Part 2: How much to risk · Chapter 6 of 14 · 12 min

The risk of ruin

The risk of ruin is the probability of touching, at least once, a loss you can no longer afford: here −30% and −50% of the account. Without an edge it grows with time: at 2% risk per trade, in a 50%-winning model a 50% loss arrives in 26% of cases in 500 trades and in 43% in 1,000. With a true edge it stays put: SX, at 10% risk, touches −50% in 9-10% of paths at any horizon, and at 1-2% almost never.

What ruin is

Ruin in the strict sense is losing all the capital. With a small risk and fixed fraction (chapter 5) it never happens: after every loss you risk less. For someone who trades, though, ruin arrives much sooner: when the account has fallen so far that it makes no sense to go on, or that you can no longer recover. A 50% loss is recovered only with +100% (chapter 4).

That is why the risk of ruin is defined with a threshold: the probability that the account touches at least once a certain loss (−30%, −50%) within a certain number of trades. You look at the lowest point of the path, not at the final result: an account that falls 50% and then climbs back has touched the threshold.

The simplest model

Start from a model without tricks: each trade is worth +1 R with probability p and −1 R with probability 1 − p. If you risk a fraction f of the initial account per trade (fixed size), the −50% threshold is equivalent to losing N = 50% ÷ f units: with f = 5% that is 10 units, with f = 1% it is 50.

For p greater than 0.5 and an endless horizon, the probability of losing N units is:

> ((1 − p) / p)^N

With p = 0.6 and f = 5% (N = 10): (0.4/0.6)^10 = 1.7%. With f = 10% (N = 5): 13.2%. With f = 20% (N = 3): 30%. With f = 1% (N = 50) the probability is about one in six hundred million.

For p = 0.5 or less the formula says 100%: sooner or later the threshold arrives. What counts then is the horizon. The probability of losing 50% (N = 50% ÷ f units) in a finite number of trades, for different values of p, in 500 trades:

Winning share pRisk 1%Risk 2%Risk 5%Risk 10%
45%58.7%92.8%99.0%99.7%
50%2.5%26.4%65.5%82.3%
55%0.0%0.6%13.3%36.5%
60%0.0%0.0%1.7%13.2%

And for p = 0.5, what changes with the horizon:

Risk per trade250 trades5001,000
1%0.2%2.5%11.4%
2%11.4%26.4%42.9%
5%52.8%65.5%75.2%
10%75.2%82.3%87.4%

A fair coin with 2% risk per trade touches −50% in one case in four in 500 trades and in almost one in two in 1,000. This is the hardest reading in the chapter: without an edge the risk of ruin grows with time and does not stop. With an edge (p = 0.55 or 0.6) the curve flattens: more trades do not raise the probability.

The real series

Let us put the real series in an account with fixed fraction, in the order in which they happened, with growing risks.

Series and riskMinimumFinal−30% touched−50% touched
LA 2%62%62%trade 517 (16 September)never
LA 5%20%20%trade 356 (26 June)trade 481 (30 August)
LA 10%1%1%trade 158 (13 March)trade 241 (28 April)
LA 20%0%0%trade 18 (24 December)trade 34 (5 January)
LD 5%81%108%nevernever
LD 10%29%47%trade 14 (2 January)trade 322 (9 September)
LD 20%0%1%trade 6 (26 December)trade 13 (1 January)
SD, from 2% to 20%100%from 110% to 199%nevernever

(Dates between December 2025 and September 2026.) LA, which has a mean of −0.03, reaches −50% at 5% risk in August and at 10% in April: the more you risk, the sooner it arrives. LD has a mean of +0.03 and closes at +10.6 R, but at 10% risk it touches −30% already at the fourteenth trade and −50% in September, and ends at 47%: a series in profit and an account in ruin. At 20% risk LD ends at 1% and touches −50% at the thirteenth trade: the risk is greater than its mean can bear (chapter 5: beyond 2μ/σ² = 6% capital falls even with a positive mean).

Figure 1. LA, capital as a percentage of the initial one with fixed fraction and risks of 2% (bars, final 62%), 5% (pink, final 20%) and 10% (blue, final 1%); the lines are at 100% and 50%. The higher the risk, the sooner the account goes below 50%.Data: gold, XAUUSDT perpetual future (real candles) · daily (D1) · from 15 Dec 2025 to 5 Oct 2026.
Read the chart in words

Daily gold chart with, below, three curves of capital with different risks and two horizontal lines at 100% and 50%.

