How to Measure Consistency in Your Trading Returns

In short. To measure consistency in your trading returns, group results into equal periods such as weeks, then check four numbers: the share of profitable periods, how widely period results swing around their average, how much of total profit comes from the single best period, and how the median compares with the mean. Two traders can finish with the same profit and have completely different consistency, and these numbers show the difference.

Consistency of returns is not the same as a prop firm consistency rule

The word "consistency" is used for two different things in trading. A prop firm consistency rule is a payout condition set by a firm, usually a cap on how much of your profit may come from your best day. That rule is covered in what a prop firm consistency rule is.

This article is about something you can measure on any account, including a personal one: how evenly your own results are spread over time. It describes your process, not a firm's requirement. A trader can pass a firm's best-day cap and still have results that are driven by one lucky week.

Why total profit hides the picture

Total profit and average profit say nothing about the path. Eight weeks that add up to +1,600 could be eight steady weeks, or six small losses and one large win. The first record suggests a process that repeats. The second may only show one outsized trade that cannot be counted on again.

Expectancy tells you what an average trade is worth. Consistency asks a different question: how much do the results around that average move from one period to the next? Both belong in a review, and neither replaces the other.

Step 1: choose a period and build the series

Everything below starts from a list of period results. Weekly buckets are a practical default, because a single day is mostly noise and a month gives too few data points for most journals.

1. Pick one period length and keep it. Mixing weeks and months breaks every comparison.
2. Sum the net result of each period after commissions, swaps, and fees, in account currency or, better, in R or as a percent of the account.
3. Keep position sizing in mind. If size changed a lot during the sample, dollar results across periods are not comparable, so use R or percent.
4. Collect enough periods. With fewer than about eight to twelve, every metric below moves a lot when one period is added. This is a rule of thumb, not a statistical threshold.

Four numbers that measure consistency

Two of these deserve a closer look.

Standard deviation is the usual measure of how much a series fluctuates. Morningstar defines it for fund returns as how much the returns have fluctuated in the past, where a higher value means returns swung more widely and a lower value means they were steadier (Morningstar glossary, standard deviation). It is the same input used in the Sharpe ratio, so you may already track it.

The coefficient of variation puts that swing in proportion to the average. NIST defines it as the sample standard deviation divided by the sample mean, and notes that it becomes sensitive to small changes in the mean when the mean is near zero (NIST Dataplot reference, coefficient of variation). For a trader, this means CV is useful for comparing two periods of profitable trading, and unreliable when the account is close to break-even or has a negative average.

A worked example: same profit, different consistency

The figures below are illustrative, not real trader data. Two traders each make +1,600 over eight weeks, so each averages +200 per week.

Important. Equal total profit does not mean equal quality of results. Judge a track record by how the profit was distributed over time, not only by the final number.

Now the four metrics, using the sample standard deviation:

Trader A lost in two weeks, but never by much, and the average week looks like a typical week. Trader B has the same average, yet the median week is a loss. Without week 3, trader B would be down 1,000. That is the pattern to look for: when the median sits far below the mean, a few periods are carrying the result.

The comparison does not say trader B has no edge. It says that eight weeks of data cannot separate an edge from one lucky trade, which is a common situation when the sample is small. More data and a review of the week 3 trade are the next steps.

How to read the numbers together

No single metric is enough, because each one hides something:

• High share of profitable periods, high best-period share: many small wins plus one dominant period. Check whether the dominant period came from a setup you can repeat.
• Low share of profitable periods, positive total: a few large wins pay for many losses. This can be a valid style, but it needs a larger sample and enough capital to survive the losing runs, as covered in maximum consecutive losses.
• Low standard deviation, low mean: steady but barely profitable. Costs can erase the edge here.
• Falling CV over time: results are getting more even relative to the average, which is the direction most traders are aiming for.

None of these numbers predicts future results. They describe the past sample, and markets and position sizes change.

Common mistakes when measuring consistency

• Mixing period lengths: comparing a weekly CV with a monthly one gives a meaningless result.
• Ignoring size changes: when risk per trade grows after a good run, later periods look more volatile in dollars even if the process did not change. Use R or percent.
• Trusting a tiny sample: three or four periods can produce any metric you like.
• Using CV near break-even: if the mean is close to zero, the ratio explodes. Look at standard deviation and the median instead.
• Cutting the worst period: removing a bad week to "clean" the data defeats the purpose of the measure.

What to log so the numbers update themselves

A journal makes these metrics a matter of reading, not recalculating. Useful fields are net result per trade after costs, risk per trade in R, the date and the account, and a setup tag. With those in place, a weekly view of closed P&L, such as the one in a trading calendar, gives you the period series directly. Reviewing it once a month fits naturally into a monthly trading review: add the four numbers to the checklist and compare them with the previous month.

This article is for educational purposes only and is not financial or investment advice. The examples are illustrative and do not describe real trading results. Trading, especially with leverage, can lead to the loss of your entire deposit.

See your weekly and monthly results side by side in the BitStat calendar, and follow the totals on the BitStat dashboard.