How do I tell strategy decay apart from a normal losing streak?

Check the rolling 20-trade metrics against the original test baseline. A normal losing streak inside a valid edge shows a win rate near baseline with a run of below-average outcomes. Decay shows a sustained shift in win rate, excursion depth, net expectancy or execution gap that does not revert after 20-30 more trades.

Can a strategy decay and recover later?

Yes, particularly regime-driven decay. A trend-following gold strategy that struggles during a compressive range can become profitable again once volatility expands, without any rule change, which is why regime tracking matters more than abandoning the rules outright.

How often should I re-test a gold strategy from scratch?

A full re-test is warranted when the quarterly review shows the underlying real-yield narrative or volatility regime has genuinely reversed, or when three consecutive monthly reviews show unexplained deterioration across at least two of the four early-warning metrics.

Does broker cost creep really cause meaningful decay?

On tight-stop gold strategies, yes. A spread moving from $0.18 to $0.30 on a $2 stop moves the cost from 9% to 15% of planned risk, which is enough to turn a marginal edge negative even if the underlying pattern recognition is unchanged.

Is crowding a real phenomenon on retail gold levels?

It is difficult to prove directly but consistent with the shrinking excursion depth and lengthening time-to-target seen on heavily publicised levels like prior-day highs and round numbers, compared with idiosyncratic structure levels that show no such drift.

What is the single most useful early-warning metric?

The intended-versus-executed gap, because it is the only metric that separates a market problem from a discipline problem, and discipline problems are both the most common cause of apparent decay and the cheapest to fix.

WHY GOLD STRATEGIES DECAY

Gold trading strategies decay for five identifiable reasons: a change in the volatility regime the stop and target were calibrated to, a shift in the real-yield narrative that quietly biased directional performance, crowding of a publicised level or pattern, broker cost creep in spread or swap, and gradual drift by the trader away from the rules that were originally tested.

Each cause leaves a distinct signature in a properly kept journal: regime change shows in ATR and time-to-target, narrative shift shows in a growing long-versus-short expectancy gap, crowding shows in shrinking excursion depth, cost creep shows in a widening gross-to-net expectancy gap, and rule drift shows directly in the comparison between intended and executed entries, stops and size.

A monthly review cadence — weekly drift checks, monthly metric reviews against a rolling 20-trade window, and quarterly macro reassessment — catches decay while it is still a sample-size question rather than after a drawdown has forced the decision, and prevents traders from discarding a genuinely valid edge over what is actually ordinary variance.