Win Rate, Payoff, Expectancy and Drawdown

Key takeaways

1. Read win rate alongside realized payoff. A higher winning percentage can still produce a worse net result.

2. Count trades consistently, include zero outcomes, and handle costs consistently. A completed sample average does not establish a future advantage.

3. R multiples use a stated initial-risk reference. The planned amount does not guarantee the loss will stop there.

4. The same outcomes can produce different drawdowns. Examine the sequence, measurement method, duration, and limits of the record.

Full transcript

Read the whole result

How often does it win?

It is a reasonable question to ask about a trading strategy.

But the answer leaves something out.

How much does it gain when it wins, and how much does it lose when it loses?

Then there is another question.

What happens between the beginning and the final result?

In the previous lesson, we separated a defined strategy from a claim that it has an edge.

Now we need a clearer way to read the results.

We will connect four measurements: win rate, payoff, expectancy, and drawdown.

Each tells you something different. Reading them together helps you see what a single impressive number can hide.

The examples in this lesson are invented to explain the calculations. They are not trading records, forecasts, or evidence for a strategy.

Win rate and counting rules

Start with win rate.

For a set of completed trades, win rate is the number of winning trades divided by the total number of trades.

If four out of ten trades finish with a gain, the win rate is forty percent.

Before calculating it, define what counts as a trade and what counts as a win.

For this lesson, one trade means a completed position, from entry until the position is fully closed. Partial exits belong to that same trade.

We will classify results after the trading costs included in the record. A positive net result is a win. A negative net result is a loss. Exactly zero is a separate outcome.

A position closed at its entry price may still be a losing trade after commissions and fees.

Include zero-result trades in the total count. If a report excludes them, that changes the denominator and needs to be stated.

Consistency matters. Counting each profitable partial exit as a separate win, while grouping several losing exits together, would distort the comparison.

Two hypothetical records

Now consider two hypothetical records.

In the first, seven out of ten trades win.

Each winner gains fifty dollars net. Each loser loses one hundred fifty dollars net.

Seven wins produce three hundred fifty dollars. Three losses remove four hundred fifty dollars.

The final result is a loss of one hundred dollars, despite a seventy percent win rate.

In the second record, only four out of ten trades win.

Each winner gains two hundred dollars net. Each loser loses one hundred dollars net.

The gains total eight hundred dollars. The losses total six hundred dollars.

The final result is a gain of two hundred dollars, with a forty percent win rate.

The higher win rate produced the worse result in these examples.

Realized payoff

That brings us to payoff.

Here, the payoff ratio means the average winning trade divided by the magnitude of the average losing trade.

Use the loss as a positive size in that division.

In the second record, the average win is two hundred dollars and the average loss is one hundred dollars.

The payoff ratio is two. The average winner is twice the size of the average loser.

This is a description of the completed results.

It is different from placing a target twice as far from entry as a planned stop.

Your actual average win and loss depend on what happened: fills, partial exits, early exits, costs, and losses that exceeded the intended amount.

A planned target-to-stop relationship does not tell you the realized payoff ratio.

Also remember what an average hides. Several small losses and one unusually large loss can share the same average as a more even group. The individual results still matter.

Calculating expectancy

Next, connect frequency and size through expectancy.

For a record containing wins and losses, multiply the win rate by the average win. Then subtract the loss rate multiplied by the average loss size.

Use the rates as decimals.

In our second example, forty percent becomes zero point four. Sixty percent becomes zero point six.

Zero point four times two hundred dollars is eighty dollars.

Zero point six times one hundred dollars is sixty dollars.

Eighty minus sixty leaves twenty dollars per trade.

You can check that another way. The total net gain was two hundred dollars across ten trades. Divide two hundred by ten, and the average is twenty dollars.

For this completed sample, the two calculations describe the same average net result.

If exactly zero outcomes are present, use win and loss rates calculated from all the trades. The zero outcomes contribute zero to the average, but they still affect the rates.

We already used net results, so we do not subtract the same trading costs again.

If you begin with results before costs instead, you must account for the applicable costs consistently before calling the result net.

The first record, with seven small wins and three larger losses, averages a loss of ten dollars per trade.

Neither average tells you what the next trade will produce.

And this distinction matters: an observed sample average is not the same thing as knowing the strategy's true expected result.

The probabilities and average outcomes of future trades are uncertain. A positive historical average can be an estimate worth investigating. It does not establish an enduring advantage by itself.

The sample, conditions, rule selection, and testing process still matter, as we discussed in Lesson Eight.

R multiples

You may also see results expressed in units called R.

If one R is the initial planned price risk for a particular trade, a net gain of two hundred dollars on one hundred dollars of initial planned price risk is positive two R.

A net loss of one hundred twenty dollars on that same initial risk is negative one point two R.

The original planned amount is the reference. It is not a guarantee that the loss stops at one R.

State the convention and keep it consistent. When the dollar risk differs across trades, an average measured in R answers a different question from an average measured in dollars.

Frequency and payoff

Now we can ask how frequency and payoff balance each other.

If the average net win is twice the average net loss size, and every trade is either a win or a loss, a win rate of one third produces a zero average under those assumptions.

Above one third, that combination produces a positive average. Below one third, it produces a negative average.

That is a mathematical relationship, not a target that creates an edge. The actual payoff and win rate have to be measured, and they can change.

Drawdown and the path

There is still something missing from the average.

The path.

Return to the ten trades with four gains of two hundred dollars and six losses of one hundred dollars.

Assume fixed position size and no change in the individual dollar outcomes.

Put all four winners first. The cumulative result rises to positive eight hundred dollars. Then the six losses reduce it to positive two hundred dollars.

The series ends profitable, but it has fallen six hundred dollars from its highest point.

That decline is a drawdown.

Drawdown measures how far a measured value has fallen below its previous peak. Maximum drawdown is the largest such peak-to-trough decline in the period observed.

It is not the same as the final net loss, or the number of losing trades in a row.

Now arrange those same outcomes differently.

Alternate a winner and a loser four times, then place the final two losses at the end.

The cumulative result reaches a highest point of five hundred dollars and ends at two hundred dollars. The largest decline is three hundred dollars.

Both arrangements contain the same ten results. Both have the same win rate, payoff ratio, average result, and final gain.

Their maximum drawdowns differ.

This is why a summary of wins and losses cannot describe the full experience of following an approach.

Measurement and recovery

When reading a drawdown figure, ask what was measured and how often.

A curve recorded only after trades close can miss losses that developed while positions were open. Daily observations can miss a larger decline during the day.

Account equity includes the effect of open positions. A closed-trade record may not.

Deposits and withdrawals also need to be identified so that money moving into or out of the account is not mistaken for trading performance.

You can express a drawdown in dollars or as a percentage of the relevant peak.

For example, a decline from ten thousand dollars to nine thousand dollars is a ten percent drawdown, assuming no deposits or withdrawals.

Returning from nine thousand to ten thousand requires a gain of about eleven point one percent on the smaller amount remaining.

The loss percentage and recovery percentage use different starting values.

Depth is not the only feature to examine. How long did the decline last? How long did recovery take? Was the previous peak ever regained within the record?

And the largest drawdown already observed is not a ceiling on what can happen next.

A longer record, different conditions, or a different sequence can reveal a larger decline. Increasing size also changes the dollar consequences.

Reading a strategy report

So when you read a strategy report, work through the measurements together.

How were trades counted, and were the results measured after costs?

How often did they win?

How large were the average gains and losses, and what unusual outcomes did those averages hide?

What was the average net result per trade?

What drawdown occurred along the way, using which measurement method?

And how much confidence does the underlying evidence deserve?

Win rate describes frequency. Payoff compares the size of wins and losses. Expectancy connects frequency and size. Drawdown helps describe the decline along the path.

Together, they make the record easier to understand. They still cannot replace an examination of how that record was produced.

In the next lesson, we will look at strategy testing: how to turn a defined idea into a useful evaluation, and how to avoid giving ourselves an answer that only looks convincing.

Strategy Testing: Can You Trust the Results?

Turn a defined idea into a useful evaluation, and examine how testing choices can create an answer that only looks convincing.

Lesson 10 / 11:06 — video and full transcript.