A strategy can be profitable on paper and still be awkward to run because it keeps risk open too long. Average trade duration is the backtest metric that tells you how long each position usually stays alive, and for a prop trader that affects costs, drawdown path, weekend exposure, and rule compatibility.
The common mistake is reading duration as a footnote. It is not. Duration is where a clean trade list turns into real exposure.
What average trade duration means
Average trade duration is the mean time between entry and exit across the closed trades in a backtest. A system with a four-hour average is not the same kind of risk as a system with a four-day average, even if both show the same net profit and max drawdown.
The number is simple. Its interpretation is not. It tells you whether the strategy behaves like a fast intraday engine, a swing system, or something in between.
| Duration profile | What it usually implies | What to inspect next |
|---|---|---|
| Minutes to hours | More dependence on spread, commission, slippage, and execution timing | Cost realism and trade count |
| Several hours to one day | More exposure to session changes and daily loss clustering | Intraday equity lows and bad-day behaviour |
| Multiple days | More exposure to swap, weekend holds, news gaps, and prop account type | Swap treatment and Swing vs Standard compatibility |
Why duration changes the cost picture
A short-duration strategy often pays the market more often. A long-duration strategy often pays the market for longer. Different problem, same point: cost assumptions matter.
For a fast strategy, spread, commission, and slippage can dominate the edge because the average profit per trade has less room to absorb friction. For a swing strategy, swap becomes harder to ignore because overnight financing is part of the trade, not an accounting detail added later.
That is why a backtest that does not state duration and costs together is incomplete. The question is not "did the curve rise?". The question is "did the curve rise after charging the costs that its holding period actually creates?".
realbacktesting's public method charges real per-symbol spread, real per-symbol commission, swap, and 1 bps slippage on cTrader broker data from 2021-2026, with execution modeled on intrabar M1 data and an 80,000 EUR model base. The exact methodology is documented on the methodology page. It does not remove uncertainty. It removes a common hiding place.
For a related failure mode, why the same cTrader backtest changes across brokers is worth reading next.
Why duration changes prop-firm risk
Prop-firm risk is path risk. A position that stays open longer has more time to pass through floating loss, news, spread changes, swap, and session transitions before it closes.
Two systems can have the same closed-trade result and still feel very different inside the account:
| Backtest summary | System A | System B |
|---|---|---|
| Closed trades | 100 | 100 |
| Average risk per trade | 1% | 1% |
| Average trade duration | 2 hours | 2 days |
| Main concern | Fast clustered losses | Long floating exposure |
Neither profile is automatically better. The right question is which one fits the account rules. A fast system may run into the daily loss limit if several trades lose in the same session. A slower system may stay within closed-trade limits while spending more time in floating drawdown.
That is why duration belongs beside drawdown, not underneath it. If you only inspect the final max drawdown number, you may miss when the pain happened and how long it stayed open. Balance vs equity drawdown for prop traders explains that difference directly.
Duration and account type must agree
A strategy that regularly holds for days is not compatible with every prop account type. If the rule set restricts weekend holding or news trading, the backtest has to match that rule set before the result means anything.
This is the practical reason duration matters for FTMO-style accounts. A strategy that holds through the weekend needs a rule set that allows that behaviour. A strategy that exits before the session closes has a different problem: whether its short holding period still survives real execution costs.
The mistake is treating account type as a setup detail. It is part of the test. If a backtest assumes weekend holds and the trader later runs it on an account that forces Friday closures, the tested strategy and the traded strategy are no longer the same thing.
The prop-account framing is laid out on the funding model, and why FTMO Swing vs Standard changes your backtest covers this specific mismatch.
A simple way to read duration in a backtest
Average trade duration becomes useful when you combine it with trade count, costs, and the equity path. By itself, it is only a clock.
Use this checklist:
- Compare duration with cost assumptions. Shorter average duration demands sharper spread, commission, and slippage treatment. Longer average duration demands explicit swap treatment.
- Compare duration with the equity path. Long floating losses matter even if the closed trade finishes green.
- Compare duration with trade clustering. Ten trades held for 12 hours each and 40 trades held for 3 hours each both create 120 trade-hours of exposure, but the clustering risk is different.
- Compare duration with the prop rule set. Weekend holds, news restrictions, daily loss limits, and max loss rules change what the same trade list means.
- Compare duration across market regimes. A strategy that held trades for hours in one regime and days in another may be more regime-dependent than the average suggests.
The last point is where averages can mislead. A single average can hide a split personality: fast winners, slow losers, or a few very long positions that carry most of the drawdown. The trade-duration distribution is more useful than the headline mean when you can get it.
What duration does not prove
Average trade duration does not tell you whether a strategy has an edge. A bad system can hold for minutes. A bad system can hold for weeks. The clock does not know whether the logic is good.
It also does not replace out-of-sample testing. A strategy may show a stable average duration in-sample and then behave differently when the market regime changes. That is why the duration question has to sit inside a broader verification process: realistic costs, held-out data, and a path-risk check.
realbacktesting is a trading-software studio for cTrader built around that broader standard: every published cBot backtest number is the cBot's own cTrader-native result, reproducible in your own cTrader, validated on real broker data with out-of-sample checks. For the wider verification workflow, start with how to verify a cTrader backtest.
Frequently asked
Is lower average trade duration better?
No. Lower average trade duration means risk is usually held for less time, but it may also mean more trades, more cost sensitivity, and more execution dependence.
Is higher average trade duration bad for prop traders?
Not automatically. Higher duration becomes a problem when the strategy's holding period conflicts with the account rules, ignores swap, or hides floating drawdown that the prop firm measures.
Should average duration be measured in trades or days?
Both views help. Per-trade duration shows the typical holding period of a position, while day-level exposure shows how much risk the account carried through time.
Can a backtest with good duration still fail live?
Yes. Duration is only one diagnostic. A live account can still differ because of broker costs, execution, market regime, rule changes, and ordinary uncertainty.
The stubborn takeaway
Average trade duration is not a side metric. It is the bridge between the trade list and the account that has to survive it.