax or a business obligation forces exits on someone else’s schedule.
• Position creep. Sizes drift upward after a good run, so the largest positions tend to arrive just as conditions turn.

The mining version of the same problem

Mining is an instructive case because the risk structure is unusually visible. Buying hardware is a capital commitment, and running it carries an ongoing electricity and hosting cost billed in fiat currency while revenue arrives in the mined asset. That is a position with a running expense attached, and it is why miners are structural sellers: coins must regularly be converted to pay bills, whatever the market is doing.

The same three questions apply. Sizing means not committing so much capital to equipment that a difficult stretch forces you to sell hardware into a weak market. An exit plan means deciding in advance the conditions under which a machine stops being worth running, rather than discovering it during a bad month. And ruin looks like being unable to cover power costs long enough to reach better conditions. The variables involved are visible on our mining dashboard, and you can model your own inputs with the mining profitability calculator. Every output there is an estimate that depends on assumptions you supply, especially those about future difficulty and energy prices.

What survivable practice tends to look like

The common thread in approaches that last is unglamorous. Risk per idea is small enough to be boring. Exits are written down before entry, not decided during. Position size follows from the distance to invalidation rather than from enthusiasm. Correlated exposures are counted as one. Leverage is treated as a different activity, not a bigger version of the same one. And decisions are reviewed on process rather than on outcome, because a good decision can lose and a reckless one can win.

None of this improves your ability to pick anything. It improves the chance that you are still around when a good decision eventually pays. Definitions for the terms used here are in our glossary, foundational material sits in the learn section and the beginner trading guides, and how we produce and source figures is set out in our methodology. Nothing here is financial advice or a recommendation to take any particular position.