What makes a good and truly profitable EA?
This page is not written to sell, but to provide you with criteria to evaluate every EA on the market, including our own. At the end of the page, there's a table comparing MindCore against each of the 10 criteria, honestly stating which ones we haven't met yet.
📡 Watch the system's gold signals live, free
MindCore logs every signal to a Telegram channel with real outcomes, wins and losses, unedited · no sign-up · not investment advice
What a truly profitable EA must have
These criteria are derived from what allows a trading system to survive across various market conditions, not what makes advertising look good.
Tested across many market conditions, not just rising gold
Anyone can profit by testing only in years when gold rises. A good system must be run through bear markets, crises and sideways markets on real minute-by-minute data, and the seller should be able to tell you how it performs in each of them.
No Martingale, no Grid, no averaging down losses
These three techniques lead to a high win rate and a beautiful graph for many months, then wipe out the account in a single day when the market moves in one direction without turning back.
Every trade has a stop-loss from opening, and fixed lot size.
SL must be with the broker from the second the trade opens, not "closed by code" which disappears if the VPS goes down. And lot size should not grow based on the system's emotions.
Profit comes from the size of winning trades, not the win rate.
True trend-following systems win 30 to 45% of trades, but winning trades are many times larger than losing trades. A 90% win rate in the gold market is almost always a sign of point 02.
Rules can be explained in plain language.
A good system can be explained in one minute: when to enter, when to exit, and why. "Secret AI" that cannot be explained usually means parameters are tuned to fit the past.
Frequency is reasonable for the timeframe.
The market provides profit based on the distance price moves, not the number of entries. A system that opens 20 trades a day on gold is paying more in spread to the broker than the market provides.
Real results confirmed by a third party, not screenshots.
Screenshots can be faked in five minutes. Results linked to a real account via Myfxbook or FX Blue with an investor password cannot be faked, and it must be a real account, not a demo.
Survives real costs: spread, slippage, and fees.
A system that is marginally profitable in backtesting will die immediately when it encounters real spread. The seller should be able to tell you how much profit every $1 of spread consumes.
State minimum capital and tolerable drawdown
All systems have losing streaks. A good seller will disclose these numbers before you buy, such as how many consecutive losses are possible, and how much capital is needed per 0.01 lot.
No guaranteed returns and no fund management
Good software is sold as software, not a promise. Anyone guaranteeing monthly profits is doing two things simultaneously: lying and breaking the law.
7 Warning Signs
- Win rate over 85% on gold is almost impossible without Martingale or no SL.
- Equity curve is a straight upward line with no drawdowns, meaning losses are hidden as open floating positions.
- Shows results only from 2024 to 2026, a period when gold rose 70%. Simply holding gold would have been profitable.
- No drawdown figures, or figures that are unusually low compared to profits.
- 3-day trial is too short to see the true nature of the system; it's designed for you to decide before seeing the reality.
- Results from demo accounts or screenshots only, with no external verification links.
- Unusually cheap price or lifetime sale. Sellers who believe in their system want to stay with customers long-term.
Your EA has to stay on to see every bar close · why we point you here - affiliate link, same price for you
Compare point by point with real numbers
All figures are from Strategy Tester, every tick mode, and 23 years of data analysis, updated September 6, 2026.
| Criteria | How typical EAs in the market usually are | MindCore | Status |
|---|---|---|---|
| 01 Multiple Market Conditions | Shows last 1 to 2 years | The 10-year test cut into 6 market conditions, profitable in every one, with the aggregate figures published in full: 43% of trades closed in profit, maximum drawdown 25.69%. | Passed |
| 02 No Martingale, Grid, Averaging Down | Most common in high win-rate groups | Position size from capital; never adds to a losing trade; adds only when the existing trade is in profit. | Passed |
| 03 SL on every trade from opening | Some "close automatically in code" or have none | SL sent to broker with open order, distance sized to market volatility. | Passed |
| 04 Profit from winning trade size | Boasts 90% win rate | 42.7% win rate. Average winning trade $2,084, average losing trade $778 (2.68x ratio). | Passed |
| 05 Rules can be explained | "Secret AI" cannot be explained | The principle is clear: follow the bigger trend, enter when it is confirmed, exit when it changes, stop loss on every trade (parameters are a trade secret). | Passed |
| 06 Reasonable frequency | Dozens of trades per day | Around 7 trades per week on average. We've measured that increasing frequency doesn't increase profit. | Passed |
| 07 Real results verified by third party | Screenshot | Our real account is running, but history is short and not yet linked to Myfxbook. We will link and show it after 3 months. | Not yet complete |
| 08 Survives real costs | Not mentioned | Measured: every $1 of average spread consumes 11% of profit. EA displays broker's spread cost on screen. | Passed |
| 09 States minimum capital and drawdown | Not stated, or stated unrealistically low | Maximum consecutive losses 52 trades. Maximum drawdown 4.6% with $10,000 capital. Recommended capital 15,000 to 20,000 THB per 0.01 lot. | Passed |
| 10 No guarantees, no money management | "Guaranteed 20% profit per month" | Sold as software. Risk warnings on every page. Forecasts are ranges with probabilities, not single figures. | Passed |
We will update this table monthly. If any item changes from met to not met, we will write it here before anyone asks.
If you only remember one sentence
A good EA is not one that never loses, but one that tells you in advance how it will lose, how much it will lose, and why it's still worth leaving on. If the seller cannot answer these three questions, no matter how good the chart looks, consider it not good enough.