Hi,

I have developed a strategy in VJFX (attached). I test it using the historical tester in Visual JForex. When I then run the strategy in the JForex platform (with the view to optimising it) i find that it doesn't take any trades on exactly the same parameters. Why? I assume that the two historical testers should give the same results??

I suspect it could be be related to me closing orders and positions that are not actually open (I do this for two reasons: I close pending orders on each new candle, this is because the I can't get orders to close on time viewtopic.php?f=139&t=50474 and I close open positions after three candles if they have not hit their SL or TP).

Although, it is perhaps bad practice/laziness that close orders/positions that are no longer pending/open, my thinking is why does this cause me problems the whole strategy to fail in JForex, especially when it still trades in Visual JForex?

Is this is known inconsistency between the two historical testers? And, do I definitely need to make sure there are no errors when using JForex?

FYI, in JForex I can see the error message is: java.lang.NullPointerException and I view the java code ands see what if block it relates to, it is definitely down to my closing of pending orders that may in fact have been filled.

thanks

I've since found that by using "Flow" on the position view, rather than "Cycle End" this problem goes away.

I'm curious, however, why do they two historical testers give different results. ie, in once instance the strategy would run and the other it wouldn't?

regards

In the Visual JForex tester, all positions are closed at the end of the test (and than commissions deducted), but in the JForex historical tester, the positions are let open.