The dilemma of tester and live demo account data.
I tried comparing the data from a frequently traded EA tester with data from a live demo account. The incredible difference is very interesting. This could imply that all our tests outside of the live data are useless.
Here are the main reasons why your test data and live real data are completely different:
1. Transaction Latency (Ping Time)
Test Device: Even if you use the "Every tick based on real ticks" mode, the MT5 test device defaults to transmitting and placing orders in the market at 0 milliseconds (perfect speed). 
Live Account: There is a data transmission time (ping) from the computer or VPS where your EA is running to the broker's server. In XAUUSD, prices change in milliseconds. When your EA generates a signal in the live market and says "Buy," by the time the order reaches the server (e.g., 50-100 ms), the price may have already changed, and you may miss the trade or enter at a worse price.
2. Price Slippage and Liquidity (Market Depth)
Test Device: A test device typically assumes there is infinite liquidity (buying and selling equivalent) at the current price level for your order volume (0.01 lot). It closes the trade perfectly the moment the price touches the level (TP/SL).
Live Account: In the live market, especially with pairs like gold (XAUUSD), there may not be enough sellers/buyers at your requested price during news hours or sudden movements. You will be matched at the next price in the order book. In grid and scalping systems, a slippage of 1-2 pips can lead to losses in live trading for a system that looked perfect in the test.
3. Spread Widening 
In a live account, spreads (the difference between the buy and sell price) are dynamic. During nighttime hours, market openings, or important data releases, spreads can reach incredible levels. Even though MT5 uses "Real Ticks" when downloading historical data, it may not perfectly record all of those millisecond-long abnormal spread widenings in the historical data. This causes the test environment to present a "smoother" and more profitable market than reality.
4. Broker Trading Rules and Rejections 
Your EA system sends very fast trade modifications (modify position) and orders in milliseconds.
The test device accepts every order and modification it sends without exception.
However, a live broker server may queue requests sent very quickly in quick succession (High-Frequency Trading behavior) to relieve server load, freeze them at levels, or requote them, delaying the order for seconds.
That's why no tester matches real data. Don't discard any of your EA projects without testing them in a live environment.
http://www.ojas-gujarat-gov-in.com/attachm...=1770884525998
http://www.ojas-gujarat-gov-in.com/attachm...=1770884549564
http://www.ojas-gujarat-gov-in.com/attachm...=1770884566219
I tried comparing the data from a frequently traded EA tester with data from a live demo account. The incredible difference is very interesting. This could imply that all our tests outside of the live data are useless.
Here are the main reasons why your test data and live real data are completely different:
1. Transaction Latency (Ping Time)
Test Device: Even if you use the "Every tick based on real ticks" mode, the MT5 test device defaults to transmitting and placing orders in the market at 0 milliseconds (perfect speed).
Live Account: There is a data transmission time (ping) from the computer or VPS where your EA is running to the broker's server. In XAUUSD, prices change in milliseconds. When your EA generates a signal in the live market and says "Buy," by the time the order reaches the server (e.g., 50-100 ms), the price may have already changed, and you may miss the trade or enter at a worse price.
2. Price Slippage and Liquidity (Market Depth)
Test Device: A test device typically assumes there is infinite liquidity (buying and selling equivalent) at the current price level for your order volume (0.01 lot). It closes the trade perfectly the moment the price touches the level (TP/SL).
Live Account: In the live market, especially with pairs like gold (XAUUSD), there may not be enough sellers/buyers at your requested price during news hours or sudden movements. You will be matched at the next price in the order book. In grid and scalping systems, a slippage of 1-2 pips can lead to losses in live trading for a system that looked perfect in the test.
3. Spread Widening
In a live account, spreads (the difference between the buy and sell price) are dynamic. During nighttime hours, market openings, or important data releases, spreads can reach incredible levels. Even though MT5 uses "Real Ticks" when downloading historical data, it may not perfectly record all of those millisecond-long abnormal spread widenings in the historical data. This causes the test environment to present a "smoother" and more profitable market than reality.
4. Broker Trading Rules and Rejections
Your EA system sends very fast trade modifications (modify position) and orders in milliseconds.
The test device accepts every order and modification it sends without exception.
However, a live broker server may queue requests sent very quickly in quick succession (High-Frequency Trading behavior) to relieve server load, freeze them at levels, or requote them, delaying the order for seconds.
That's why no tester matches real data. Don't discard any of your EA projects without testing them in a live environment.
http://www.ojas-gujarat-gov-in.com/attachm...=1770884525998
http://www.ojas-gujarat-gov-in.com/attachm...=1770884549564
http://www.ojas-gujarat-gov-in.com/attachm...=1770884566219
OF COURSE, IT'S LUCK.