Backtesting/Optimization - page 83

JamalJohnson
36
JamalJohnson  

Why MT4 cuts my backtest

Why does mt4 cut off backtests when it decides that I have too much risk? It is the low leverage? Have they changed something in the later builds?

Im doing a martingale similar to the Blessing EA and it just decides to stop my test.

Any suggestions to work around this problem?

Mladen Rakic
177216
Mladen Rakic  

...

In the experts properties change the initial deposit to bigger amount (at least in back-tests and demo accounts we all can be millionaires:))

Also, try setting your EAs properties to minimize risk (smaller initial lot size for martingale , smaller total risk %, smaller total martingale steps - just guessing what could a martingale EA have to lessen the risk)

JamalJohnson:
Why does mt4 cut off backtests when it decides that I have too much risk? It is the low leverage? Have they changed something in the later builds?

Im doing a martingale similar to the Blessing EA and it just decides to stop my test.

Any suggestions to work around this problem?
JamalJohnson
36
JamalJohnson  
mladen:
In the experts properties change the initial deposit to bigger amount (at least in back-tests and demo accounts we all can be millionaires:)) Also, try setting your EAs properties to minimize risk (smaller initial lot size for martingale , smaller total risk %, smaller total martingale steps - just guessing what could a martingale EA have to lessen the risk)

Thank you for speedy reply.

Question

Will an older build of MT4 be better? I remember that my back-tests would not stop until the money was gone. As it is now, mt4 stops my tests with 85% of the initial still in the account. It ridiculous that it decides that I don't wish to risk more. I cant get a real feel for the risk because the software decides that I don't need to. It's ridiculous.

Ill try the larger account size. thanks

sunshineh
1312
sunshineh  

Doing the optimal backtest...

Hi,

I have the possibility to get backtest data.

What data's do I need for the best "true" result?

Tickdata or Minute-Data or ALL Timeframes??

Bid AND Ask

and of course Open, High, Low, Close

Or is there a better system to do backtests f.e. ninja trader, ...

biddick
339
biddick  

To get the result that is closest to real time trading ,You need a platform with walk forward optimization and testing :Walk forward optimization - Wikipedia, the free encyclopedia For example Wealth lab is powerful platform with differential genetic programming functions.Don't forget to include spread in all your tests.Good luck.

sunshineh
1312
sunshineh  

Thank you for your interesting comment! So as I see, I can do a walk forward optimation with my metatrader plattform and good datas by hand.

The only miss, as I can see, is that the metatader platform uses always the actual spread for all calculations. Or is there a possibility to use the real tickdatas with bid and ask in mt5?

biddick
339
biddick  

Unfortunately Metatrader4 doesnt' have walk forward optimization and testing.

dimmie
3
dimmie  
N0talent:
Thanks for the advice...im useing tickdata from Ducas with 99% modling quality atm. I cant quite say, how close this get to reality, but i hope its close enough ^^

Hi,

Are you using the programs provided by birt on eareview to get the 99% modelling quality?

spirit100
14
spirit100  

Dalaytime for sending orders about 30 min??? in Backtesting / Reallive

Hello Traders,

at the moment i backtest some strategies. In the visual screen you can see when your signals are generated. in the result report i see the orders are send 30 minutes later than the signal was. is that korrekt and only in backtest mode?

How is the delay in Realmode? have you some experience? my broker is alpari.

regards

Min Li
2715
Min Li  
spirit100:
Hello Traders,

at the moment i backtest some strategies. In the visual screen you can see when your signals are generated. in the result report i see the orders are send 30 minutes later than the signal was. is that korrekt and only in backtest mode?

How is the delay in Realmode? have you some experience? my broker is alpari.

regards

In back testing it assumes perfect execution so not possible for any delay. The only reason I can think of is that your EA probably choose to submit trades after a bar completely closed (in other words, new bar open) and you are testing 30m chart. The signal can only identified at the end of 30 minutes, not the beginning. So not really late.

If above is not the reason and it does open trades 30m late in back testing, in forward trading it should be the same thing.