How do you calculate position size?

Decide the money you are prepared to lose on the trade, often a fixed share of the account. Divide it by the loss per share or contract if the stop is hit. Round down. For futures, the loss per contract is the stop distance in points times the contract multiplier.

What is the Kelly criterion in trading?

It is a formula from a 1956 paper by John Kelly for the fraction of capital to bet so that wealth grows fastest over the long run, given a known edge. In trading the edge is only an estimate, and full Kelly bets are large, so many practitioners use a fraction of it or none.

Position sizing: how many shares or contracts to trade

Fixed fractional position sizing step by step for shares and futures, why size follows the stop, and what the Kelly criterion says and where it fails.

Part of the Styles of Trading and the Honest Reality track on Agenticks. About 12 minutes, written for a intermediate reader.

Read first: Risk management basics for traders.

Last reviewed 2026-11-08.

Risk notice. Leverage multiplies losses as well as gains. Trading on margin or with leveraged products can lose more than the money you deposit, and a broker can close positions without notice.

Key takeaways

Fixed fractional sizing risks the same share of the account on every trade: size equals the money at risk divided by the loss per unit at the stop.
Size follows the stop distance: a wider stop means fewer shares or contracts for the same money at risk.
For futures, the loss per contract is the stop distance in points times the contract multiplier, and a large contract may simply not fit a small budget.
The Kelly criterion gives the growth-maximizing bet for a known edge, but real edges are estimates, and full Kelly bets are large and produce deep drawdowns.

Position sizing answers one question: how many shares or contracts to trade. It is easy to treat as an afterthought, picking a round number of shares or whatever the account allows, but it decides how much each trade can cost you. Two traders with identical entries and exits can end up with completely different results purely because of size. The previous lesson on risk management set a budget for each trade. This lesson turns that budget into a number of shares or contracts, then looks at a famous formula for how large that budget could be, and why it is usually treated with caution.

Fixed fractional sizing Fixed fractional sizing means risking the same share of the account on every trade. It takes three steps: 1. Risk budget = account size x risk per trade. On a $20,000 account at 1%, that is $200. 2. Risk per unit = the distance from entry to stop, times what one point is worth per unit. For shares, one point is $1 per share. 3. Size = risk budget / risk per unit, rounded down. Rounding up would risk more than the budget. Because the budget is a share of the current account, it shrinks after losses and grows after gains. After a 10% drawdown, the same $20,000 account holds $18,000, and 1% is now $180.

Worked example: shares, with three different stops

A $20,000 account risking 1% per trade, buying at $50. Only the stop changes. Illustrative numbers.

Risk budget (1% of $20,000)	$200
Stop at $49 ($1 a share) ($200 / $1; position $10,000)	200 shares
Stop at $48 ($2 a share) ($200 / $2; position $5,000)	100 shares
Stop at $46 ($4 a share) ($200 / $4; position $2,500)	50 shares
Loss if each stop is hit	$200 in every case

The position value changes fourfold while the money at risk stays the same. That is what it means for size to follow the stop.

The stop sets the risk, the size keeps it in budget

Choose the stop where the trade idea fails. Then let the size fall out of the arithmetic. Narrowing a stop just to trade more shares changes the trade, not only its size.

This is why traders talk about risk in money rather than in shares. "I bought 100 shares" says almost nothing about risk. "I risked $200, or 1%" says everything. It is also why a volatile stock with a wide natural stop is traded in a smaller size than a quiet one: the wider stop costs more per share, so fewer shares fit. Sizing this way lets you compare trades on very different instruments, and record each result in units of risk, R, as in the lesson on expectancy.

Sizing futures The method is identical for futures, except that one point is worth the contract multiplier. CME Group's E-mini S&P 500 is $50 per index point and the Micro E-mini is $5, as covered in what are futures. Take a $50,000 account risking 1%, a $500 budget, buying at an index level of 5,000: Stop 10 points away. Each Micro risks 10 x $5 = $50, so $500 / $50 = 10 Micros. Each E-mini risks 10 x $50 = $500, so 1 E-mini fits exactly. Stop 20 points away. Each Micro risks $100, so 5 Micros. Each E-mini risks $1,000, double the budget, so no E-mini fits at all. That last line is a common trap. Contracts cannot be split, so a large contract on a small account forces a choice between risking far more than planned or not trading. Smaller contracts exist partly to solve that. And notice that 5 Micros control 5 x 5,000 x $5 = $125,000 of exposure on a $50,000 account. The stop keeps the planned loss at $500, but a gap through the stop can cost more, and with futures losses can exceed the account.

