How Much to Risk Per Trade, and Why That Number

The short answer

Nobody can tell you how much to risk per trade, and this publication is not going to pretend otherwise. What we can tell you is that the familiar answers — one per cent, two per cent — have no source we have been able to find. They are conventions. They may well be sensible conventions, but a number repeated confidently for thirty years without a derivation is not evidence, and you should know that before you adopt one.

What is established is the structure of the decision, and it is more useful than the number:

1. The amount a mistake is allowed to cost is decided first, in advance, as a fixed figure.
2. The point at which the idea is wrong is decided second, from the market and not from the size you want.
3. The position size is then whatever those two imply. It is an output, not an input.

Reverse that order — pick the size first, then put the stop wherever it fits — and the cost of being wrong becomes whatever happens to be left over. That reversal is the single most common sizing failure, and unlike a feeling, it is visible in a record.

Why we will not quote you a percentage

We looked for the origin of the one per cent rule. We did not find a study, a regulatory document, or a published derivation. The figure appears in trading education material, usually without attribution, and the versions differ: one per cent, two per cent, "never more than you can afford to lose", sometimes with a scaling rule attached and sometimes not.

Two per cent of what, as well, is rarely specified. Of the account? Of liquid net worth? Of the capital you consider at risk in this activity as a whole? Those produce wildly different numbers for the same person, and the ambiguity is doing a lot of quiet work in the advice.

So this article will not supply a figure. We apply the same rule here as in our investigation of the statistics this field repeats: if we cannot read it in a document, we do not publish it as a fact.

What the regulators' numbers do and do not tell you

European rules do set hard limits on retail leveraged positions, and these are often quoted as if they were guidance about risk. They are not. They are ceilings.

When ESMA agreed its product intervention measures in March 2018, it set initial margin requirements by instrument class — 30:1 for major currency pairs, 20:1 for non-major currency pairs, gold and major indices, 10:1 for commodities other than gold and non-major equity indices, 5:1 for individual equities and other reference values, and 2:1 for cryptocurrencies. It added a margin close-out rule at 50 per cent of the minimum required margin, and negative balance protection on a per-account basis as an overall guaranteed limit on retail client losses.

Three things follow.

A ceiling is not a recommendation. The maximum a regulator will permit is the point at which a supervisor judged the product unacceptable, not a suggestion about what to use.

Negative balance protection bounds the account, not the plan. It stops you owing more than you deposited. It does nothing about the deposit.

The close-out rule is the last line, not a risk policy. If the broker's automatic close-out is what ends a position, every decision you were supposed to make has already been skipped.

The regulators also published the figures those limits were a response to: national authorities' analysis showed 74 to 89 per cent of retail investor accounts typically lose money on CFDs, with average losses per client ranging from €1,600 to €29,000.

The one piece of evidence about size

There is a little direct evidence connecting position size to outcomes, and it comes from the French regulator's four-year study of 14,799 retail forex and CFD clients.

The AMF reported results broken down by average transaction size, taking leverage into account. Across the whole population the average result was about minus €10,900. Among the 62 per cent of clients whose average transaction size exceeded €10,000, the average result was about minus €14,876, and the losses continued to deepen through the larger size bands. The same study found that clients who placed at least 250 orders over the four years — 52 per cent of the population — lost an average of about €18,741.

That is correlational, from one country, over 2009 to 2012, and it cannot separate size from whatever else distinguishes a client who trades larger. It is, nevertheless, the only measurement we have found that relates size to retail outcomes at all, and it does not point in a comfortable direction.

How the number is actually derived

Here is the arithmetic, with hypothetical figures used throughout. Nothing below is a recommendation, and every number is invented for the purpose of showing the structure.

Suppose a hypothetical trader has decided, in advance and in writing, that a single mistake may cost £200. Suppose the level at which the trade idea is wrong sits 40 points away from where the position would be opened. The size is then simply:

£200 ÷ 40 points = £5 per point.

That is the whole calculation. Everything difficult about it happens before and after.

Before, the £200 has to come from somewhere defensible. It is not a percentage someone quoted; it is an amount you have decided you can lose repeatedly without it changing your behaviour, your finances or your sleep. If losing it four times in a week would make you trade differently, it is the wrong number, and only you have the information needed to set it.

After, the number has to survive. The failure mode is not arithmetic. It is that the stop gets moved, or the size gets increased because the signal looks unusually good, or the figure quietly drifts upward over a profitable month. Those are the subject of sizing when it is uncomfortable.

Note what the formula does when the invalidation level is far away: the size gets small. That is the correct behaviour and it is the part people override. A wide stop with a normal size is not a wide stop — it is a bigger bet.

The questions that actually decide your number

Since no source can give you a figure, these are the questions it has to come out of. They are not rhetorical; each one has an answer that is specific to you and that nobody writing on the internet has access to.

1. How many consecutive losses is your method capable of producing? Not how many you expect — how many it has produced, in records you kept. Most people have no idea, which is an argument for keeping a usable record.
2. What size of loss changes your behaviour? This is the binding constraint and it is psychological rather than financial. A loss that makes you abandon your own rules has cost more than its money.
3. Is this money you can lose entirely without consequence? Every figure in the regulators' studies above was a real person's deposit.
4. Does the number survive a bad run on paper? Work it through for the worst sequence you can document, before the sequence happens rather than during it.

What this publication will say

Risk per trade should be a fixed, written amount, decided before you look at a chart, derived from your own circumstances, and used to produce position size rather than the other way round. The widely quoted percentages are conventions without a published basis; use one if you like, but use it knowing that is what it is.

And the limits the regulators impose are the outer boundary of what a supervisor will allow a firm to sell you, which is a very long way from a view about what is sensible. The wider behavioural context for all of this is in our overview of trader behaviour and the evidence behind it.

Trading leveraged products carries a high risk of losing money quickly. Nothing on this page is advice or a recommendation, and no figure in it should be copied.