An equity curve filter is a rule that changes a strategy's size or pauses it after the strategy's own recent losses. It can make a prop backtest easier to survive, but it is not evidence unless the filter itself survives the same scrutiny as the entry rule.
The appeal is obvious. A strategy hits a rough patch, so the filter reduces exposure until conditions improve. The catch is less obvious: the filter sees the same noisy past that tempted the trader to build it. It can turn a bad period into a smaller loss. It can also sit out the recovery that paid for the original strategy's edge.
What an equity curve filter actually changes
An equity curve filter is an account-level overlay, not a new entry signal. It uses the strategy's realised equity, drawdown, or recent results to decide whether the next eligible trade is taken at normal size, reduced size, or not at all.
That distinction matters in a prop backtest. The filter changes the sequence of exposure, so it can change daily loss pressure, drawdown duration, and the final result even when the underlying entries never move. A test that reports only the filtered curve has hidden the decision that did the changing.
| Filter decision | What changes | What must remain visible |
|---|---|---|
| Trade normally | No overlay intervention | The original strategy result |
| Reduce size | The next loss and the next gain both shrink | The sizing rule and trigger |
| Pause trading | Eligible trades are skipped | Every skipped trade and re-entry condition |
Why prop traders are drawn to the idea
Prop rules make the path matter. A system can end a long sample ahead and still create a run of losses that the account cannot tolerate along the way. An equity curve filter is often an attempt to make that path less violent.
There is nothing irrational about testing that idea. There is also no free pass. If a filter reduces the worst decline, it may do so by giving up some recoveries or by moving losses into a different part of the sample. The relevant comparison is not “filtered curve looks calmer.” It is whether the same predeclared overlay remains intelligible across the full history and untouched validation.
For the account context, keep the current firm rulebook beside the test. The funding model and get-funded account context explain why floating equity and the route through a drawdown matter as much as a final total. Firm rules can change; their own documentation is the final reference.
Test the overlay as a separate hypothesis
The clean way to test an equity curve filter is to write its rule before looking at the filtered result. Record the input, threshold, action, and return condition. Then preserve a baseline run in which the underlying strategy is never filtered.
Keep the trigger mechanical
“Pause when the curve looks weak” cannot be reproduced. A usable test rule states exactly what is measured and when the decision is made. For example, it might use a defined drawdown from the strategy's prior equity high, evaluated only after a completed trade. The point is not that this particular choice is superior. The point is that another trader can run the same rule without guessing what “weak” meant.
Compare path metrics, not only the finish
The overlay should be judged against the baseline on the metrics the account actually experiences. Inspect maximum floating-equity drawdown, worst daily loss, time spent below a prior equity high, number of skipped trades, and the distribution of results across different windows.
A small recovery calculation makes the asymmetry clear. A 20% drawdown needs a 25% gain to return to the prior high. A 50% drawdown needs a 100% gain. Reducing drawdown can be valuable; the test still has to show what was sacrificed to achieve it.
Hold the filter out of sample too
Choosing a trigger after inspecting every possible threshold is optimisation, even if the rule is called “risk management.” The filter has parameters, and parameters can memorise a sample.
Freeze the overlay before the untouched segment is examined. Then evaluate baseline and filtered versions on that segment with the same costs and execution assumptions. Backtest acceptance criteria are useful here because they force the failure conditions onto paper before the cleaner curve gets a vote.
The common ways an overlay flatters itself
An equity curve filter is easiest to overstate when its counterfactual is missing. These checks make the trade-off visible.
| Check | Why it matters |
|---|---|
| Show the unfiltered baseline | Reveals whether the overlay improved risk, return, neither, or only the chart |
| Keep skipped trades in the ledger | Shows whether the filter avoided losses or missed recoveries |
| Test a small set of predeclared settings | Limits the search for the one flattering threshold |
| Inspect multiple historical windows | Exposes dependence on one fortunate sequence |
| Include realistic costs in both runs | Stops the overlay from receiving a frictionless comparison |
The last point is mundane but important. If the base strategy and overlay are not run with identical data, execution, spread, commission, swap, and slippage assumptions, their difference is not attributable to the overlay. The methodology page and proof of the test conditions are the useful standard: make the conditions visible enough that a skeptical reader can challenge them.
What an equity curve filter cannot prove
A calmer historical curve does not prove that the next drawdown will arrive in a recognisable pattern. Equity filters react after losses have appeared. They do not know whether a difficult patch is temporary noise, the beginning of a regime change, or the first part of a recovery.
That limitation is exactly why the baseline remains essential. A filter can be a reasonable hypothesis about exposure management. It is not a licence to discard the underlying strategy's bad history or a promise that a prop account will survive.
realbacktesting is a trading-software studio for cTrader built around test conditions a trader can inspect. Its published cBot figures are cTrader-native backtests that a buyer can reproduce in their own cTrader; that is evidence to check, not a live track record or a performance warranty.
Frequently asked
Does an equity curve filter create an edge?
Not by definition. It changes when and how much a strategy trades after its own results; whether that improves a result is a testable claim, not a property of the label.
Should a filtered backtest replace the original strategy result?
No. The original and filtered runs answer different questions. Showing both is the minimum needed to see the cost of the overlay and to reproduce its effect.
Can a filter reduce prop-firm drawdown risk?
It can reduce historical drawdown in a particular test, but it can also miss profitable recoveries. Compare the two paths under the current firm rules and do not treat the historical reduction as a guarantee.
The stubborn takeaway
An equity curve filter does not make a strategy safer because it turns off after pain. It becomes credible only when the unfiltered curve, the skipped trades, and the out-of-sample result are all still allowed to disagree with it.