How to Set Risk Per Trade
To set risk per trade, decide the longest losing streak you expect to meet and work out what each candidate percentage would do to the account across it. Then pick the largest figure whose result you could sit through without changing anything about the plan the next morning.
This is one number, chosen once, applied to everything. It is also the number with the largest single effect on whether an account survives long enough for anything else to matter.
Before you start
Your account balance as of today, which is the base the percentage applies to. Not the peak, not the figure before the last drawdown — what is actually there.
An honest estimate of your win rate, or the knowledge that you do not have one. Without a record, assume you are at a coin flip and size accordingly.
A losing streak length you are prepared to sit through without changing anything. This is the input the whole decision turns on, and it has to be decided while nothing is going wrong.
The steps
1. Write down the streak you expect to meet
At a 55% hit rate over 100 trades, a run of 6 losses is more likely than not and a run of 8 has a 14.49% chance. Pick the number you will actually plan around.
2. Compute what each candidate percentage does across that streak
Six losses at 1% leaves 94.15% of the account. At 3% it leaves 83.30%. At 5% it leaves 73.51%. Those are the three outcomes you are choosing between.
3. Add the recovery requirement to each
94.15% needs 6.22% to get level. 83.30% needs 20.05%. 73.51% needs 36.04%. The third is a different category of problem from the first.
4. Pick the largest figure whose outcome leaves the plan unchanged
Not the one that maximises returns. The one where the streak’s outcome would leave you following the same plan the next morning.
5. Check it against the ruin arithmetic
With a win rate of 52% and even-money outcomes, risking 2% gives a 1.83% chance of ruin. At 1% it is 0.0334%. At 4% it is 13.52%. Doubling the size multiplied the danger sevenfold.
6. Apply it to correlated positions as a group
Three trades on the same theme each risking 1% is 3% on one idea. Budget the correlated group against the single-trade figure rather than each leg separately.
7. Write it down and stop revisiting it
A percentage raised after a losing run raises the probability of ruin at the exact moment the account is smallest. The figure is decided once, in calm conditions, and then it is a constraint rather than a preference.
How to tell it worked
You can state the figure without checking, and it has not changed in 90 days.
Multiplying it by your account gives the same money figure your last trade actually risked, within rounding.
A run of 4 losses produced no change to the percentage, which is the test the whole decision exists for.
And correlated positions were counted as one, so the open risk never exceeded the single-trade figure by more than you intended.
What the percentage cannot cover
Costs are paid whether the trade wins or loses. On this site’s shared series a round trip measures
about 2% of the median bar range of 0.493, so a small stop means the cost is a meaningful share of the
risk budget. The figures are in research/series-measurements.json.
A gap through the stop produces a larger loss than the percentage permitted. No sizing rule prevents that; only a smaller figure reduces it.
The figure most people arrive at
One percent is the common answer and it is not arbitrary. It survives a run of 10 losses with 90.44% of the account intact, needing 10.57% to recover — an ordinary stretch rather than a crisis.
Half a percent is the answer for anyone without a measured edge. It halves the ruin probability by far more than half, and it costs only the speed at which a real edge would compound.
Two percent is defensible with a measured edge and a tested process. Above that, the arithmetic starts requiring assumptions about your win rate that almost nobody has the sample to support.
And the honest default while learning is the smallest size your broker permits. The purpose of the first hundred trades is to produce a record, not a return, and a small size is what makes finishing them possible.
The original data
Of the 24,971 unique videos in research/search-study-corpus.jsonl, 55 have an instruction-shaped
title about risk management, at a median of 2,522 views across 54 channels. Position sizing appears
in 34 at 6,157 and win-rate targets in 161 at 13,711. The counts come from site/rank_howto.py, which
deduplicates by video id.
55 videos at a 2,522 median against 161 on win rates at 13,711. The subject that decides whether an account survives has a sixth of the audience of the subject that decides how it feels, which is a consistent finding across the whole risk category on this site.
The answer to the question on that chart is that increasing size after losses raises the probability of ruin at the worst moment. Going from 2% to 4% takes the ruin figure from 1.83% to 13.52% on the same edge, with a smaller balance and a sequence already running against you. The arithmetic says risk less after losses, not more — because ruin is measured from where you are rather than from where you started.
When it fails
The failure is a figure that drifts upward without a decision, and it drifts during good runs. A winning stretch makes the current size feel timid, one position is taken slightly larger, then another, and within a few months the account is running at double the risk it was designed for. Nobody raised the percentage; it simply stopped being applied. The next ordinary losing streak then produces a drawdown that the plan never contemplated.
The second failure is raising it after losses. That is the one move the arithmetic forbids.
A third is ignoring correlation. Three positions on one theme is one risk.
A fourth is sizing from a peak balance. The percentage applies to today.
A fifth is forgetting costs. They come out of the same budget.
And a sixth is choosing the figure from a return target. The return is a forecast; the streak is not.
Related
Risk per trade covers what the figure does in the wider framework. Position sizing is how it becomes a number of units. And drawdown is what the whole decision is trying to keep survivable.
The number that settled this for me was the gap between the 2% row and the 4% row on the ruin table. I had assumed doubling the risk roughly doubled the danger. On the same edge it multiplies it by about seven, and that asymmetry is the whole argument for trading small enough to be bored.
— Michael Whitman
This page is educational, not financial advice. Test every idea on your own charts before risking money.