Drawdown and risk of ruin
Measure a peak-to-trough loss, then see how a string of planned losses shrinks the account. · 11 min
Hook · The hole and the climb back
The percent down and the percent back are not twins.
Equity peaks at 100 and later prints 70 before it recovers. Enter the drawdown in percent.
30. The peak is the reference. A later new high ends that drawdown. Getting back from 70 to 100 is a 43% gain on the trough, not a 30% gain.
Squeeze · A string of planned losses is a path
A drawdown is peak to trough. It is not the distance from the day you opened the account, unless that day was the peak. Recovery math is asymmetric: lose half and you need a double to return. Risk of ruin, here, is the chance of hitting a floor if you keep risking a fixed fraction. Lose 10% of what is left, ten times, and the account is about 35% of the start, because each loss hits a smaller base: 0.9 to the tenth power. Risk 2% instead and ten losses leave about 82%. The backtest lesson already warned that a pretty curve can be a lie. This lesson reads a curve that is honest and still painful. Correlation, from the portfolio lesson, makes the losses arrive together, which deepens the trough without changing the per-trade formula. A plan that cannot survive its own losing streak is a plan that only works on the highlight reel. None of these percents is a forecast of your next month.
Body · Compound the losses
Each loss hits what is left. Do not subtract the percents.
You risk 2% of what is left, and you lose ten trades in a row. Enter the percent of the starting account that remains, to the nearest percent.
0.98 to the tenth is about 0.817, so about 82% remains. Subtracting 2% ten times would say 80% and would be the wrong operation. The base shrinks.
Same ten losses, but you risk 10% of what is left each time. Enter the percent that remains.
0.90 to the tenth is about 0.349, so about 35% remains. The streak was the same length. The bet size did the damage. This is the classroom picture of ruin arriving faster.
Wicks · The climb back
The trough is the new base.
The account fell from 100 to 50. Enter the percent gain required to get back to 100.
50 needs another 50, which is 100% of the trough. The drawdown was 50%. The recovery is not 50%.
Ten losses of a fixed fraction arrive in one week because five positions were the same theme. Correlation changed…
The formula for one trade did not change. The portfolio lesson's point is that the losses are not independent. The trough is a cluster, not a polite sequence.
- The per-trade formula
- The chance they arrive together, which deepens the trough
- Nothing, because math is per trade
Challenge · Fresh peaks
A new drawdown, then a sizing choice.
Equity peaks at 80 and troughs at 60. Enter the drawdown in percent.
20 divided by 80 is 25%. Using the trough as the base would call it 33% and would be the recovery, not the drawdown.
A backtest shows a 40% drawdown and you plan to feel fine at 10%. The mismatch is…
The curve is a path, including the hole. If you will abandon the rule at 10%, you will not live the backtest that went to 40%. That is a real constraint, not a pep talk.
- The test was too long
- The path you might live through is larger than the pain you wrote down
- Drawdown is always 10%
Reveal · One streak, worked
Made-up account, $10,000. The plan risks 1% of what is left. Fifteen losses in a row, no wins in between, leave about 86% of the start, near $8,600, because 0.99 to the fifteenth is about 0.86. The same fifteen losses at 5% leave about 46%, near $4,600. The win rate that would be required to climb out is a different question. This lesson only shows how fast the hole opens. A margin account that is forced to sell in the hole can turn the trough into a sale you did not schedule. The journal would record the streak as fifteen followed losses if the rule was followed, not as fifteen reasons to double.
At 1% risk, that fifteen-loss streak left about…
Small fractions compound slowly. 0.99 to a high power is still most of the account. Five percent was the version that cut it in half. The streak length was the same.
- 85% gone
- About 86% of the account still there
- The account at zero
Case · Case file: the hole in the curve
The bars after the last one are hidden. Decide, then the path is shown.
Check · Fresh holes
New peaks. Same asymmetry.
Equity peaks at 120 and later prints 90. The drawdown is…
- 30 points, so 30%
- 90 divided by 120, so 75%
- 25%
Answer
25%. 30 divided by 120 is 25%. The peak is the base.
An account falls from 100 to 80. The gain that gets back to 100 is…
- 25%
- 20%
- 80%
Answer
25%. 20 on a base of 80 is 25%. The drawdown was 20%. The way back is steeper.
Ten losses at 2% of what remains leave more of the account than ten losses at 10% because…
- Ten times 2% is exactly 20, and ten times 10% is exactly 100
- Each loss hits a smaller base, and the slice is thinner at 2%
- The broker waives the larger losses and keeps only the small ones
Answer
Each loss hits a smaller base, and the slice is thinner at 2%. The fraction compounds. A larger slice of a shrinking base reaches a floor faster.
A test's worst hole is 35% and the plan says you will quit at 10%. You are not going to live…
- Only the early winners, and none of the later hole
- The tested path, because you exit the rule earlier than the hole
- A smooth 10% gain, because drawdowns are not allowed
Answer
The tested path, because you exit the rule earlier than the hole. The test. You will have left before the path the test includes.
Key terms
- Drawdown
- The drop from an equity peak to a later trough, before a new peak. A 30% drawdown needs more than a 30% gain to get back, because the base is smaller.
- Risk of ruin
- The chance a sizing plan hits a floor you cannot continue from. In this lesson the floor is a classroom zero. Bigger bets reach it in fewer losses.