如何在信任之前验证黄金 EA 的可验证交易记录

Before a Fed decision

The next Fed rate decision is due on Wednesday, 28 October 2026, at 14:00 Washington time (19:00 Zurich time). In the weeks with a decision, spreads can widen, fills can slip and gold can move fast, so a weak EA tends to show its flaws then. For the full checklist to run before a news day, read How to run a gold EA on news days. No strategy has passed yet; every result is public at baroqtech.com/record.

验证清单

• Third-party tracked? You need a live tracker link that reads the account directly. Screenshots and "private results" don't count.
• 真钱？账户类型应显示为真实（live）。模拟记录是另一回事。
• History and privileges verified? Both need to be verified, not just one.
• Enough trades, over enough time? A handful of trades over a few weeks is too early to say anything.
• Costs inside the numbers? Commission and swap should be charged in every figure.
• 净值回撤？未平仓的亏损也算，不只是已平仓的交易。
• 资金转入转出？检查入金和出金，以及自定义的起始日期是否隐藏了最初几个月。在回撤期间入金，会掩盖回撤究竟有多深。
• Is this the only account running it? Run one EA on many accounts and one of them will look good by chance.
• Close to the backtest? Live results far below what the test promised are a warning.

过去的良好记录并不预示未来的结果。通过这份清单并不能保护您免于亏损。

“已验证”证明了什么，没证明什么

“已验证”证明的是，跟踪平台读取了一个真实的交易账户。它不证明这个 EA 有效，也不证明它会一直有效。

Verified track record: a trade history that a tracking service has read straight from the broker account, usually through an investor (view-only) password. The trades shown are the trades that happened.

交易权限：另一项检查：出示记录的人是否真正控制该账户。有历史而没有权限，只能说明这些交易发生过，不能说明它们属于卖家。

各跟踪平台的标签检查什么

• Myfxbook: Track record means "the account's trading history shown on Myfxbook matches the account's trading history provided by the broker". Trading privileges means "the individual who added the trading account possesses the master password".
• MQL5 Signals: It counts weeks "since the first trade performed on the trading account". It shows the trading mode as "demo or real", and notes that "a provider can afford much greater risks on a demo account than they would on a real one."
• FX Blue: FX Blue sorts published accounts by where the data came from (About FX Blue verification, read 26 October 2026):
  • Verified means "FX Blue has collected the account data directly from the broker, using our account sync, or comes from a similar source where the user cannot manipulate the data." FX Blue adds that it cannot rule out a broker manipulating the data.
  • Semi-verified accounts "are uploaded using our publisher EAs for MT4 or MT5". FX Blue says the results "are almost certainly genuine. However, it is theoretically possible for users to manipulate the data which is sent to FX Blue."
  • Everything else is unverified, including backtesting reports. For those, FX Blue says "it is impossible to verify whether the results have been manipulated, or whether they are from a demo account or a real one."

FX Blue doesn't say that it checks who controls the account. Treat a semi-verified label as weaker evidence than a verified one.

没有任何跟踪平台能告诉您的事

• 是哪个 EA 下的这些单。
• 用了哪些设置。您购买的版本可能不是用记录背后的那些设置运行的。
• 它是否会一直有效。记录描述的是过去。

模拟与真实：绝不要混在一起

Demo vs live account: a demo account uses practice money at the broker, and a live account uses yours. Fills, spreads, slippage and pressure are all different.

Never add the two into one number. The desk is BaroqTech's MT5 dashboard, and it only reads your account. It shows demo and live on two separate sides, in two colors, and no total ever adds practice money to real money.

多少笔交易之后，一份记录才有意义？

A short record can look excellent by luck alone. The fewer the trades, the wider the range that chance alone can produce. If a record is only a few weeks old, the desk's verdict will read "too early", whatever the profit says. That's the minimum for reading a record at all, not a reason to trust one. BaroqTech's judge is the locked set of tests every trading idea has to pass, and it asks for many trades in the period it judges. One idea on the public record, daily momentum, was turned down as "too few trades to judge".

阅读成本：佣金、隔夜利息、点差

Gold strategies often trade short moves, so costs take a big share of each trade, and swap adds up overnight. A record that leaves costs out shows a result the account never earned.

BaroqTech's tests charge spread, commission and slippage on every test trade, plus overnight financing on long positions held past the rollover. They then run again with costs doubled. An edge that only survives cheap trading is not an edge.

Worked example, in words

Picture a system that wins most of its trades with small gains and loses a few with large losses. Once spread, commission and swap come off every trade, the small gains can shrink until the few large losses outweigh them. A high share of winning trades says nothing until the costs are counted.

用正确的方式数交易

Partial close: closing part of a position and leaving the rest open, for example taking profit in steps.

One position is one trade, however many times it was partly closed. Counting each partial close as a trade inflates the trade count and the share of winners, while the money stays the same. The desk will merge partial closes into one trade.

危险信号

净值回撤：账户净值从峰值起的最大跌幅，包含未平仓交易。

BaroqTech 如何用同样的规则要求自己

BaroqTech makes quantitative analytics software for gold (XAUUSD) traders on MetaTrader 5, and it holds its own trading ideas to this checklist. Every idea goes through six years of gold prices, the judge (locked tests nobody adjusts afterwards), a demo account, and then a live account with small size, a stop on every trade and a daily loss limit. The full rules are on the method page.

No strategy has passed yet; every result is public at baroqtech.com/record. Each idea is judged on costs, on prices it hasn't seen and on a luck test, and an idea that fails any of them is turned down. Every run is in the log.

The Gold EA Checking Kit puts every check on one page. It is included with the Lifetime Pass.

Related: Why gold EAs pass the backtest and fail live · How to spot a fake gold EA

常见问题

BaroqTech 提供的是软件。它不是投资建议，也不管理资金。交易黄金差价合约等杠杆产品存在较高的亏损风险，您可能损失全部本金。测试或模拟账户的结果并不能预测真实账户的结果。请阅读完整的风险提示和免责声明。