Problems with backtesting - suggestion required

[Financialabs]

Hi,

when doing backtesting on my EA I noticed that "time value" of open/close orders on "Result" tab is not aligned with the graph, especially with the bar where the blue/red arrow appears.

Does it mean that the backtesting is not accurate?

How can I check accuracy of backtester?

Thanks

Alex

[Keith]

Financialabs wrote >>

Hi,

when doing backtesting on my EA I noticed that "time value" of open/close orders on "Result" tab is not aligned with the graph, especially with the bar where the blue/red arrow appears.

Does it mean that the backtesting is not accurate?

How can I check accuracy of backtester?

Thanks

Alex

Generally the tester does what the ea tells it to; if the results are not what you expect then more than likely the ea is not

working as you expect, you may need to check throught the logic very carefully.

Also, it would not be unusual to see the time stamp lag a little behind the arrows,

likely the arrows are painted at close of bar, and open/close executed slightly later on the next bar. How large is the error?

HTH

Keith

[onyx]

kminler:

Generally the tester does what the ea tells it to; if the results are not what you expect then more than likely the ea is not

working as you expect, you may need to check throught the logic very carefully.

Also, it would not be unusual to see the time stamp lag a little behind the arrows,

likely the arrows are painted at close of bar, and open/close executed slightly later on the next bar. How large is the error?

HTH

Keith

I realise this is an older thread yet it deserves some technical clarification, all my EAs have their order times one bar ahead of where they visibly should be on the graph.

I long ago stopped using Bar 0 or such small variations to forecast for my EAs, yet even so I cannot fully understand why the order time does not line up on the x-axis.

I do understand there is a great deal of uncertainty between the most recent time index and the closing price drawn, which would be the most likely cause, yet still would it not be better to have these order indicators corrected after the fact, so they lined up logically with their related time on the X-Axis? At least an option of having this happen would be appreciated in future versions.