Average time in trade is the mean duration between opening and closing your trades. It describes your style, scalping, intraday or swing, and above all it reveals the gaps between what your plan says and what you actually do.
The distribution, not the average
Contents
Average time in trade is computed without inventing anything: for each closed trade, take the gap between close time and open time, and average them. It is one of the few statistics in a journal that depends neither on capital, nor on market, nor on result.
Its first use is descriptive. A median duration of a few minutes describes scalping, a few hours intraday, several days swing. That sounds trivial, and it is not: many traders discover, looking at this figure, that they do something other than what they think they do.
Its real value appears when you separate winners from losers. A trader holding losers longer than winners has a problem that is identified, quantified and fixable, and no other statistic shows it as directly. It is the one reading of this term that changes anything.
Average time in trade = Σ (close − open) ÷ number of closed trades
Still-open positions are excluded, otherwise their duration would grow on its own and skew the average at every consultation.
The duration counted is clock time, not market time: a trade held from Friday evening to Monday morning counts all 64 weekend hours.
The average says nothing about the distribution. Two mixed styles produce an average matching no real trade, and that is this statistic's main flaw.
| What you compare | What it reveals |
|---|---|
| Winners longer than losers | The healthy case. You let what works run and cut what does not. |
| Losers longer than winners | The most frequent and most expensive signal: hope on losers, haste on winners. |
| Very tight durations | A method executed mechanically. Rare, and generally a good sign. |
| Two separate populations | Two methods in the same journal. They have to be measured separately. |
| An average rising month after month | Trades are being held longer and longer, often without it being decided. |
The second row is the only outcome this term genuinely serves to produce. Holding losers longer than winners is the most widespread behaviour in trading, and it is almost always involuntary: you cut a gain to secure it, you leave a loss because it might still come back. The figure shows it without having to interrogate your emotions.
The fourth row is the one that invalidates the average. When two populations coexist, the average describes neither, and continuing to read it amounts to following a figure that corresponds to nothing. In that case the only way out is to separate the two methods in the journal and measure each on its own.
The average announces 6h 12m, and almost no trade lasted that. The diagram at the top of this page shows why: there are two distinct populations, very short trades and very long ones, and the average lands in the gap between them. It describes behaviour this trader never had.
This is the simplest use and the most often surprising. A trader who describes themselves as intraday and whose average exceeds a day is not doing intraday, whatever their intentions. The statistic settles a question introspection does not.
The gap is rarely fixed by willpower, and often by a setting: a closer target, a scheduled exit at the end of session, an alert. But you have to know the gap exists.
This is the only use that produces a correction. Compare the average duration of your winning trades with that of your losers: if the latter are longer, you have an expensive and perfectly identified habit, which is far easier to treat than a vague unease.
The correction is mechanical rather than psychological: a stop that does not get pushed solves half the problem, and a target that does not get pulled in solves the other.
The central flaw is that the average assumes a single population. Almost no trader has one: you scalp an open and hold a swing on another market, and the average of the two describes nothing. It is the most frequent case, and also the one tools display without ever flagging it.
The second is clock time. A trade opened Friday evening and closed Monday morning lasted three days for the calculation, while the market was shut two thirds of that. On swing trading, this bias mechanically inflates every duration and makes cross-market comparisons misleading.
The third is reading it without splitting by result. A global average says nothing actionable; the ratio between the duration of winners and that of losers says everything. Most journals display the first and not the second, leaving the reader with a descriptive figure instead of a diagnosis.
The classic mistake. Reading the average when the distribution has two humps. A trader who scalps in the morning and holds swings shows an average of several hours while none of their trades lasted that. The figure is arithmetically correct and false as a description, which is the worst possible combination.
| Indicator | Formula | What it measures | When to use it |
|---|---|---|---|
| Average time in trade | Σ (close − open) ÷ closed trades | The typical duration of a trade | Checking your real style |
| Winner vs loser duration | The two averages, separately | An exit habit, quantified | Fixing the habit of holding losses |
| Exit efficiency | gain banked ÷ gain available | What you capture of the move | Knowing whether you exit too early |
| Best / worst day | the extremes of daily P&L | Your outlier days | Spotting a day without discipline |
Count how many of your trades last under five minutes, under an hour, over a day. If two groups appear, the average is unusable and they have to be treated as two distinct methods. That count takes ten minutes and avoids months of conclusions drawn from a figure describing nothing.
It is the only reading that produces a correction. If your losers last longer than your winners, you hold losses and cut gains, and the correction is mechanical: do not push a stop, do not pull in a target.
An average climbing month after month signals that you are holding positions longer and longer, almost always without having decided to. The absolute value does not mean much, the drift does.
It is the mean duration between opening and closing your closed trades. It describes your execution style: a few minutes for scalping, a few hours for intraday, several days for swing.
Because you probably have two styles in the same journal. An average between ten-minute scalps and three-day swings lands in the gap between the two groups and describes no real trade. In that case the two methods have to be measured separately; the global average teaches nothing.
That is the healthy case, yes, and the reverse is the most widespread behaviour in trading. Cutting a gain to secure it and leaving a loss because it might come back produces exactly the opposite profile, and the correction is mechanical: do not push a stop, do not pull in a target.
Yes, the duration is clock time. A trade opened Friday evening and closed Monday morning counts three days, while the market was shut two thirds of that. On a longer style this bias inflates every duration the same way, which makes it bearable as long as you do not compare markets with different opening hours.
Not much. Taken alone, it describes without indicating anything. Crossed with the result, splitting winners and losers, it points at a precise exit habit and a concrete correction. That is the difference between an inventory statistic and a diagnostic one.
Tradoshi computes average time in trade from the open and close times returned by your broker, with nothing to type in. It appears in the Track record widget on the dashboard, next to best day, worst day and monthly returns.
The reading that matters is in the Reports screen, and it is rare: average duration exists there in three versions, Average duration, Average duration (winners) and Average duration (losers). That is exactly the comparison described above, available without computing anything, and the gap between the last two columns is the diagnosis.
To go deeper on this: read the full guide. Every term is gathered in the glossary.