XAUUSD EA Holidays: Why Your System Fails During Market Closures

Why EAs Keep Running During Holidays by Default

An EA evaluates trading conditions based on price data, indicators, time, and spread. It has no concept of a calendar holiday unless the developer specifically built calendar awareness into the code. Building this is technically complex — holidays vary by jurisdiction, some fall on different dates each year (Easter, Thanksgiving), and different brokers have different holiday schedules. Most EA developers make a practical decision: leave holiday management to the trader.

This is actually a reasonable design choice. Automating holiday detection would require maintaining a holiday database, handling jurisdictional variations, and dealing with cases where your broker is open but liquidity is still degraded. A trader who understands their EA and reviews it weekly is better positioned to make holiday decisions than code trying to generalize across hundreds of broker schedules.

The consequence is that if you don't intervene, your EA will trade during Christmas week with the same session filter and same lot size as it uses during a normal November Tuesday. The conditions are fundamentally different; the EA doesn't know that.

What Thin Markets Do to EA Signals

The way holiday thin markets degrade EA performance is more subtle than just "spreads are wider." The underlying price behaviour changes too. With fewer institutional participants, price movements are driven more by smaller orders. Trends are less sustained. Breakouts are less reliable — price can spike through a level and immediately reverse because there isn't enough volume behind the move. For a breakout-based EA, this means more false signals.

The spread filter catches some of this: entries that would have triggered during wide-spread conditions get rejected. But the entries that do pass the filter are still in a degraded environment. The win rate typically falls during holiday weeks for strategies that rely on sustained directional moves, because the market structure that creates those moves (institutional order flow, deep liquidity supporting price continuation) is absent.

Historical evidence for this is usually in individual trade logs rather than published backtests, because the standard backtesting data doesn't accurately model the spread widening and execution degradation of live holiday trading. If you've been trading gold EAs for more than a year, look at your trade history for Christmas week 2024 vs. a comparable week in October 2024. The comparison is usually informative.

How to Implement a Holiday Pause

The mechanics of pausing are simple. When you want to pause the EA before a holiday period: right-click the chart the EA is on, select "Expert Advisors" from the context menu, and remove the EA. This stops all new trade entries immediately. Any currently open trades remain open with their stops and limits active — pausing the EA doesn't affect existing positions.

After the holiday: reattach the EA using the same parameters (this is why a settings backup is important — you need to restore the same configuration). The EA will start its normal session evaluation from fresh and will begin trading again during the next valid session window.

An alternative to full removal: change the EA's lot size to 0.01 (the minimum) before the holiday period. This way the EA continues running and you can see how it's behaving, but with negligible real risk. After the holiday, restore to your normal lot size. This is more of a monitoring approach than a genuine pause — useful for learning holiday patterns without significant exposure.

The Spread Filter as a Partial Defence

A well-configured spread filter provides real but incomplete protection during holidays. Goldie Razor V2.8.4 uses an 18-pip maximum spread filter. During a typical Christmas week where spreads regularly reach 25–40 pips, the filter will reject the majority of would-be entries. This means the EA mostly sits idle during the worst conditions — which is the correct behaviour.

The incomplete part: spread widens are not constant. During a Christmas week, spread might be 30 pips at 10:00 UTC, then briefly tighten to 12 pips for 15 minutes before widening again. In that 15-minute window, the EA's spread filter is satisfied and it may enter a trade. The position then faces the rest of the trading day in thin market conditions. So even with a good spread filter, a manual pause is the more reliable protection during the highest-risk periods.