forex-scalping.com

A strategy enters a losing stretch. The question arrives: is it broken, or is this normal variance?

The answer lies in pre-defined criteria. Before trading, compute the expected range of drawdown and losing streaks from historical data or simulation. A stretch within that range is variance. A stretch beyond it, such as a drawdown 50 percent deeper than the historical worst, signals a problem.

Other tests: has the market changed in a way that affects the logic, such as a shift in volatility or spreads? Has execution quality deteriorated? Are trades being taken exactly as designed?

Changing parameters after a few losses is almost always premature. Stopping the strategy entirely after crossing a defined threshold is reasonable.

A review checklist, completed in calm conditions, replaces the emotional judgement of the moment.

Where is the line between variance and failure for the main strategy, and was it set before trading began?

I learned this one the expensive way. A few years back I rewrote the exit rules on a range strategy after something like eleven losing days, and the version I replaced it with did worse than the original over the following quarter. When I finally went back and checked, that bad stretch was well inside what the backtest had produced several times before.

Now I write two numbers on a card before a strategy goes live. The first is the deepest drawdown from a thousand shuffled runs of the test trades, and the second is the longest stretch of days below the equity peak. If live results stay inside both, I keep trading at normal size and only make notes. If one of them is breached, I cut to half size and do a full review over the weekend, not mid-session.

The checklist idea is good, and I would add one honest question to it: am I still taking every signal exactly as defined? Twice now the strategy looked broken and the real problem was me filtering trades by mood.