A trading system can produce convincing results for weeks, months or even years, and then enter a period of difficulty that appears inexplicable. This is more common than most traders assume, and it happens to well-prepared operators, not only beginners. This article covers why it happens and how to respond.
First, Rule Out the Obvious Explanation
Before asking why a trading system stopped working, you have to eliminate the most mundane possibility: the system was never as good as it appeared.
Sometimes the problem is not that a system has lost its effectiveness, but that it never had enough of it. A strategy can look profitable over a limited period, particularly if it was tested on scarce data, on a particularly favourable market phase, or on conditions unlikely to repeat.
To rule that out properly, it helps to define what a good trading system actually looks like.
The Characteristics of a Sound Trading System
The first is clarity of rules. A system cannot rest on vague guidance like enter when the market looks strong or exit when the move loses energy. Those formulations may work as intuitions, but they are not enough to build a repeatable method. A solid system defines entry conditions, exit conditions, stop loss management, targets, position size and exclusion criteria.
The second characteristic is verifiability. A trading system has to be testable, measurable and analysable. That does not mean trusting a backtest blindly, but it does mean collecting enough data to establish whether the method has a statistical edge.
The third is compatibility with the trader. A system can be good in the abstract and unsuitable for the person using it. A fast intraday strategy may be perfectly valid and completely unmanageable for someone who cannot watch the market continuously. Equally, a trend-following system may be effective over the long run but hard to sustain for someone who cannot tolerate extended sideways periods or long strings of small losses. Our overview of the main types of trading system covers which profiles suit which approach.
A sound system also needs realistic risk management. If a strategy only works with excessive leverage, overly wide stops or disproportionate position sizes, the problem is not merely operational — it is structural. A valid system has to allow survival through the negative phases, because no method avoids drawdown entirely.
Finally, a serious system should be simple enough to be applied with discipline. Simple does not mean trivial; it means comprehensible. If a strategy requires too many conditions, too many indicators and too many exceptions, the trader eventually loses the ability to tell when a signal is genuinely valid. In those cases the system often looks sophisticated while actually being fragile — a distinction covered in detail in our guide to spotting a fragile trading strategy.
Reason 1: The Market Regime Changed
Every trading system is built, more or less explicitly, to exploit particular conditions. Some methods work well in directional markets, others in ranging ones. Some benefit from high volatility, others become more effective when moves are orderly and progressive. The problem arises when the market changes character.
An example makes it concrete. A trend-following system can produce excellent results during a directional phase in FX, commodities or indices. When the market moves into congestion, however, the same system starts accumulating false signals, late entries and repeated stop losses. That does not necessarily mean the system is wrong. It means the context in which it performed best is no longer present.
This distinction matters enormously for how you respond. A system underperforming because its regime has gone should be reduced in size or paused, not rebuilt. A trader who rewrites the rules in response to a regime change typically ends up with a system optimised for the conditions that are about to end.
Reason 2: Overfitting
The second reason is overfitting, or excessive optimisation. This problem usually originates in the construction phase. The trader tests many parameter combinations until finding the one that would have worked best in the past. The result can look excellent: a smooth equity curve, contained drawdown, highly convincing statistics. But that result may have been obtained by fitting the system too closely to the historical data.
In practice, the system has not learned a repeatable market logic, it has memorised the past. Applied to new conditions, it loses effectiveness because the rules were calibrated on details with no genuine predictive value. This is one of the most insidious risks for intermediate traders in particular, precisely because it arrives with the tools that mark progress: backtesting software, optimisation routines and advanced platforms. Our article on overfitting in algorithmic trading and AI covers how the same failure scales up when models get more powerful.
The diagnostic signature of overfitting is specific: the system degrades immediately and permanently once it leaves the data it was built on, rather than degrading in one market condition and recovering in another. If live results have been worse than backtest results from the very first month, this is the likely cause.
Reason 3: The Trader Changed, Not the System
The third reason is a change in the trader rather than in the method. Blame usually lands on the system rather than on the person operating it. But a trading system can stop working simply because the trader has started applying it differently: anticipating entries, moving stops, skipping signals after a losing run, increasing size to recover, closing profits too early, or taking trades the rules never authorised.
This is the most common cause of all, and the hardest to see from the inside, because each individual deviation feels reasonable at the time. Nobody decides to abandon their system. They make a series of small, defensible exceptions that collectively add up to trading a different strategy than the one that was tested.
How to Diagnose Which One It Is
So how should you respond to a trading system that has stopped working?
The first and most important step is keeping a detailed trading journal. Recording entry, exit and result is not sufficient. You need to note whether the trade respected the rules, whether the size was correct, whether the stop was planned in advance, and whether the trade belonged to the intended setup at all. Only that level of detail lets you distinguish between a strategy in difficulty and a strategy betrayed by its execution.
With that record in place, the diagnosis becomes tractable. Separate your trades into those that followed the rules and those that did not, and calculate the results for each group. If the rule-following trades are still performing and the deviations account for the losses, the system is fine and the execution is the problem. If the rule-following trades have also stopped working, then it is either a regime change — check whether the market conditions the system needs are actually present — or overfitting, which you can test by looking at whether performance ever matched the backtest.
The response follows from the diagnosis, and the three responses are genuinely different. Execution drift is fixed by process and discipline, not by changing the strategy. A regime change is handled by reducing exposure and waiting, or by adding a filter that recognises the regime. Overfitting requires rebuilding the strategy on a sound logical basis, and it is worth running the rebuilt version through the scenarios in our guide to stress testing a trading strategy before committing capital to it again.
What all three have in common is that they cannot be diagnosed from the equity curve alone. A declining equity curve looks identical in all three cases. The journal is what makes them distinguishable — which is why keeping one is less about self-improvement and more about basic instrumentation.
