Larry Hite: Sizing Risk Through a Sequence of Losses

Five consecutive trades each lose 1% of the account’s value at entry. The fifth loss feels like a verdict on the method, yet the arithmetic describes a 4.90% drawdown, not a complete account failure. Larry Hite’s emphasis on accepting risk directs attention toward how a process handles that sequence before the first trade begins.

The publisher’s description of The Rule presents Hite’s approach through numbers, trend following and the relationship between limited losses and larger winners. Wiley’s contents for Market Wizards names his interview chapter “Respecting Risk”.

The calculations below are independent hypothetical examples. They aren’t Hite’s performance figures, his position limits or instructions about an appropriate risk level for any reader.

Larry Hite and the fifth losing trade

Starting with 100 units, five losses of 1% of the remaining balance leave 100 multiplied by 0.99 to the fifth power, or approximately 95.10. The decline is about 4.90%. Five losses of 5% instead leave approximately 77.38, a decline of 22.62%.

I compare those two ending balances because the entry signals could be identical. Position size changes how much of the account remains available after the same five adverse outcomes. It also changes the recovery required: 77.38 needs about 29.24% to return to 100.

A longer sequence, a large gap or correlated positions can produce substantial damage even at the smaller hypothetical loss rate. The calculation isolates one variable so its effect is visible; it doesn’t model every source of risk.

The Ralph Vince profile offers related reading on position sizing. A method’s entry logic and its allocation rule form a combined process. Evaluating either in isolation can leave the eventual account path difficult to understand.

A stop distance is an assumption about execution

Suppose a hypothetical entry is 50.00 and an intended exit is 48.00. With 200 shares, the planned price loss is 400 before costs. If the first executable price after an overnight gap is 45.00, the loss is 1,000 before costs.

The 48.00 level remains useful as a planned response, but it didn’t cap the realised loss at 400. The market supplied no execution at the expected price in this scenario. A risk worksheet that calls 400 the guaranteed maximum would misdescribe the position.

A trader studying this setup might compare several assumed fills, including 48.00, 47.50 and 45.00. Those give price losses of 400, 500 and 1,000 for the same share count. The point is to make the sensitivity visible before an exceptional outcome occurs.

The ATR-based swing stop guide offers a separate framework for relating planned distance to recent movement. It doesn’t make future gaps predictable, and its inclusion here doesn’t establish that Hite used that particular tool.

A low win rate needs sufficient winning size

Imagine one hundred trades with thirty-five winners averaging 3R and sixty-five losses averaging 1R. R is a fixed unit of planned initial risk. Gross results total 105R minus 65R, or 40R, giving an average of 0.40R per trade.

If the average winner falls to 1.5R while everything else stays unchanged, winners contribute 52.5R and losses still cost 65R. The result becomes minus 12.5R. The same 35% win rate now accompanies a negative average.

I focus on that reduction from 3R to 1.5R because it shows what early exits can change in a hypothetical payoff distribution. It doesn’t prove that every early exit is wrong. A new exit rule might alter both the win rate and the losing distribution, and the whole result would need measuring again.

Costs also belong in the comparison. An average cost of 0.10R reduces the first example from 0.40R to 0.30R per trade. If realised losses exceed the assumed 1R during gaps, the gross estimate changes before those costs are deducted.

The large winner may arrive after the difficult stretch

Take a short illustrative sequence of six losses at 1R each, followed by one gain of 8R. The final result is plus 2R before costs. The first six observations alone show minus 6R. Knowing the ending makes that initial stretch look easier to tolerate than it would have felt in real time.

The example demonstrates how a distribution can depend on infrequent large outcomes. Another sequence might end without the 8R winner. A researcher can’t justify continuing a deteriorating method merely by assuming that such a winner is due.

There is a separate statistical issue. Real trades may share market conditions, so consecutive outcomes needn’t be independent. A string of losses could reflect chance, a changed environment, execution problems or a weak original hypothesis. The sequence alone can’t select the explanation.

