Kelly Criterion: Definition, How Formula Works as Risk Management Strategy

Kelly’s Criterion

Percentage Risk

The previous chapter examined three position sizing techniques: unit per fixed amount, percentage margin, and percentage volatility. Each offers a different way of answering the question of how much capital to commit to any given trade. The combination of equity estimation models and position sizing techniques can produce a wide variety of outcomes, and finding the right combination for a particular trading style requires experimentation and honest self-assessment.

Before moving on to Kelly’s Criterion, however, there is one more practical position sizing approach that deserves careful attention. It is called the Percentage Risk method. Many active traders consider it the most intuitive and user-friendly of all the available techniques, and it has the additional virtue of being directly connected to the stop loss, which is already a standard part of most trading setups.

What Makes Percentage Risk Different

The methods examined in the previous chapter determined position size based on either the capital deployed as margin or the typical daily price movement of the instrument. The percentage risk method takes a different and arguably more natural starting point. It asks a single question before every trade: how much am I willing to lose on this position if it goes against me?

That acceptable loss figure, expressed as a percentage of total capital, becomes the anchor for the entire position sizing calculation. The stop loss price establishes how much would be lost per share or per lot if the trade is exited at the stop. Dividing the maximum acceptable loss by the loss per unit then determines how many units to buy. The position size emerges directly from the risk rather than from the margin or the capital available.

This approach has an important psychological dimension that the other methods lack. It forces the trader to define and accept the worst-case outcome before entering the trade. Most market participants think about potential gains first and potential losses second, if they consider losses at all before entry. The percentage risk method reverses that sequence deliberately and productively.

A Worked Example: Setting Up the Trade

To illustrate how the percentage risk method works in practice, consider the following intraday trade setup on a stock futures contract.

A trader named Rahul is monitoring a mid-cap automobile sector stock that has been trading around a well-established support level of 410 rupees. This level has been tested on two separate occasions in recent weeks, and on both occasions the stock bounced meaningfully upward from it. The stock is currently sitting at 410.50 rupees, and the repeated holding of this support level suggests that buyers are active at this price. The logical target, based on prior price behaviour, is approximately 418 rupees.

Rahul’s trade details are as follows.

Stock: Automobile sector futures. Trade direction: Long, meaning the position profits if the price rises. Entry price: 410.50 rupees. Target price: 418 rupees, giving a potential gain of 7.50 rupees per share. Stop loss price: 406 rupees, where a support zone provides a logical point to exit if the trade moves against expectations. Stop loss value: 4.50 rupees per share, representing the loss that would be realised if the stop is triggered. Reward to risk ratio: 7.50 divided by 4.50, which equals approximately 1.67. This means the potential reward is 1.67 times the potential loss, which is a respectable ratio for a single-day trade. Lot size: 1,200 shares per lot. Margin required per lot: 68,000 rupees.

Why Buying the Maximum Number of Lots Is a Mistake

Rahul’s trading account holds 4,00,000 rupees. If he simply bought as many lots as his available margin permits, the calculation would be as follows.

4,00,000 divided by 68,000 = approximately 5.8 lots, which rounds down to 5 lots.

Five lots would require a total margin of 3,40,000 rupees. If the trade went against Rahul and all five lots hit the stop loss at 406 rupees, the loss would be calculated as follows.

Loss per share: 4.50 rupees. Shares per lot: 1,200. Number of lots: 5. Total loss: 4.50 multiplied by 1,200 multiplied by 5 = 27,000 rupees.

As a percentage of total capital, that loss amounts to 27,000 divided by 4,00,000, which equals 6.75 per cent.

Losing nearly 7 per cent of total capital on a single intraday trade is far beyond what sound risk management permits, regardless of how confident the trader feels about the setup. Even the most carefully analysed trade can go wrong. A 6.75 per cent loss on one position means that the account would need to generate a gain of more than 7 per cent simply to recover to its starting point, before any further progress can be made. This is precisely the recovery trauma dynamic described in an earlier chapter.

The percentage risk method prevents this outcome by working backward from an acceptable loss limit rather than forward from available margin.

Applying the Percentage Risk Rule Step by Step

Most disciplined traders limit their risk on any single trade to between 1 and 3 per cent of total capital. The specific figure depends on the trader’s overall strategy, their historical win rate, and their personal comfort with drawdowns. Rahul decides to set his maximum risk per trade at 1.5 per cent of total capital.

Maximum acceptable loss = 1.5 per cent of 4,00,000 rupees = 6,000 rupees.

This figure, 6,000 rupees, is the absolute ceiling on what Rahul is willing to lose on this trade. It does not matter how strong the setup looks or how confident he feels. The rule holds.

The next step is to calculate the loss per lot if the stop loss is triggered.

Loss per share at stop loss: 4.50 rupees. Shares per lot: 1,200. Loss per lot: 4.50 multiplied by 1,200 = 5,400 rupees.

If the stop loss is triggered on one lot, Rahul loses 5,400 rupees. To find the maximum number of lots he can buy whilst staying within his 6,000 rupee risk limit, he divides the maximum acceptable loss by the loss per lot.

6,000 divided by 5,400 = 1.11 lots, which rounds down to 1 lot.

Rahul therefore buys 1 lot, blocking a margin of 68,000 rupees. His maximum possible loss if the stop is triggered is 5,400 rupees, which represents 1.35 per cent of his 4,00,000 rupee account. This sits comfortably within his 1.5 per cent risk boundary.

The position size is determined entirely by the risk framework, not by the attractiveness of the trade or the amount of margin available. This is the discipline that separates systematic traders from those who size positions based on gut feeling.

Recalculating Equity and Risk for the Next Trade

Once the first position is open and the margin is blocked, the equity base for the next trade is recalculated. Using the Reduced Total Equity Model introduced in an earlier chapter, the process is as follows.

Starting capital: 4,00,000 rupees. Margin blocked for the open trade: 68,000 rupees. Remaining equity available for sizing the next trade: 4,00,000 minus 68,000 = 3,32,000 rupees.

Maximum acceptable loss for the next trade: 1.5 per cent of 3,32,000 rupees = 4,980 rupees.

The stop loss value and lot size of the next trade are then used to determine how many lots can be purchased within this revised 4,980 rupee risk limit. The process repeats for every subsequent position, with the equity base and the risk limit both adjusting dynamically as trades are opened, closed, or locked into profit through trailing stop losses.

This rolling recalculation is what makes the percentage risk method particularly well suited to traders managing multiple simultaneous positions. The risk exposure on each new trade is always calibrated to the current state of the account, not to the account’s original starting value.

What Happens When Confidence Is High

A question arises naturally at this point and is worth examining directly. If a trader has conducted thorough analysis and is highly confident in a particular setup, should they increase the position size beyond what the percentage risk rule permits? Should genuine conviction override the formula?

The instinct to do so is understandable. A trader who has studied a chart carefully, identified a clear support level, noted a favourable reward to risk ratio, and seen similar setups resolve profitably in the past will naturally feel that this particular trade deserves more capital than the standard rule allocates. It feels like leaving money on the table to apply the same 1.5 per cent risk limit to a high-conviction trade as to an average one.

This is exactly the question that Kelly’s Criterion was developed to answer. Rather than relying on subjective confidence, Kelly’s Criterion provides a mathematical framework for determining the optimal position size based on the historical probability of a trading system producing winning trades and the ratio of average wins to average losses. It quantifies how much more, or less, should be risked on any given trade based on the statistical edge of the underlying strategy.

The formula and its practical application in a trading context are examined in detail in the section that follows.