ForumSections

[Scalper EA] What should I consider to bring this EA Live? - page 2

New comment
[Deleted]
This raptor has something against me ofcourse i know and can see in the chart the lot size is increasing.

I have made a tick scalper which seems to be very good in backtest. However, modeling quality is only 25%. Where can I download the tick data to make 99% modeling quality?

Recently I read an article on back testing using data of one minute or less. The article claimed that you could not get better than 25% when back testing on one minute or less. And it went into great detail explaining why. Sorry I can't cite the reference right now but I'll keep looking for the article.
MisterDog:
Recently I read an article on back testing using data of one minute or less. The article claimed that you could not get better than 25% when back testing on one minute or less. And it went into great detail explaining why. Sorry I can't cite the reference right now but I'll keep looking for the article.

It would be probably right. I guess the only good way to know if it works is forward test: excesive optimism is the first cause of busted accounts LOL.

I'll get some VPS. Do you guys know of a nice VPS located in the UK? If you do, send me a private message. Thanks!

flaab:

It would be probably right. I guess the only good way to know if it works is forward test: excesive optimism is the first cause of busted accounts LOL.

I'll get some VPS. Do you guys know of a nice VPS located in the UK? If you do, send me a private message. Thanks!

A forward test is no different to a backtest except it is a lot slower . . . assuming you have the right tools to do a relevant backtest.
RaptorUK:
A forward test is no different to a backtest except it is a lot slower . . . assuming you have the right tools to do a relevant backtest.

Really? I am downloading Dukascopy tick data to test it more accurately. So forward test live data is different from a real acount??

How could I do a relevant back test?

Thank you!

flaab: So forward test live data is different from a real acount??
Yes. You'll be using YOUR broker's data. There is no slippage in the tester (your open and TP/SL will be different.) Variable spread. Time delay opening and closing order. Possible loss of connection.
WHRoeder:
Yes. You'll be using YOUR broker's data. There is no slippage in the tester (your open and TP/SL will be different.) Variable spread. Time delay opening and closing order. Possible loss of connection.
I understand better now thanks. Do demo accounts mirror the behavior of a real account?
flaab:
I understand better now thanks. Do demo accounts mirror the behavior of a real account?
Not all . . . some Demo accounts have better Spreads than the live account you may be able to open with the same Broker.
WHRoeder:
Yes. You'll be using YOUR broker's data. There is no slippage in the tester (your open and TP/SL will be different.) Variable spread. Time delay opening and closing order. Possible loss of connection.
Variable spread can be done using tick data . .

I have backtested with tick data downloaded and results are pretty much the same (See chart attached where 90% quality is reached).

I have also tested with wider TP/SL/Trailing stop values. Like...60TP, 80SL and 15-20 pips trailing stop, and results are also the same.

I don't know how to further verify the performance of this EA. A forward test in demo account would be next, but I don't know if it is of much use.

To what extend is a test on a demo account representative?

EURUSD 2005-2012.

1234
New comment