Position Sizing When It Is Uncomfortable

The short answer

Position sizing is a division sum. The amount a mistake may cost, divided by the distance to the point where the idea is wrong, gives the size. There is nothing difficult in it and every trading platform will do it for you.

Which raises the real question: if the arithmetic is trivial, why is sizing the thing people most often get wrong?

Because the calculation is not where the decision happens. It happens in four specific situations, each of which pushes in the same direction, and none of which is addressed by knowing the formula. This article is about those four.

The calculation, stated once

Using hypothetical figures throughout — none of this is a recommendation:

Suppose a trader has fixed, in advance, that a single mistake may cost £150, and the level at which this idea is wrong lies 30 points from the intended entry.

£150 ÷ 30 points = £5 per point.

That is it. Where the amount in the numerator should come from is the subject of how much a mistake is allowed to cost, and it is a question about your circumstances rather than about markets.

Notice the property that makes this formula useful and unpopular: when the invalidation level is far away, the size comes out small. The formula is doing its job. Overriding it because the position then feels "too small to bother with" is not a refinement — it is abandoning the method while keeping the vocabulary.

The four situations

One: the size is set after the entry decision

This is the commonest and the quietest. The sequence inverts: a size is chosen — often the one used last time, or the one that makes the potential gain feel worthwhile — and the stop is then placed wherever that size permits.

The arithmetic is identical. What changes is which quantity is the output. Done in the correct order, the cost of being wrong is a decision. Done in reverse, it is a residue.

A record will show you which one you did, because the two orders leave different traces: in the correct order the size varies trade to trade and the loss amount does not. If your position size is near-constant and your losses vary widely, you have been sizing backwards. That is one of the things a usable record is for.

Two: the position is smaller than the conviction

You size correctly, the trade starts working, and the position now feels too small for how right you were. The pressure to add is strongest exactly when the evidence for adding is weakest — a position moving in your favour is not new information about the idea, it is the same idea with a worse entry available.

There is no study we can point to that measures this in retail accounts. What we can say is what the adjacent evidence supports: the one well-documented association between size and outcome, in the French regulator's study of 14,799 retail forex and CFD clients, runs the wrong way for anyone adding on conviction. Clients whose average transaction size exceeded €10,000, 62 per cent of the population, lost about €14,876 on average against a population-wide average of about €10,900, and losses deepened through the larger size bands.

Three: the previous trade lost and this one looks obvious

This is the situation with the most precise evidence behind it, and it is not the one people expect. Alex Imas showed that the contradictory findings on risk-taking after losses resolve once realised losses are separated from paper ones: after a loss that had been booked, people took less risk; where the same loss was left unrealised, they took more.

The implication for sizing is specific. The dangerous moment is not after you have closed a loser and feel the sting. It is while the loser is still open and the decision to size the next position is being made alongside it. The full argument is in what happens in the hour after a loss.

Four: the platform allows far more than the plan

Leverage limits are frequently read as a sizing guide. They are not. ESMA's product intervention measures set initial margin requirements at 30:1 for major currency pairs, down through 20:1, 10:1 and 5:1 by instrument class, to 2:1 for cryptocurrencies, with a margin close-out at 50 per cent of minimum required margin and negative balance protection per account.

Every one of those numbers is a ceiling imposed on a firm, arrived at because national regulators found that 74 to 89 per cent of retail investor accounts typically lose money on these products. Sizing up to the limit because the limit exists is using a supervisor's view of the worst acceptable case as a plan.

What does not help

Formulae that promise an optimal size. Any rule of that kind requires inputs — your true win rate, your true average win and loss — which you do not have. You have estimates from a small sample of your own trades, and plugging a noisy estimate into a formula that is sensitive to it produces confident nonsense. If you have fewer than a few hundred recorded trades, you are estimating, and the three ways people misjudge their own accuracy is a useful corrective on how confident those estimates deserve to be.

Resolving to be disciplined. The one study that looked for a trader personality found none, and nothing in the literature suggests resolve is the operative variable. A rule you can be seen to have broken is worth more than a resolution you cannot.

Sizing by feel on the grounds that you have experience. Experience improves many things. It does not give you access to a probability you have never recorded.

What does

Three things, all of them boring, and all of them checkable afterwards:

1. Write the loss amount down before the session, not during it. It is a number, it does not change intraday, and if it does change you have a record of when.
2. Let the size be computed, every time, by the same division. Including when it comes out awkwardly small. Especially then.
3. Record which trades departed from it, and why. Not to punish yourself — to find out whether the departures cluster. In our experience of the evidence, the useful question is never "am I disciplined?" but "under which conditions am I not?", and that has an answer only if you wrote things down.

The rules that make any of this hold under pressure have to exist before the pressure does, which is the whole subject of what a plan has to decide in advance. The broader behavioural picture these failures sit inside is set out in our guide to what has been measured about trader behaviour.

Trading leveraged products carries a high risk of losing money quickly. Nothing here is advice, and no figure above should be used as one.