Try it: shares or contracts for a risk budget

Use a point value of 1 for shares, 5 for a Micro E-mini S&P 500 and 50 for an E-mini. The defaults are the illustrative share trade above.

The Kelly criterion, and its caution Fixed fractional sizing says how to size once you have chosen the fraction. It does not say what the fraction should be. In 1956 John Kelly, an engineer at Bell Labs, worked out the fraction of capital a gambler with an edge should bet to make wealth grow fastest over the long run. In the simplest case, an even-money bet won with probability p and lost with probability q = 1 - p, that fraction is p - q. More generally, the Kelly bet is the edge divided by the odds. For a trade that wins b times the amount risked, that works out to p - q / b. Now plug in a trading example. Suppose a strategy wins 50% of the time and its winners average 1.5 times its losers. Kelly says p - q / b = 0.50 - 0.50 / 1.5 = 0.167, risking about 16.7% of the account per trade. That is more than sixteen times the 1% used in this lesson, and a run of losses at that size would be devastating. The bigger problem is that p and b are estimates. If the true win rate were 40% rather than 50%, the same formula gives 0.40 - 0.60 / 1.5 = 0, meaning no edge at all. The 16.7% bet would then be a large bet on nothing. MacLean, Thorp and Ziemba sum up the trade-off: Kelly maximizes long-run growth, but its suggested bets can be very large, it can be very risky in the short term, and errors in the estimated edge matter a great deal. Betting more than Kelly adds risk and lowers growth; at about twice the Kelly fraction, long-run growth falls to roughly nothing in their continuous-time analysis. That is why practitioners who use it often bet a fraction of it, such as half, and why many traders skip it entirely and keep risk per trade small. A backtest's win rate is an estimate too, and sample size decides how much to trust it.

Worked example: what Kelly suggests, and how fragile it is

Win rate p, loss rate q = 1 - p, and payoff b = average win / average loss. Illustrative numbers.

Estimated p = 50%, b = 1.5 (0.50 - 0.50 / 1.5)	Kelly 16.7%
Half Kelly on that estimate (16.7% / 2)	8.3%
If the true p is 45% (0.45 - 0.55 / 1.5)	Kelly 8.3%
If the true p is 40% (0.40 - 0.60 / 1.5)	Kelly 0%
A 10 point error in the win rate	Turns a 16.7% bet into a bet with no edge

The formula is exact. The inputs never are.

budget stop down

You can size a position

You can now turn a risk budget into shares or contracts, explain why size follows the stop, and say what the Kelly criterion promises and why it is used with caution.

Check your understanding

A $20,000 account risks 1% per trade. Entry is $50 and the stop is $48. How many shares?: 100. 1% of $20,000 is $200. The loss per share is $2. $200 / $2 = 100 shares.
The same trade's stop moves from $48 to $46. What happens to the position size if the risk stays at 1%?: It halves, to 50 shares. The loss per share doubles to $4, so $200 / $4 = 50 shares. The money at risk stays $200.
A $50,000 account risks 1% ($500). The stop is 20 points away. How many E-mini S&P 500 contracts ($50 a point) fit?: None. 20 x $50 = $1,000 per contract, double the budget. Five Micros at $100 each would fit instead.
Why do many practitioners bet less than the full Kelly fraction, if at all?: Because the win rate and payoff are only estimates, and full Kelly bets are large and swing hard. Overestimating the edge leads to overbetting, which lowers long-run growth and deepens drawdowns. A fraction of Kelly trades some growth for a smoother ride.

Common questions

How do you calculate position size?: Decide the money you are prepared to lose on the trade, often a fixed share of the account. Divide it by the loss per share or contract if the stop is hit. Round down. For futures, the loss per contract is the stop distance in points times the contract multiplier.
What is the Kelly criterion in trading?: It is a formula from a 1956 paper by John Kelly for the fraction of capital to bet so that wealth grows fastest over the long run, given a known edge. In trading the edge is only an estimate, and full Kelly bets are large, so many practitioners use a fraction of it or none.

Terms defined in this lesson

Continue

Sources

Written and reviewed by the Agenticks Learn editorial team. How we write and check lessons.

Agenticks Learn is general education. It is not financial, investment, tax or legal advice, it does not take your situation into account, and nothing on it is a recommendation to buy, sell or hold any security or to open any account. Examples use illustrative numbers. Investing involves risk, including losing money you put in, and past performance does not predict future results. Tax rules and limits change; the official sources linked in each lesson are the final word.