New$10 Sign Up Bonus & $30 Free Trial Reward
Start challenge

Risk of ruin

Also called probability of ruin, gambler's ruin

The probability that a run of losses shrinks an account to the point where trading must stop. It rises with the share of capital risked per trade and falls as the trader's edge improves.

Reading
2 min
Sources
3Research
Updated
On this page· 5 sections
  1. How it works
  2. Example
  3. Watch out
  4. Sources
  5. Related terms

A method can win more often than it loses and still end an account. Risk of ruin puts a number on that chance. The mathematics comes from one of the oldest problems in probability, the gambler’s ruin, for which Abraham de Moivre gave a derivation in 1711[3].

How it works

The classic model is plain. A gambler starts with i units and makes a series of independent bets, each winning one unit with probability p or losing one unit with probability q = 1 − p. Play stops when the fortune reaches a target N or zero[1].

With a target, the probability of ruin is 1 − (1 − (q/p)^i) ÷ (1 − (q/p)^N) when p and q differ, and 1 − i/N when both are 0.5[1].

Without a target, against an opponent with unlimited money, the answer is starker. If q is smaller than p, the probability of ruin is (q/p)^i; if q is larger, or the two are equal, ruin comes with probability tending to one[3]. Unless the bets are strictly better than fair, ruin is certain[1].

Size enters through i. Risking half as much per bet doubles the number of units the same capital holds, and the ruin probability falls with the power of i.

Betting a fraction of current capital changes the arithmetic. In Kelly’s 1956 model, capital after W wins and L losses at a fraction ℓ is (1 + ℓ)^W × (1 − ℓ)^L times the starting capital[2]. On paper that shrinks without reaching zero. A real account stops earlier, at a loss limit or a margin floor, so ruin there means reaching that floor.

Example

Say each trade risks a fixed amount, one unit, and either wins or loses one unit. The account can take 10 such losses before it reaches the level where trading stops.

  1. Win rate 55%: q/p = 0.45 ÷ 0.55 ≈ 0.818. Risk of ruin with no target: 0.818^10 ≈ 13.4%.
  2. Half the risk per trade, so 20 units of room: 0.818^20 ≈ 1.8%.
  3. Win rate 52%, back to 10 units: (0.48 ÷ 0.52)^10 ≈ 44.9%.
  4. Win rate 50%: ruin is certain in an endless game. With a target of +10 units, it is 1 − 10/20 = 50%.

A drop of three points in the win rate more than triples the risk, from 13.4% to 44.9%.

Watch out

The model’s conditions are strict: bets of one fixed size, payoffs of one unit up or down, and outcomes independent of the past[1]. Real trades vary in size and payoff, so the numbers show the shape of the risk rather than a precise figure.

A positive edge does not remove ruin. With p above 0.5 the probability (q/p)^i shrinks as the room grows, but it stays above zero[3].

The win rate is an estimate. Step 3 shows how far a small error in it moves the result.

Sources

  1. ResearchColumbia UniversityGambler's Ruin Problem (Karl Sigman, lecture notes, 2009)Checked November 9, 2026
  2. ResearchBell System Technical Journal (copy hosted by Princeton University)A New Interpretation of Information Rate (J. L. Kelly Jr., Bell System Technical Journal, July 1956)Checked November 9, 2026
  3. ResearchDartmouth College (The CHANCE Project; 2nd edition published by the American Mathematical Society)Grinstead and Snell's Introduction to Probability (Section 12.2, Gambler's Ruin)Checked November 9, 2026
All terms in Tronexopedia

Your challenge starts here

Start your challengeFree TrialNew
Registration bonus

Get $10 free in your Wallet when you sign up

Register and use your $10 Wallet credit to reduce the price of your first challenge.

The $10 bonus is credited to your personal Wallet and can only be used to reduce the price of a challenge purchase. Credits cannot be withdrawn or used for payouts.

Start now & get $10