Position sizing from the stop, not the account

Ask most people how many contracts they can trade and they look at the margin. Margin is the broker's deposit requirement, and it will let a $10,000 account hold a position that one ordinary move can erase. The size of a trade comes from somewhere else entirely: from the stop. Decide how much you can afford to lose if you are wrong, measure how far away "wrong" is, and the size falls out of the division. Everything on this page is that one sentence, with numbers.

The two numbers

The first is your risk per trade: the fraction of the account you will lose if the stop is hit. The conventional range is 0.5% to 2%, and the top of it has a name — Alexander Elder's "2% rule: don't risk more than 2% of your account equity on any one trade." Nothing about the range is magic; it exists because of the arithmetic on the drawdown page, where a run of losses that is statistically ordinary turns into a hole you cannot climb out of once each loss is large.

The second is the stop distance in dollars: how many points from entry the stop sits, times what a point is worth. On the Micro E-mini S&P 500 a point is $5; on the E-mini it is $50. An 8-point stop risks $40 per MES contract and $400 per ES contract.

Then:

contracts = (account × risk per trade) ÷ (stop in points × dollars per point)

rounded down. If it rounds to zero, the answer is zero — the trade is too large for the account at that stop, and the honest response is a smaller contract, a wider account, or no trade. It is never a tighter stop chosen to make the division work.

What the formula gives

Hypothetical accounts, an 8-point stop, rounded down:

Read the right-hand column. At 1% risk and an 8-point stop, an account needs $40,000 before a single ES contract is the right size, and $50,000 before it is comfortably so. Plenty of $10,000 accounts trade ES. Each of them is risking 4% of the account per trade at that stop, and by the drawdown page's arithmetic a routine run of six losses is a quarter of the account.

Why margin is the wrong number

Day margin on an MES contract at the desk's broker was about $50 when last read (Aug 27, 2026). On that number alone, a $10,000 account could hold 200 contracts. A 10-point move — a quiet morning — against 200 MES contracts is $10,000. The whole account, on a move that happens most days.

Margin measures how fast the broker can liquidate you, not how much you should trade; the lower it is, the more dangerous it is to size from. How a futures contract works has the rest of that arithmetic. The desk's own sizing never reads margin as a limit; it reads it only to check that the position the stop allows is also one the account can post.

Where the stop comes from

The stop is not the number you can afford. It is where the trade is wrong — the price at which the reason for entering no longer holds — and that is a property of the market's noise, not of your account. A stop placed inside the range the market wanders through on a normal day will be hit by the wandering. Systematic traders usually set it as a multiple of recent volatility, so that it widens when the market is loud and narrows when it is quiet; a wider stop then means fewer contracts, not more risk. That is the whole discipline: the market sets the stop, the stop sets the size, and the risk per trade stays the same every time.

Kelly, and why nobody uses all of it

There is a mathematically optimal fraction to risk, and it has been known since 1956. John Kelly, working at Bell Labs on a problem about information rates, showed that a bettor with an edge grows wealth fastest by risking a fixed fraction of it each time:

f* = p − (1 − p) ÷ b

where p is the probability of winning and b is the payoff ratio — how much a win pays per dollar risked. A system that wins half its trades at 1.5-to-1 has f = 0.5 − 0.5 ÷ 1.5 ≈ 16.7% of the account per trade. A 55% system at even money has f = 10%.

Nobody who understands the formula bets that much, for two reasons. The first is that Kelly's fraction assumes you know p and b exactly, and you never do — a backtest gives an estimate, and the true values are usually worse. Bet more than the true Kelly fraction and growth falls; bet much more and it turns negative, so an error in your favour is far more expensive than an error against you. The second is the ride. Edward Thorp, who used the formula in blackjack and then in markets for decades, reports that betting half of Kelly keeps about three quarters of the growth rate with far smaller swings, and that the probability of ever halving your capital is one half at full Kelly and one eighth at half. Most practitioners, in his words, "strongly prefer the increased safety and psychological comfort" of the half. Fractions of a fraction — a quarter Kelly, or the plain 1% rule — are what people actually run.

What the desk does

Every hand on the desk shows its stop and its worst-case loss per trade — stop distance times dollars per point — before it can be dealt. A seat starts at one contract; a larger size is offered only when the account covers the margin for it plus ten consecutive stops at that size, and that rule is enforced on the server, not just in the buttons. Sim first, always, so the arithmetic above is watched on paper before any of it is real. None of that makes a system profitable; it only makes the losing stretch survivable, which is the precondition for finding out.

• Drawdown is the number that decides whether you survive
• How to read a track record (and the ten ways it lies)
• Leverage: the arithmetic of being wrong

← Risk, the common thread · All of Learn