How Much to Risk Per Trade: The Case for 0.5% to 2%

Risk one percent per trade. The instruction has been repeated so often that it now functions as wallpaper, and the questions that ought to follow it are almost never asked: why one percent and not three, why not a quarter, and what exactly is the number defending you against? There is an answer, and it is arithmetic. Working through it takes perhaps a minute. After that the sensible range more or less selects itself.

Somebody outside markets usually puts the question better than anybody inside them does. What is the number protecting? Not the account, which is the answer most traders reach for first. The household. The mortgage, the holiday that is half booked, the general business of not having to explain anything awkward over dinner. Everything below is a formal way of saying that the correct fraction is whichever one lets the rest of your life carry on unaffected while the market does whatever it intends to do that month.

What a fixed percentage does

Fix the fraction and the sum at stake moves with the account: it shrinks while you are losing and expands while you are winning. That single property is what puts ruin out of mathematical reach, since every loss is taken as a share of whatever remains, and no finite run of losses can therefore reduce the balance to nothing.

There is a second benefit, quieter but useful. Results become comparable across time. A trade that lost 1% in a thin month and a trade that lost 1% in the best year you have had are errors of identical size, whatever the dollar figures happened to be on the day.

The losing-streak table

Potential profit does not set the percentage. An ordinary losing streak does. A strategy winning half of its trades will throw up ten consecutive losses about once in every thousand trades, which for anyone trading actively means the streak arrives inside the span of an ordinary trading career.

Risk per trade	After 10 straight losses	After 20 straight losses
0.5%	95% of the account	90%
1%	90%	82%
2%	82%	67%
5%	60%	36%
10%	35%	12%

Sit with the bottom two rows for a moment. At 5%, ten losses strip 40% from the account, and climbing back requires a 67% gain. At 10%, twenty losses leave 12% standing, and the recovery needed is 720%. Almost nobody returns from there. What should trouble you is that none of it requires a single bad decision beyond the size itself.

The top three rows describe accounts that survive. The bottom two describe accounts that do not. That gap, compressed into a sentence, is the whole reason professional consensus has settled between half a percent and two.

Choosing your place in the range

Tidiness would suggest one percent for everybody. Three things decide it.

The lower end of the range, around 0.5%, belongs to anyone whose record still runs to fewer than a couple of hundred trades, because the edge remains unproven at that stage and small size is simply what the information costs. It suits high-frequency approaches too, where streaks arrive quickly in calendar time. Large accounts belong there as well. Half a percent of a large balance is still a serious sum of money.

One percent is the default for a sound reason: it withstands any streak you can reasonably expect, and it still leaves a genuine edge enough room to compound at a respectable pace over the years you will be doing this.

Two percent asks more of you. It is defensible where the edge has been tested over a long sample and where the losses in the journal cluster tightly around the planned figure. Suppose they do not. A journal showing losses of 1.5R and 2R scattered through it is telling you that the trader already risks more than 2% on those trades without ever having chosen to, and in that case the percentage is the wrong dial to turn.

An illustration

Take one case. A trader moves up to 2% on the strength of a good few months, and the edge may well be real. Assume execution is not yet clean, though: stops get moved now and then, and a moved stop at 2% risk quietly becomes a 4% loss that nobody ever decided to take. Dropping to 1% would do nothing at all to that trader’s win rate. It would halve the cost of the errors. The errors are the part still outside their control.

That is the more useful way to read the range. It measures the margin you are leaving for the mistakes you have not yet stopped making, and ambition has very little to do with it.

Putting it in place

1. Choose a figure inside the range and write it somewhere permanent. One percent, if you cannot decide.
2. Recompute the dollar amount from the current balance once a month, so that a good week does not quietly enlarge every position you open afterwards.
3. Convert it to a share count on every trade with the position size calculator. Use the real stop distance.
4. Review the journal in R. Losses that consistently come in worse than −1R mean the execution needs attention before the percentage is worth revisiting at all.

Once the sample grows large enough to put error bars around your expectancy, the expectancy and sizing guide explains how to size from the edge itself. Until that point, the rule of thumb is doing exactly the job it was meant to do.