Your EA hasn't traded in three weeks. The forum's already telling you it's "abandonware". A 30-day Myfxbook chart looks like a flat line. The vendor isn't replying as fast as you'd like. You start drafting the chargeback email.
Slow down. Half the time, "frozen" is a feature. The other half, it's a real problem — and the way to tell them apart is not what most retail traders think.
I run two EAs that have been "frozen" for stretches in 2026: Gold Guardian (1 trade in the last 30 days) and GBP Master (0 trades this month). Both are still doing exactly what they were built to do. Here's how to read silence in an EA without panicking — and how to spot the silence that should make you panic.
What "frozen" technically means
An EA is "frozen" when it's running, connected to the broker, and not opening trades. The terminal is on, the chart is loaded, the autotrading button is green, the smiley is there. No errors in the journal. Just no trades.
That state can be caused by very different things, only one of which is actually a problem:
- Filtered out: the EA's entry conditions are not met (volatility, session, spread, ATR, news filter, regime classifier, whatever the logic uses)
- Risk-off mode: the EA detected something it doesn't like (recent loss streak, drawdown threshold, equity curve break) and parked itself
- Spread / broker filter: spreads are too wide vs the EA's max-spread setting (very common on cTrader / raw accounts during NY close, weekends, news)
- Configuration drift: set file changed, magic number conflict, broker symbol naming changed (e.g. EURUSD vs EURUSD.r), VPS time zone shifted
- Vendor abandonware: the system was discontinued, the developer disappeared, or the strategy stopped working in current market conditions and nobody updated it
The first three are features. The fourth is a fix-it-in-an-hour problem. Only the fifth is the disaster everyone's worried about.
The vendor lie everyone bought: "trades every day"
Half the EA marketplace is built on a fantasy: that a "good" EA must trade frequently. Daily preferred. Weekly minimum. The marketing copy reinforces it: "Over 500 trades a year!" "Generates 3-5 signals daily!"
That's a sales pitch, not a strategy.
If your edge is real, you only take it when the conditions for it appear. Conditions don't appear on a fixed schedule. London is choppy in August. Gold goes nuclear in March and dead in September. Some weeks the regime is wrong for your model and the right answer is to do nothing.
An EA that trades every day in every market regime is not finding edges; it's manufacturing trades to look productive. That's how you get a 92% win rate with a hidden martingale recovery system that sails through 11 months and detonates in the 12th. The math behind why those systems blow up is well documented.
Selective EAs feel boring. They don't generate clicks. They don't fill Telegram channels. And they're the ones that survive five-year forward tests.
Real example 1: Gold Guardian — 1 trade in 30 days, +6%
Gold Guardian had a long stretch in early 2026 with very few trades. Then on May 7-8, it took one trade. That single trade closed +6.19% on the equity. Total trades this year: 62. Total trades over 14 months live: 187.
That's roughly 13 trades per month average, but not evenly distributed. Some weeks: 6 trades. Some months: 1.
Why? Gold Guardian is built for very specific gold conditions. The model wants a particular volatility-and-trend combination, and when those don't show up, it doesn't take the trade. The 96% win rate isn't from a recovery system — it's from refusing setups that don't match the pattern.
The cost is that some months feel dead. The benefit is that the equity curve doesn't blow up when the dead months end. The 14-month live record is on Myfxbook with every trade visible — including the month where it took 1 trade and the month where it took 14.
Real example 2: GBP Master — 0 trades this month, +165% lifetime
GBP Master is more extreme. It hasn't taken a trade this month at all. YTD it's slightly down (-1.42%). Over the full live record: +165%, 1,179 trades, 92% win rate, profit factor 2.18.
The GBPUSD volatility regime since late 2025 has been compressed. The EA's edge needs a particular kind of GBP movement that just hasn't been there. So it sits.
This is the part the YouTube reviewers don't get: an EA that refuses to trade in conditions where its edge doesn't exist is doing exactly the right thing. The bad EA in the same conditions would force trades and bleed slowly until the strategy "stops working" — except the strategy never worked, the conditions just hid it.
Will GBP regimes return? Historically yes. Could they not? Also yes — that's why it's not the only EA in any portfolio I run.
The portfolio principle that solves this
No single EA trades every regime. Selectivity is what keeps drawdowns small — but selectivity also means dead months. The fix isn't a more aggressive EA. The fix is running EAs with different regime preferences in parallel.
When GBP Master is sleeping, Alpha Pulse AI is taking AI-driven trades on majors. When Gold Guardian is filtering out, the others are still active. Diversified by edge, not by symbol.
8-slot portfolio · regime-diversified · public Myfxbook v2
How to tell selective from broken in 5 minutes
This is the framework I'd give a friend. Run through it before assuming an EA is dead:
1. Check the long-term Myfxbook (3+ years if available)
Are there other multi-week silent stretches in the history? If yes, this is the EA's normal rhythm. If the chart is wall-to-wall trades for 3 years and now 30 days flat, something changed.
2. Check the journal and experts log
Open MT4/MT5 → Toolbox → Experts and Journal. Are there entries showing the EA is alive (initialized, ticks processed)? Are there filter messages ("spread too high", "volatility filter", "session filter")? If yes, the EA is awake and choosing not to trade. If silent or showing errors, you have a configuration problem.
3. Check the broker side
Is autotrading enabled (green button)? Is the symbol the right one (some brokers use suffixes like .r, .pro, .ecn)? Is the spread within the EA's max-spread setting right now? Open the symbol's order window and confirm spread is reasonable.
