What is the formula for trading expectancy?

Hit rate times average gain, minus loss rate times average loss. All four figures are already in your journal. An example with freely chosen numbers: with a hit rate of 40 percent, an average gain of 2 R and an average loss of 1 R, 0.4 times 2 gives 0.8 and 0.6 times 1 gives 0.6, so expectancy is 0.2 R per trade. Risking one percent of a 10,000 euro account makes one R worth a hundred euros, so expectancy is twenty euros per trade. What matters is calculating from net results, meaning after all costs, otherwise you produce a figure that does not exist in your account.

What is the difference between expectancy and profit factor?

Both describe the same trades and answer different questions. Profit factor is a ratio, namely the sum of gains divided by the sum of losses. It is dimensionless and therefore suits comparing two strategies or two segments, but says nothing about magnitude. Expectancy translates the same data into an amount per trade and is therefore the only one of the two you can extrapolate with. In practice: profit factor answers which of two variants is better, expectancy answers whether a variant is worth the effort at all. Those are two different decisions, and they do not always come out the same way.

Why should I calculate expectancy in R?

Because an expectancy in currency depends on your account size and therefore changes as the account grows or shrinks. A time series in currency mixes two effects, namely the quality of your trades and the development of your capital, and stops being readable. Calculated in R the figure stays untouched by that and is comparable across periods, instruments and account sizes. That is exactly what you need to judge whether something improved. Converting back to an amount is trivial whenever you want it, namely expectancy in R times your risk amount per trade, so keep the metric in R and convert on demand instead.

Trading Expectancy: What an Average Trade Is Worth

Q: Can a positive expectancy still not be enough?

Yes, because the metric applies per trade and says nothing about how many trades are available. An expectancy of 0.05 R at two trades a month produces an annual figure that barely justifies the work, even with the sign in your favour. Judging it therefore requires a second figure, namely expectancy multiplied by the number of trades your approach realistically produces, since only that product describes what happens across a year. It follows that an approach with a small expectancy and many opportunities can make more sense than one with a large expectancy and very few, provided costs allow it, because costs grow with the count.

Expectancy tells you what one average trade of your strategy is worth, expressed as an amount or in R. It is the third of the three core metrics and answers a question profit factor and R multiples leave open: what should a hundred further trades under unchanged rules be expected to produce? That is an extrapolation from your existing distribution rather than a promise about the future.

How do you calculate expectancy?

From four figures, all already in your journal: hit rate, average gain, loss rate and average loss. The calculation is hit rate times average gain, minus loss rate times average loss.

An example with freely chosen numbers, showing only the mechanics. Suppose your hit rate is 40 percent, your average gain 2 R and your average loss 1 R. Then 0.4 times 2 gives 0.8, and 0.6 times 1 gives 0.6. The difference is 0.2 R per trade.

That 0.2 R is your expectancy. Risking one percent per trade on a 10,000 euro account makes one R worth a hundred euros, so expectancy is twenty euros per trade. Across a hundred trades that is two thousand euros, provided the distribution holds and costs are already accounted for.

That last condition matters. Calculate the metric from net results, otherwise you are calculating an expectancy that does not exist in your account.

How does it differ from profit factor and R multiples?

All three describe the same trades but answer different questions. Drawing that distinction is the real value here.

Profit factor is a ratio: the sum of gains divided by the sum of losses. It is dimensionless and therefore suits comparing two strategies or two segments, but says nothing about magnitude. How to calculate it is covered under calculating your profit factor.
An R multiple normalises a single trade against the risk you took in it. It makes trades of different sizes comparable but always concerns one trade only.
Expectancy translates both into an amount per trade and is therefore the only one of the three you can extrapolate with.

In practice: profit factor answers which of two variants is better. Expectancy answers whether a variant is worth the effort. Those are two different decisions, and they do not always come out the same way.

Why calculate in R rather than in currency?

