After the latest terminal update SELL STOP and BUY STOP orders in the tester are executed at buy and sell prices, but not at order placing price - this is only in opening prices in testing mode.
Here are screenshots of two tests of the same Expert Advisor with the same parameters, the only difference is inall ticks and all ticks based on real ticktesting mode. Your terminal is 1340, your account. Please advise which of these tests can be used as a reference and shown to clients as backtest results.
Stanislav Dray: Here are screenshots of two tests of the same Expert Advisor with the same parameters, the only difference is inall ticks and all ticks based on real ticktesting mode. Your terminal is 1340, the account is yours. Please advise which of these tests you can use as a reference and show it to your clients as a backtest.
The second screenshot (1t) doesn't define the simulation quality at all. It doesn't look like this test was run on ticks at all...
Alexey Kozitsyn: The second screenshot (1t) does not define the modelling quality at all. It doesn't look like this test was run on ticks at all...
I seem to have clearly described the test conditions and the difference between the tests specifically highlighted. I want to use 2-core 32bit machine, i have enough free space for history on disk, no errors during testing, i have no problems with history resuming at startup, testing period 2014-2016. Adviser does not use lot increment (each order with stoploss).
After the latest terminal update SELL STOP and BUY STOP orders in the tester are executed at buy and sell prices, but not at order placing price - this is only in opening prices in testing mode.
Here are screenshots of two tests of the same Expert Advisor with the same parameters, the only difference is in all ticks and all ticks based on real tick testing mode. Your terminal is 1340, the account is yours. Please advise which of these tests you can use as a reference and show it to your clients as a backtest.
The second screenshot (1t) does not define the modelling quality at all. It doesn't look like this test was run on ticks at all...
Compilation error
The second screenshot (1t) does not define the modelling quality at all. It doesn't look like this test was run on ticks at all...
When testing based on real ticks the quality of history is not analyzed. For there is no need.
Look how many ticks there are in comparison with "all ticks" mode (where the quality is 98%)
When testing based on real ticks, the quality of the story is not analysed. For there is no need.
Look how many ticks there are compared to "all ticks" mode (where the quality is 98 percent)
Not all characters are displayed incopyright and not all times
Result:
Not all characters are displayed incopyright and not all times
Result: