Before releasing the tool, what do you check for two weeks on a demo account

I am Yohaku, creating MT4/MT5 indicators and handy tools. What I make is decided from the complaints that appear in Gogazine reviews and questions.

Before releasing any tool, I test it for two weeks on a demo account. I will write in detail what I look at this time. I hope this serves as a reference for buyers when considering how thoroughly a tool has been verified.

Conclusion first

• Three main things I verify before release: there are no compile warnings on both MT4 and MT5, orders and displays operate exactly as specified, and it runs without defects for a long time.
• Even if warnings disappear, defects can appear after long-term operation. Therefore, in addition to automated tests, I keep it running on a demo account.
• For tools that place orders, two weeks; for tools that do not place orders, continuous automated testing for three consecutive days after full pass (was decided on October 8, 2026).

What I verify before release

What I am currently doing with the tool is as follows.

• Create both MT4 and MT5 versions and ensure zero compile warnings for both
• Automatically test eight order types (market order, limit order, stop order × buy/sell, adding two methods that move the entry line back to market after adjustments) and partial close
• Automatically test switching timeframes, currency pairs, multiple charts, and removal of tools
• Restart MT4/MT5 to see if panels and settings revert
• After removing the tool, count whether drawn lines and buttons remain
• Continue running on a demo account to check for any defects

Since a human might overlook it, I write a program that automatically sends button clicks and line drags to verify results based on the provider's information.

Results of trying it out

On October 7, 2026, I conducted order tests on demo accounts for MT4/MT5 with GOLD (XAUUSD) and USD/JPY. The eight order types and nine partial-close items were tested in four configurations for MT4/MT5 × two instruments, totaling 36 runs. I confirmed that order types, limit/stop prices align with the lines, the directions of stop loss and take profit, and that lot sizes do not exceed allowable loss, and all 36 passed.

On the same day, I also ran tests for startup, timeframe switching, currency pair switching, multiple charts, and removal. All 19 items passed in four configurations. After removal, there were zero components drawn by the tool, and the saved values returned to zero as well.

For continuous operation, I confirm that after restarting the terminal, the panel opens in its original state.

Why even bother with long-term operation

Because some defects do not appear in short checks even if compilation passes. In practice, automated tests and continuous operation have revealed the following:

• Panel text and buttons become hidden behind a colored background, visible only as colored rectangles
• When switching currency pairs, the price line of the previous pair remains
• While moving the stop-loss line with the mouse, the line flickers

All of these pass compilation, and they are not obvious at a glance. They only appear when you actually click or run for hours. Therefore, I take the time before releasing.