Because an expectancy in currency depends on your account size and therefore changes as the account grows or shrinks. A time series in currency mixes two effects: the quality of your trades and the development of your capital.

Calculated in R the figure stays untouched by that. It is comparable across periods, instruments and account sizes, which is exactly what you need to judge whether something improved.

Converting back to an amount is trivial whenever you want it: expectancy in R times your risk amount per trade. So keep the metric in R and convert on demand rather than the other way round.

What does a positive expectancy tell you, and what not?

It tells you your existing distribution of gains and losses summed to something positive. It does not tell you the next ten trades will be positive, and it says nothing about order.

That is the most common misreading of this metric. A positive expectancy is perfectly compatible with long losing streaks, because it describes an average rather than a sequence. How deep such declines can run therefore deserves separate attention.

The reverse holds too: a single poor month does not disprove a positive expectancy. Whether it has changed can only be judged across enough trades rather than across a period.

Why can a good expectancy still not be worth the effort?

Because the metric applies per trade and says nothing about how many trades are available. An expectancy of 0.05 R at two trades a month produces an annual figure that barely justifies the work, even with the sign in your favour.

So judging it requires a second figure: expectancy multiplied by the number of trades your approach realistically produces. Only that product describes what happens across a year.

An uncomfortable but useful consequence follows. An approach with a small expectancy and many opportunities can make more sense than one with a large expectancy and very few, provided costs allow it. Costs are the decisive constraint here, because they grow with the count.

How do costs shift the result?

Directly and in full, because they come off every individual trade. In the example above with an expectancy of 0.2 R, average costs of 0.05 R per trade consume a quarter of it.

Costs here mean more than commissions. The gap between intended and actual fill price belongs in there too and is regularly underestimated, covered under measuring slippage. With leveraged products, ongoing holding costs come on top.

A practical recommendation: calculate expectancy twice, once from pure price results and once after all costs. The distance between the two figures is your cost ratio, and at high trading frequency it is often larger than expected.

How many trades do you need?

Enough that the four inputs are stable, and that is more than most people assume. Thirty to fifty trades under unchanged rules is a rough lower bound, and it applies per segment once you split by setup.

The average gain is the most sensitive input, because a few large trades can shift it substantially. So look not only at the mean but at how much of it comes from your largest individual trades. If expectancy falls sharply without your three best trades, it rests on individual cases rather than on a pattern.

How to approach analyses like this in a structured way, without interpreting noise every month, is covered under reviewing your trades.

How do you read expectancy over time?

In blocks rather than as a single number. An expectancy across your entire history is an average over every rule change, market regime and learning step contained in it, which hides exactly the development you want to see.

Blocks of equal size work well, say fifty trades each, rather than calendar months. The reason is comparability: an active month and a quiet month contain different numbers of trades, and comparing them measures the count as much as the quality.

One rule devalues everything else if it is missing: note every change to your rules with a date. Without those markers you cannot attribute a jump in the series and have no way of telling whether it came from a change you made or from the market.

What do you do with a negative expectancy?

Break it apart rather than discard it. The overall figure is almost always a blend of several setups, market regimes and standards of execution, and within that blend a part is often positive.

Three splits tend to pay best. By setup shows whether one idea is financing another. By rule compliance shows whether the problem sits in the system or in the execution. And by market regime shows whether the approach only works structurally in certain environments.

How to set up those segments without fragmenting your sample is covered under tagging your trades. Analyse one dimension at a time throughout.

Which journal fields does this need?

Only two additional ones if result and date are already recorded: initial risk per trade, and costs kept separate from the price result. Initial risk produces the R, and the separation produces the net calculation.

Initial risk has to be there before the trade, because it cannot be reconstructed later, only remembered. A trading journal with an exchange connection supplies result, date and costs on its own; initial risk is the only field that has to come out of your planning.

Which metrics carry beyond these, and in what order to introduce them, is covered in the overview of the metrics that matter most.