Back test unreliable

PanagiotisCharalampous
16 Jun 2020, 15:14

Hi Adam,

Can you provide more information about this issue like cBot source code and steps to reproduce it?

Best Regards,

Panagiotis 

Join us on Telegram

acrigney
17 Jun 2020, 03:07

RE:

Mate I think you would have to post your code to make accusations like this. Have you got some randomness in your code logic?

shapebubble said:

Hi ya

 

I've been doing a lot of back tests on Ctrader and have noticed that it seems flawed. The system returns different result each time, doesn't trade at all the opportunities and reliably returns losses. I came to this realisation because I noticed that it was returning reliable losses and so had an algo made the did the opposite. However it still returned losses and did not do the same trades. If I have 2 ctraders open and run the same test with the same algo over the same dates I get different results. If I close the program, open again, enter the same data and rescan I get different results.

 

I can't see how the system is usable if it is so unreliable. I have tested both Clickalgo algo's and a freelancers and Ctrader just seems so unreliable.

 

Does anyone else have this problem?

 

Thanks

 

Adam

shapebubble
17 Jun 2020, 12:07

RE:

Hi ya

I've done a little reading about it and it might be because I'm using ctrader and fxpro ctrader.

The results of the optimisation can be wildly different in their effect on following data though, is that normal? For example I just did a scan and picked the best optimisation for max equity and got around £300 profit on the following date period. I scanned the same time period again and picked the best  equity again and got £100 loss on the same following time period. Is that normal because it seems to be for me.

Thanks

shapebubble said:

Hi ya

 

I've been doing a lot of back tests on Ctrader and have noticed that it seems flawed. The system returns different result each time, doesn't trade at all the opportunities and reliably returns losses. I came to this realisation because I noticed that it was returning reliable losses and so had an algo made the did the opposite. However it still returned losses and did not do the same trades. If I have 2 ctraders open and run the same test with the same algo over the same dates I get different results. If I close the program, open again, enter the same data and rescan I get different results.

 

I can't see how the system is usable if it is so unreliable. I have tested both Clickalgo algo's and a freelancers and Ctrader just seems so unreliable.

 

Does anyone else have this problem?

 

Thanks

 

Adam