forex-scalping.com

The slow bleed of a strategy that used to work

Strategies decay. A scalping approach that earned steadily for a year starts to flatten, then drifts down. Nothing dramatic happens. No single event explains it.

Several causes tend to hide behind the decay. Liquidity providers update their algorithms, so the micro-patterns that once repeated shift slightly. Brokers change routing or pricing, and a quarter-pip of extra cost appears. Competition grows: other traders discover the same edge, and the edge gets crowded.

The slow nature of the decline makes diagnosis hard. A normal drawdown looks the same as a structural decline for the first few weeks.

A few checks help separate the two. Compare the average cost per trade now with the cost a year ago. Costs that crept up explain a lot. Compare the distribution of entry slippage, win rates by hour, and average trade duration. A shift in any of them points to a changed environment.

Another approach is to keep a small benchmark: a simplified version of the strategy that has not been altered, running on small size. If the benchmark has declined along with the main system, the market changed. If only the main system declined, the cause is probably something in its own parameters or management.

Some decay is permanent. Some reverses when conditions return. Telling them apart takes data, and data takes record-keeping, which is why the boring habit of logging everything pays off in the end.

Keeping a dated changelog helps too. Every parameter change, every broker change, every infrastructure move gets a line with the date. When performance shifts, the log provides candidate causes to check first. Without it, the mind reconstructs history from memory, and memory is generous about which changes mattered. A decline that starts the week after a VPS migration is a very different problem from one that begins the week after a rule change.

What measurable change, in costs or fills or timing, preceded the last decline in performance?

Slow markets and active traders don't mix well. When nothing is happening, the temptation to make something happen grows. Lowered thresholds, more pairs, extra signals, looser filters.

The outcome is predictable. Trade count rises while quality falls. Spreads in quiet hours are often wider relative to the available range, so each trade costs more. Edges, if present, shrink or disappear. The net effect is a drain.

Overtrading isn't only about volume. It also shows up as constant switching between timeframes, adding new indicators, and looking for justification. Each is a way to feel productive.

A few measurements make the problem visible. Count trades per hour and compare with results per hour. Many scalpers find a handful of hours produce nearly all of the profit, and the rest produce a slow bleed.

A pragmatic rule is to cap trades per hour or per session and to cut off trading in periods when the log shows no edge. Silence in the market is information, not a defect to be fixed.

One option is a trade budget for each hour. If the plan allows four trades per hour, the fifth is off limits, no matter how good it looks. In slow periods, the budget is rarely used up, and the idle time is fine. The budget only bites on the days when overactivity is the problem.

Which hour of the day contributes the most losses in the trade log, and what is the trade count in that hour?