A predefined review can examine whether the signals, costs and fills still match the tested method. That review differs from changing rules after every disappointment. It gives adverse evidence somewhere to go without allowing either blind persistence or impulsive reinvention to become the default.

Six markets can still share one adverse event

Suppose six hypothetical positions each carry planned risk of 0.5% of account equity. Their simple sum is 3%. If all six respond adversely to the same event, that planned exposure can be realised together, with gaps potentially adding more.

Six different markets can still carry overlapping exposure. Currency, rates, financing and commodity exposures can overlap. Different instruments may express similar economic bets, particularly during stress.

An exposure table can record the position, direction, planned loss and plausible shared driver. A second table might group hypothetical losses by event, rather than by instrument. That makes the concentration visible even when each individual trade appears modest.

The sector comparison guide provides another lens on shared market behaviour. Its rankings don’t measure all portfolio dependencies. Strong relative performance in several related markets can coexist with substantial exposure to the same reversal.

Changing size after a streak changes the experiment

Imagine four trades with outcomes of plus 1R, plus 1R, plus 1R and minus 1R. At a constant risk amount of 100, the sequence earns 200 before costs. If the risk amount doubles to 200 only for the last trade, the same signal outcomes earn 100.

That simple example shows why a trading record needs both the normalised outcome and the money amount. A favourable signal history can produce a disappointing account result when the largest positions happen to coincide with losses.

The reverse can also happen. A lucky increase before a winner can improve a short record. Assessing the sizing rule requires a wider comparison of outcomes. Evaluating only the final account return may conceal how much depended on one allocation decision.

The Ulcer Index discussion provides related information about the path below prior peaks. Different allocation sequences can create different drawdown experiences even when the underlying trades are identical. The path belongs beside the final result.

The operating record has to include mistakes

A hypothetical system might generate an entry at 60.00 while the actual fill occurs at 60.40. If the intended exit remains 58.00, the price distance grows from 2.00 to 2.40, a 20% increase. A spreadsheet retaining only the ideal entry understates the risk taken.

Missed signals and delayed exits deserve the same treatment. They may reflect practical limits on monitoring, order placement or liquidity. Describing a method as systematic doesn’t eliminate those implementation details.

A useful record distinguishes the proposed rule, its signal and the action that occurred. That separation makes diagnosis possible. Otherwise a loss caused by a faulty fill assumption can be blamed on psychology, while a weak signal can be excused as an execution error.

Hite’s risk emphasis is productive when it reaches this level of detail. The number on the worksheet needs to describe something that could actually happen in the account.

The actual contract size can impose another constraint. Suppose a hypothetical futures contract carries 25 per point and the proposed adverse distance is 8 points. One contract represents 200 of planned price risk before costs. A model calling for 100 of risk cannot achieve that amount with half a contract if fractional contracts aren’t available. Rounding up changes the risk; choosing no position changes participation. The operational record needs the actual choice. A smooth percentage-allocation formula can conceal these discrete limits in smaller accounts.

Books for studying Larry Hite

The Rule is Larry Hite’s book; Michael Covel wrote its foreword. Market Wizards is Jack D. Schwager’s interview collection and includes Hite. Trend Following is by Covel and supplies related reading. The linked Market Wizards is an Updated paperback, and Trend Following is the Updated Edition. The 35%-win-rate exercise above is independent of all three.

As an Amazon Associate, I earn from qualifying purchases.

The Rule. Hardcover.

(US) The Rule
(UK) The Rule
(DE) The Rule

Market Wizards. Paperback.

(US) Market Wizards
(UK) Market Wizards
(DE) Market Wizards

Trend Following. Paperback, Updated Edition.

(US) Trend Following
(DE) Trend Following

The sixth trade starts with the remaining balance

After five 5% losses, the hypothetical account contains 77.38 of its original 100 units. Any next decision starts there. A risk process earns its place by making that possibility explicit before the sequence begins and by retaining room to examine whether the method still deserves to continue.

Learn the pattern. Ride the trend. Keep the gains.

Educational content only. Not investment advice. Trading involves risk. You are responsible for your decisions.