Figure 2. LD, capital as a percentage of the initial one with risks of 5% (bars, final 108%), 10% (pink, final 47%, minimum 29%) and 20% (blue, final 1%); the lines are at 100% and 50%. The same series, with a mean of +0.03 R, can end in profit or in ruin depending on the risk chosen.Data: gold, XAUUSDT perpetual future (real candles) · daily (D1) · from 15 Dec 2025 to 5 Oct 2026.
Read the chart in words

Daily gold chart with, below, three curves of the capital of LD with different risks and two horizontal lines at 100% and 50%.

On random paths

The real order is only one. To see what could happen with other orders, I draw at random from the results (bootstrap, chapter 2) 500 trades, 3,000 times, and count how many times the account touches −50% and −30%. Fixed fraction, probability of touching −50% / −30% in 500 trades:

SeriesRisk 1%Risk 2%Risk 5%Risk 10%
LA (mean −0.03)1.9% / 34.2%39.7% / 74.2%87.7% / 95.0%98.7% / 99.4%
LD (mean +0.03)0.0% / 4.5%6.2% / 27.7%49.4% / 71.0%84.4% / 92.0%
SD (mean +0.09)0.0% / 0.0%0.0% / 0.0%0.0% / 1.5%2.5% / 14.0%
SX (true edge)0.0% / 0.0%0.0% / 0.1%0.6% / 6.0%9.3% / 27.1%

Two things to look at.

  1. The effect of the horizon reveals whether there is an edge. For LA at 1% the probability of touching −50% is 0.0% at 250 trades, 1.9% at 500, 22.6% at 1,000: it keeps rising. For SX at 10% it is 9.6%, 9.3%, 9.9%: it stays put. If your probability of ruin grows with the trades you add, you do not have an edge that contains it.
  2. Risk makes more difference than results. LD is an almost break-even series: at 1% it touches −50% almost never (0.0%), at 5% one time in two (49.4%), at 10% almost always (84.4%). The same series, three different worlds.

SX, which has a built-in edge, also shows that the risk of ruin is not zero: at 10% risk it touches −50% in one path in ten, despite 60% of trades winning. The formula value is 13.2%: the same order of magnitude (the simulation gives 9.3% because with fixed fraction the account falls a little less than with fixed units).

Why the calculation is done twice

The flaw of every table on the risk of ruin is that it depends on p and on how much you earn, and these numbers you do not know: you estimate them from a series with its error (chapter 2). That is why the calculation is done twice:

  1. with the edge you hope for (for example p = 0.55-0.6), to know how exposed you are if everything goes as planned;
  2. with zero edge, or slightly negative (like LA), to know what happens if there is nothing.

With p = 0.5 at 500 trades the table says: risk 1% → 2.5% chance of touching −50%; risk 2% → 26.4%; risk 5% → 65.5%. With a small risk the no-edge case hurts but is not the end; with a large risk the loss is almost certain. With a slightly negative edge (LA), even 1% reaches the −30% threshold in 34% of 500-trade paths: no fixed fraction saves you for long, because a small risk slows the fall without stopping it. What stops the fall is a rule to quit and check the system (chapter 13).

What to do with it

  1. Choose the ruin threshold beforehand (−30%? −50%?) and the probability you accept of touching it (for example under 5% in a year).
  2. Compute two scenarios: with the edge you hope for and with zero edge. The risk per trade must be chosen so that the second is bearable too.
  3. Stay small: with p = 0.5 at 500 trades 1% gives a 2.5% chance of touching −50%; 5% gives 65%.
  4. Look at the horizon: if the probability of ruin grows with the trades, the edge you think you have is not enough.
  5. Add a rule to quit (chapter 13): small risk does not save a system without an edge.

Procedure in four steps

  1. Set the ruin threshold and the maximum probability you accept.
  2. Estimate p and the mean gain from your series, with the error (chapter 2), and take the zero-edge scenario too.
  3. With the bootstrap on your results, or with the tables, find the risk per trade for which the probability stays under the limit in both scenarios.
  4. Add a rule to quit, because small risk alone does not protect from an edge that is missing.

Exercises

Chapter exercises

1. A system wins 60% of trades at ±1 R. With a 10% risk per trade (simple model, −50% threshold), what is the probability of ruin?

2. A system with no edge (p = 0.5) risks 2% per trade. How likely is it to touch −50% in 500 trades?

3. What does a probability of ruin that grows with the number of trades indicate (LA at 1%: 0%, 1.9%, 22.6% at 250, 500, 1,000)?

4. LD has a mean of +0.03 R and closes the series in profit, but with 20% risk per trade it ends at 1% of the account. Why?

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.