You know the drill. Someone drops a screenshot of a 3,000% backtest, slaps “AI-powered” on the name, and expects you to hand over $500.
You’ve seen it a hundred times. So have I. And I’m tired of it too.
So here’s what I’m going to do instead: a full Alpha Pulse AI performance breakdown. Every single number — the wins, the losses, the months that worked, and the one that didn’t. Because after testing dozens of “AI bots” on MQL5, I can tell you that a perfect equity curve is practically a confession that someone is selling you smoke.
If these numbers impress you, great. If they don’t, that’s fine too. At least they’re real.
Alpha Pulse AI Performance — Myfxbook Verified, April 7 2026
No cherry-picking. No “since inception” tricks. No convenient start date. Just the account as it stands right now:
- Total Gain: +15.25%
- Monthly Average: +4.23%
- Max Drawdown: 8.60%
- Profit Factor: 1.29
- Win Rate: 55%
- Total Trades: 105
- Average Trade Length: 2h 52m
- Starting Balance: $7,046.58
- Current Balance: $8,123.22
Don’t trust me. Verify: Alpha Pulse AI Baseline on Myfxbook. Every trade. Every losing streak. Nothing hidden.
April has been particularly strong: +8.00% gain, 80% win rate on 5 trades. But I’m not going to pretend one good month defines the EA. I got too hyped about early results once before — a couple of fantastic weeks followed by a couple of rough ones that brought me back to reality. That’s how this works. The full picture is what matters.
55% Win Rate — And Why That’s Actually Good
If you see “55% win rate” and think that’s mediocre, I get it. You’ve been conditioned by vendors advertising 90%+ win rates. Those numbers come from two places: cherry-picked backtests or martingale strategies that will eventually blow your account to pieces.
Here’s the math that actually matters:
- Average Win: $82.74 (2,212 pips)
- Average Loss: -$79.19 (2,228 pips)
- Expectancy per Trade: +$10.25
The wins are slightly larger than the losses, and they happen more often. PF 1.29 means for every dollar lost, $1.29 comes back. Across 105 trades, that’s $1,076 in net profit on a $7K account.
Is it going to make you rich by Friday? No. Is it going to blow up at 3 AM while you sleep? Also no. And right now, in a market where most EAs are designed to separate you from your money, real Alpha Pulse AI performance like this — consistent and verified — is the edge nobody wants to sell you. Because it’s not sexy enough for a landing page.
Every trade recorded. Every loss visible.
No screenshots. No curves without context. Verify the live Myfxbook signal or check the full product details.
8.60% Max Drawdown — The Number That Separates Survivors
Most traders obsess over win rate. The ones who survive obsess over drawdown.
8.60% max drawdown on a live account trading Gold — one of the most volatile instruments in the market — means the risk management is doing its job. Gold moved over $400 per ounce in March 2026 alone. The account stayed within 8.6% of its peak.
Why does this matter more than any return number? Because drawdown is where discipline dies. A 30% drawdown on a gorgeous equity curve means absolutely nothing if you can’t psychologically survive it. You’ll turn the EA off at -25%, miss the recovery, and declare “it didn’t work.” I’ve seen it happen hundreds of times.
8.6% is livable. If that drawdown already makes you uncomfortable, then you’re oversized — and that’s a risk management problem, not an EA problem.
Best and Worst — No Hiding
- Best trade: +$246.68 (Feb 9)
- Worst trade: -$165.00 (Feb 16)
- Best in pips: +8,178 (Mar 20)
- Worst in pips: -8,535 (Mar 23)
Yes, the worst trade in pips is larger than the best. That happens in real trading. If someone shows you a system where every loss is smaller than every win, they’re either curve-fitting their backtest or outright lying. The system works because the math works across 105 trades, not because every individual trade is perfect.
The Bad Period — Because There Was One
I’ll be honest: early on, I had too much hype. A couple of really strong weeks made me think it was already running at full potential. Then came a couple of rough weeks that brought everything back to earth.
That’s not a failure. That’s what real Alpha Pulse AI performance looks like across time — not a perfect line going up, but a system that survives bad conditions. If someone sells you an EA where every month is positive, they’re either lying or running something that will eventually implode. Good EAs survive bad markets. Great EAs recognize them and reduce exposure.
Alpha Pulse AI does something that most EAs physically can’t: before every trade, it sends real market data to an AI model — price action, volatility, session context — and gets back an actual analysis. Not a moving average crossover with “AI” in the name. An actual decision from GPT-5.4, Gemini 3.1 Pro, or Claude Opus 4.6 about whether conditions favor a trade right now.
When conditions are garbage, the EA often stays out. That’s why you see 2-3 trades per week, not 20. There are many models, many pairs, many timeframes… that’s exactly why you have to take it slow.
Longs vs Shorts — The AI Learned Something
- Longs Won: 50% (32/63 trades)
- Shorts Won: 61% (26/42 trades)
The EA has been notably better at shorting Gold than buying it. This makes sense — XAUUSD has had sharp pullbacks during the broader rally, and the AI models seem to identify reversal conditions more accurately than continuation setups.
This isn’t something I programmed. It’s something the AI learned from analyzing conditions in real time. A static EA would trade both directions with the same logic. This one adapts — and the data shows it.
One EA is a psychology test, not a trading plan.
Alpha Pulse handles Gold. But running a single strategy alone means your entire equity depends on one instrument’s behavior. Download the free USDJPY module and start diversifying at zero cost.
Who This Is For (And Who Should Stay Away)
This IS for you if:
- You understand that 55% win rate with positive expectancy beats 90% win rate with hidden martingale risk
- You want verified results — not screenshots, not backtests, not “trust me bro”
- You can let the EA run for months without touching it after one bad week
- You want an EA that evolves as AI models improve — without manual updates or new purchases
- You accept that some weeks will be flat or negative. That’s not a bug. That’s reality.
This is NOT for you if:
- You need guaranteed monthly returns. Nobody can guarantee that. If they say they can, they’re running a scam
- You’ll panic and turn the EA off during the first drawdown. At that point, you’re not trading — you’re gambling with extra steps
- You expect 90%+ win rates. That’s martingale territory, and you know how that ends
- You want easy money. This isn’t easy money. There’s still risk. If someone sold you the idea that bots print money, they scammed you
What’s Actually Under the Hood
Real AI trading means passing real market information to a model and having that model actually participate in the decision. That’s it. That’s the definition. Everything else is marketing makeup.
Before every trade, Alpha Pulse AI sends current market data to one of these models:
- GPT-5.4 — OpenAI’s latest. Strong analytical depth
- Gemini 3.1 Pro — Google’s model. Fast processing, free API tier (up to $300 in credits)
- Claude Opus 4.6 — Anthropic’s deepest thinker. Best for multi-layered analysis
- Grok 4.20 — xAI’s model. Pattern recognition focused
You choose which model to run. You switch whenever you want. When these providers release better models, your EA uses them automatically. That’s the auto-evolution advantage — since Gemini 3.1 Pro and GPT-5.4, the analysis quality has improved dramatically compared to earlier versions.
The presets have optimized the frequency of API calls and how the analysis is structured. This matters because API costs are real — and I’d rather be honest about costs than hide them like everyone else.
Cost Transparency
- EA Price: $397 one-time. No subscription. No recurring fees. No “premium tier” upsell
- AI API Cost: $0/month with Gemini free tier, or $2-5/month with premium models at normal trading frequency
- VPS: $15-30/month recommended for consistent execution (ForexVPS works well)
Total ongoing cost: $15-35/month. On a $7K account making 4% monthly, that’s covered in the first trade. On a smaller account, use Gemini free tier and skip VPS costs until you scale. There’s a real cost problem with AI trading — most vendors hide it. I’d rather you know upfront.
FAQ
Can I verify these results independently?
Yes. Every number in this post comes from the live Myfxbook signal. Every trade, every drawdown, every losing streak. Nothing hidden. Proof without context is worthless — that’s why I’m showing you context.
Why only 105 trades? Isn’t that a small sample?
It’s enough to establish a pattern, not enough to prove lifetime consistency. I think we’re at the point where, if it’s not yet consistently profitable long-term, it’s very close. At 200+ trades, the statistics become much more robust. I’ll keep publishing updates.
What happens during high volatility events?
The AI models analyze volatility as part of their decision. During March 2026 market turbulence, the EA reduced trade frequency and kept drawdown controlled. That’s the difference between a static EA and one that actually reads conditions.
What broker do you recommend?
IC Markets for raw spreads on Gold, or Axi Select if you want to scale capital without challenge fees. As an active Axi affiliate, I can escalate any account issue directly to my manager — resolution in days, not weeks.
Can I run this on a funded account?
Yes. The EA’s conservative risk settings (2% default) and controlled drawdown (8.6% max observed) make it suitable for most funded account programs. Drop to 0.5-1% risk for stricter prop firm rules.
Resources
105 trades in. The direction is positive and the risk is controlled. Follow along and see where we are at 200.