Martingale and anti-martingale sizing
A martingale sizing scheme increases position size after each loss, classically doubling it, so that one eventual win recovers all previous losses plus the initial stake. Anti-martingale schemes do the opposite and increase size after gains. The martingale turns a high probability of small wins into a small probability of a catastrophic loss.
Senzoukria · Glossary · Updated September 2026
The arithmetic
Start by risking 100 dollars. After a loss, risk 200; after a second loss, 400; then 800, then 1,600. After five consecutive losses, the account has lost 3,100 dollars and the next trade must risk 3,200 to recover everything plus the original 100. After six losses, the cumulative loss is 6,300 and the next stake is 6,400. The required size grows as 2ⁿ, while the reward of the whole sequence stays equal to the first stake.
Streaks of that length are not rare. With a 50% win rate and independent trades, a sample of 100 trades contains a run of six or more losses about 55% of the time.
Why it fails in practice
- Every account has a limit: margin, a maximum position size, a daily loss limit or a drawdown floor. Once the progression hits it, the strategy cannot place the recovering trade and the accumulated loss is locked in.
- The expected value of the scheme is the expected value of the trades. Doubling after losses does not create an edge; it reshapes the distribution into many small gains and rare large losses.
- Losing streaks cluster when conditions turn against a strategy, which is exactly when the martingale is at its largest size.
- In futures, larger positions also face more slippage, so the recovering trade is the most expensive one to execute.
Anti-martingale
Anti-martingale methods raise size after wins and cut it after losses. Fixed fractional sizing is a mild, continuous form: risk is a percentage of current equity, so it grows with gains and shrinks during drawdowns. Pyramiding into a winning position is another. These methods do not create an edge either, but they concentrate size where the account can afford it rather than where it is weakest.
In Senzoukria
Several independent caps stop a size progression in the desktop. Settings can set a maximum order size in contracts, enforced by the backend for every order from the ticket, the chart quick ticket or the context menu. The autopilot has its own maximum contracts, 2 by default, and a maximum daily loss, 500 dollars by default, after which new automated entries are blocked. The simulated demo account can be given a contract cap and a daily loss limit that apply before each entry. The prop firm simulation checks the maximum contracts of the account rules you enter.
Related
In the same section
- Pyramiding
- Matching engine
- Marking the close
- Max pain
- Marketable limit order
- MAE
- Market-if-touched order
- Maximum contracts
This page in other languages
Frequently asked questions
- Can a martingale work with a high win rate strategy?
- A high win rate makes long streaks rarer but does not remove them, and the size needed to recover grows exponentially with each loss. Over enough trades, a streak long enough to hit the account's limit becomes likely, and that single event can exceed all the small gains before it.
- Is averaging down a form of martingale?
- It is closely related: both increase exposure as a position loses. Averaging down adds to the same open trade rather than sizing up the next one, but it shares the problem of concentrating risk precisely when the trade idea is failing.