4. Check the news filter
Many EAs pause around major news. NFP, FOMC, CPI, BoE meetings. If you're in a heavy news week, the filter may be the entire reason for the silence.
5. Check vendor activity
Has the vendor posted updates in the last 30 days? Replied on MQL5 / forum / email? Is the product still listed? If the vendor went quiet at the same time the EA went quiet, that's the actual red flag — not the silent equity curve.
If 1-4 all check out and the vendor is alive, you have a selective EA in a wrong-regime period. That's annoying, not broken.
When silence IS the problem
Real abandonware looks like this:
- Vendor stopped replying (any channel) for 60+ days
- Product page on MQL5 / website removed or deactivated
- The EA's market regime has returned but it still doesn't trade
- Other identical-strategy EAs from different vendors are taking trades; yours isn't
- License server stopped validating (you get an "unable to verify" error)
- Forum complaints from other users with the same product all hit at the same date
That last one is the cleanest signal. If 5 random buyers in different countries on different brokers all stopped getting trades at the same calendar date, the strategy stopped working — not the conditions.
What this means for buying decisions
Stop using "trade frequency" as a quality metric. It's the easiest stat to fake and the most misleading one to read. A 50-trade-a-day EA with 40% win rate and a recovery system is far worse than a 2-trade-a-month EA with real selectivity.
Use these instead:
- Years of live forward record — backtest doesn't count. Live, on a verified third-party tracker (Myfxbook with all privacy off, or FX Blue), with broker name and account numbers visible
- Drawdown history under real money — every visible trough on the curve, not the cherry-picked recent stretch
- Survivorship over regime shifts — did the EA survive the 2024-2025 regime change? The COVID volatility spike? That's the test
- Vendor responsiveness now — message them with a real question and time the reply. If you can't reach them pre-purchase, you definitely won't post-purchase
If you're new to portfolio thinking, start with the free piece
The free USDJPY EA is the simplest entry to multi-EA logic — one slot, public Myfxbook, no upsell required. It's the same architecture used inside MultiStrategy Pro, just with one strategy instead of eight. Build the habit before scaling.
Free · MT5 · public Myfxbook · no email-only paywall
The honest stat about my own products
Right now (May 2026):
- Gold Guardian: 1 trade in the last 30 days, +6% on that one trade. Operating but very selective.
- GBP Master: 0 trades this month. Has not been wrong, has just not had its setup.
- Alpha Pulse AI: 19 trades this week, +2.75%, 36% win rate (pulled back from prior week's run). Active.
If I were marketing instead of telling the truth, I'd hide GBP Master and only show Alpha Pulse. The reason I keep all three on the same equity dashboard is exactly because their dead months don't overlap. That's the whole point.
And if they all did go dead at the same time? Then the right answer is to not run any of them that week — not to switch to a worse EA that pretends nothing's wrong.
Bottom line
Frozen ≠ broken. The EAs you should be most suspicious of are the ones that never stop trading. Real selectivity costs trade frequency in exchange for survivorship. If you can't tolerate dead months, you're going to end up holding a system that bleeds during them — there's no third option.
Look at the long-term curve, check the vendor, check the journal, check the regime. Then decide if you're holding selective infrastructure or a corpse. Most of the time, it's the first one.
FAQ
How long is "too long" for an EA to not trade?
It depends on the strategy. A scalper going silent for 2 weeks is a red flag. A swing EA on a major pair going silent for 6-8 weeks during a low-volatility regime is normal. Always benchmark against the EA's own historical pattern, not a general rule.
Can I force the EA to take more trades?
Sometimes — if the EA exposes "aggression" or filter parameters in its set file. But you're trading off the exact selectivity that's protecting your equity. You're essentially manufacturing trades the model didn't want. Don't be surprised when the equity curve gets worse.
Should I switch EAs if mine is silent for a month?
Not if everything else checks out (vendor alive, journal clean, broker config OK, regime verifiably different). One month of silence in a 3-year live record is signal noise. One month of silence after a vendor stops replying is a different story.
Why do my Gold Guardian and GBP Master have so few trades right now?
Gold Guardian: gold conditions in 2026 have been less consistent for the model's preferred setup pattern. GBP Master: GBPUSD volatility has been compressed since late 2025. Both EAs have multi-year live records showing identical low-trade stretches in the past, followed by recoveries when the regime returned. The selectivity is the design.
What's the difference between selective and broken?
Selective: long-term curve has many flat stretches but trends up overall. Vendor active. Other identical-strategy EAs are also being selective right now. Broken: the curve was always smoothly rising and now isn't. Vendor silent. Other systems on the same strategy are still trading. Use the curve and vendor activity together — neither alone is enough.
Is portfolio trading the answer to dead-month frustration?
Mostly yes. Running 3-8 selective EAs with different regime preferences means at any given week, some are taking trades and some aren't. That's exactly the design behind MultiStrategy Pro v2 — 8 slots, regime-diversified, public live Myfxbook on the v2 forward test. Won't eliminate dead weeks completely, but it makes them rare.
Trading involves real risk to capital. Live forward test results referenced (Gold Guardian +530%, GBP Master +165%, Alpha Pulse +18% YTD) are from real broker accounts on public Myfxbook. Past performance does not guarantee future results. Selectivity protects against losing trades; it does not guarantee winning ones. Never run an EA on capital you can't afford to lose.
If you want one short email a week with the actual state of these EAs (frozen weeks included, no marketing varnish): Join the newsletter. About 230 readers, sent only when there's something honest to say.
