Part 3: Time, stop and size · Chapter 9 of 14 · 8 min
Size by volatility
With a 1.5 ATR stop the dollar risk per unit varies across the plan's 79 trades from $18.2 to $141.6 (median $28.4, ratio 7.8). With a fixed 1% risk per trade the worst loss is −1.02% and the drawdown 5.9%; with fixed size (1% at the median stop) the account ends higher (204% against 176%) but the risk per trade runs from 0.64% to 4.99%, the worst loss is −3.57% and the drawdown 6.6%. The advantage of fixed size, here, is sample luck.
The formula
Book 11 explained how to size a position: you decide how much of the account you are willing to lose if the stop is hit (for example 1%) and derive the size from the stop distance:
> size = risk in dollars ÷ stop distance in dollars
With the stop measured in ATR the distance is k × ATR. With a $10,000 account, 1% risk ($100) and a 1.5 ATR stop:
| Hourly ATR | 1.5 ATR stop | Size (ounces of gold, or equivalent units) |
|---|---|---|
| $13.2 (10th percentile) | $19.8 | 5.05 |
| $18.6 (median) | $27.9 | 3.58 |
| $33.0 (90th percentile) | $49.5 | 2.02 |
The size adapts to volatility: with ATR at $33 you buy or sell about 56% of what you buy with ATR at $18.6. That is why a plan that «risks 1%» has trades that differ in dollars, and why the size must be recalculated at every entry, not once.
How much the dollar risk varies
Across the 79 trades of our example plan, the dollar risk per unit (that is the stop distance, 1.5 ATR) is:
| Measure | Value |
|---|---|
| Minimum | $18.2 |
| 25th percentile | $25.2 |
| Median | $28.4 |
| 75th percentile | $35.5 |
| Maximum | $141.6 |
The trade with the widest stop has a distance 7.8 times that of the one with the tightest stop. Whoever always opened the same quantity would therefore risk 7.8 times more dollars on one than on the other.
Fixed risk or fixed size: the same plan, two ways
We applied two rules to the same 79 trades:
- Fixed risk: at each trade you risk 1% of the account, and the size is recalculated on the stop (as in book 11).
- Fixed size: the quantity is always the one that risks 1% with the median stop ($28.4). When the stop is wider, the loss in dollars is larger.
| Fixed risk 1% | Fixed size | |
|---|---|---|
| Final account (starting from 100%) | 176% | 204% |
| Maximum drawdown | 5.9% | 6.6% |
| Worst trade | −1.02% | −3.57% |
| Deviation per trade | 1.94% | 2.83% |
| Actual risk per trade | 1% (about) | from 0.64% to 4.99% |
Read the chart in words
Daily gold chart with, below, two account curves: a pink one with fixed risk and a blue one with fixed size, higher and more irregular.
Fixed size ended 28 points higher. It would seem the better rule. But you have to read why.
Why fixed size won (here)
Fixed size weighs trades with a wide stop more. In our sample those trades did well: the ten with the widest stop have a mean R of +0.60 and are worth $605 per unit out of $2,113 in total (29%). Splitting the 79 trades into thirds by stop width:
| Stop | Trades | Mean R ± error |
|---|---|---|
| Small | 27 | +1.00 ± 0.75 |
| Medium | 26 | +0.33 ± 0.69 |
| Wide | 26 | +0.87 ± 0.77 |
The errors are large (±0.7) and there is no order: the middle third is the worst. Fixed size won because by chance the wide trades were good, not because high volatility pays more. Had the wide trades gone badly, fixed size would have lost far more than the recalculated size.
That is why the comparison on result does not decide. On risk, instead, the difference is clear and does not depend on luck:
- the risk per trade with fixed size goes from 0.64% to 4.99%, almost eight times between one and the other;
- the worst loss is 3.5 times larger (−3.57% against −1.02%);
- the deviation per trade is 46% higher.
Size recalculated on the stop does not improve the average gain in a proven way: it makes the risk what you decided.
And if you reduce size when volatility is high?
A popular rule: halve it when the ATR is above the 80th percentile of the last 500 candles. On the 79 trades it triggers 18 times. Result: final account 155% (against 176%), drawdown unchanged (5.9%). The 18 trades with high ATR have a mean R of +1.44 ± 0.93, those without +0.53 ± 0.47. Here reducing size in the most agitated moments took away gain, because by chance those were the best moments. With 18 trades you cannot say that a high ATR pays more; you can say that this sample does not justify the rule. The question stays open and we take it up in chapter 10.
What to do
- Calculate the size at every entry from risk in dollars and stop in ATR.
- Write the risk per trade in the plan (the share of the account) and not the quantity.
- Do not judge the size rule by the gain of a 79-trade sample: look at the risk per trade and the worst case.
- Do not reduce size in agitated moments on an impression: test the rule and declare how many trials you made.
- Know your worst case: with fixed size the worst loss here was 3.5 times that with fixed risk.
Procedure in four steps
- Decide the share of the account to risk (for example 1%).
- At each entry calculate the stop in ATR and in dollars.
- Derive the size: risk in dollars ÷ stop in dollars.
- Check every month the worst loss and the share of risk per trade.
Exercises
Chapter exercises
1. Account $10,000, 1% risk and a stop of $25 per unit. How many units do you open?
2. You always open the same quantity and the ATR doubles from one trade to the next. What happens to the dollar risk?
3. Over the 79 trades fixed size ends at 204% and fixed risk at 176%. Which reading is correct?
4. With fixed size the worst trade is −3.57% and with fixed risk −1.02%. What does it imply?
Have a question or want to share your exercise?
Post in the community, or join the free signals room on Telegram.
Educational content, not financial advice. Trading involves risk.