The most common EA mistake is pausing for the wrong reasons. Flip each card to reveal the verdict — PAUSE, DO NOT PAUSE, or REDUCE — with exact instructions.
Click each card to reveal the verdict
"Pause for infrastructure issues. Trust the EA for market conditions."
You should pause your gold EA when an infrastructure condition makes automated trading unsafe — not when market conditions are unfavourable. The key distinction: infrastructure problems (broker maintenance, news events causing extreme spread widening, connectivity issues) are legitimate pause triggers because your EA cannot function correctly. Market conditions like drawdown, volatility, or poor recent performance are not pause triggers because the EA's edge still applies — the problem is usually the trader's psychology, not the EA's performance.
The core question to ask before pausing any EA: "Is this a problem with how the EA can operate, or a problem with how the EA is performing?" These two categories require completely different responses.
If the EA cannot operate correctly — because spread is 50 pips during news, because the broker is in maintenance, because your VPS connectivity is unstable — then pausing makes sense. The EA's logic cannot function as designed in these conditions.
If the EA is operating correctly but performing below recent expectations — three losing days, a 10% drawdown, lower win rate this week — then pausing is almost certainly wrong. The EA's logic still applies. The issue is either normal statistical variance (losing streaks happen even for profitable strategies) or a fundamental change in market conditions that requires investigation, not an immediate pause. This distinction is the foundation of all discipline around the pause decision.
FOMC announcements, NFP releases, CPI data — these scheduled events cause spread widening of 3–10× normal levels in the minutes before and after the release. XAUUSD spread on an ECN broker normally runs at 0.8–1.5 pips. During NFP, it can reach 30–80 pips. A scalping EA targeting 15 pips profit cannot function profitably when entry spread alone is 30 pips.
This is a pure infrastructure issue — the EA's logic is sound, but the execution environment makes it impossible to implement that logic at normal cost levels. Pre-schedule pauses 45 minutes before major events using calendar reminders. Review the week's economic calendar every Sunday. For broader timing guidance, the guide on pre-scheduled no-trade windows vs reactive pausing covers how to set these up systematically.
Brokers periodically perform server maintenance — typically announced 24–72 hours in advance. During maintenance windows, order processing is unavailable or unreliable. An EA with open positions during maintenance cannot manage those positions: cannot move stops, cannot take profits, cannot exit. This is genuine unmanaged risk.
Before any scheduled maintenance: (1) close all open positions manually, (2) disable Algo Trading, (3) wait until maintenance is confirmed complete (not just scheduled to end — confirm via broker support chat or MT5 connection indicator), (4) re-enable Algo Trading 30 minutes after confirmed completion.
Christmas Eve, Christmas Day, New Year's Eve, US Thanksgiving, and Easter Monday reduce market participation by 40–70%. The remaining liquidity comes disproportionately from retail traders rather than institutional players, creating erratic spread patterns and false signal generation. An EA trained on data from normal institutional sessions will generate signals that do not apply to holiday market microstructure.
The safer option: fully pause on major holidays. The reduced-risk option: halve the lot size and set tighter spread filters. The profit potential on a holiday with thin liquidity rarely justifies the added risk.
If your VPS has connectivity problems, your EA cannot communicate with the broker. Open positions become orphaned — the EA cannot apply its exit logic if price moves against you. If you notice your VPS is experiencing instability (MT5 not reconnecting, logs showing disconnection errors), the safe move is to log into your broker platform directly and manually manage any open positions until VPS stability is restored.
If your EA does not have a built-in MaxSpread filter and you notice XAUUSD spread is currently at 25–50 pips (visible in the MT5 Market Watch), temporarily pausing until spread normalises is sensible. However — and this is important — if your EA already has a MaxSpread parameter set correctly, you do not need to manually intervene. The EA's filter handles this. Trust the filter and consider whether a manual intervention is actually adding value or just feeding your anxiety.
This is the most important pause guidance, and the most frequently ignored. When an EA enters a losing streak — 3, 5, or even 8 consecutive losses — the psychological pressure to intervene is intense. The losses feel like evidence that something is wrong. They feel like a signal to stop.
They are not. Consecutive losses are a mathematical certainty for any strategy, including those with 65–70% win rates. A strategy winning 65% of the time has a 4% chance of losing 5 in a row. Over 200 trades, you will have multiple such streaks. They are not evidence of EA failure — they are evidence of statistical reality.
The traders who pause during drawdown make three compounding errors: they lock in the losses at that point, they miss the recovery trades that follow (which restore the account to expected equity), and they then face the psychological difficulty of knowing when it is safe to re-enter — which they almost always time incorrectly, either re-entering after more losses (too late) or not re-entering at all.
The one exception: if drawdown exceeds the EA's historical maximum drawdown by more than 30%, that is a signal worth investigating. Not pausing immediately, but investigating whether broker conditions have changed, whether parameters need adjustment, or whether a fundamental market regime shift is happening. For this deeper investigation, the guide on recovery mode as an alternative to pausing provides additional decision frameworks.
There are three ways to pause an EA in MT5, and they have meaningfully different effects:
Pausing and turning off an EA are fundamentally different decisions with different protocols:
| Dimension | Pausing | Turning Off |
|---|---|---|
| Duration | Hours to days | Weeks to months |
| Trigger | Specific known event | Fundamental issue found |
| Analysis required | None — predefined rule | 4-branch diagnostic first |
| Resume condition | Known: after event ends | Unknown: after issue resolved |
| Appropriate when | Infrastructure problem | EA fundamentals in question |
When you think you need to "turn off" your EA, the first step is always to ask whether a temporary pause would be more appropriate. If you cannot identify a specific date when you would re-enable the EA, that is a sign you're considering turning off rather than pausing — which requires the more thorough analysis described in the guide on when pausing is not enough.
Duration guidelines for the most common pause triggers:
The goal is to make pausing systematic rather than reactive. Reactive pausing — responding to bad trades in real-time — is almost always emotional. Systematic pausing — following a pre-built protocol — is almost always rational.
Here is a weekly protocol that takes 15 minutes on Sunday and prevents 90% of problematic pause decisions:
This simple system removes in-the-moment decision-making entirely. You follow the plan. The plan is made when you are calm, not when you are watching a loss. This is the core insight of volatility situations and whether to pause — having a predefined rule for every situation means you never need to decide in real-time.
Goldie Sniper, Goldie Razor V2 and V2.8.4, Blind Sniper, and Hybrid Scalper. Full suite for every market condition.
Not sure which EA fits your account size or trading style? Email us and we will help you choose.
Goldie Razor V2.8.4
M15 breakout + H4 EMA filter — built for XAUUSD on MT5