Auto BackTesting Help | Forex Factory

Options

Permalinks
Print Thread

Similar Threads

MetaTrader Expert Advisor BUILDER for backtesting and auto trading 4 replies

Backtesting help 3 replies

Need help trying to devise method of backtesting system using straddling EA 0 replies

Backtesting Systems - help. 0 replies

Need help on backtesting and micro-lot 2 replies

Auto BackTesting Help

  • #1
  • First Post: May 2, 2007 12:28pm May 2, 2007 12:28pm
Can anyone help me with this? I'm planning to learn to program on MetaTrader's Editor so that I can backtest like 4years of data, instead of like 100 manual trades.

Can anyone with experience leave their two cents?
  • #2
  • May 2, 2007 12:33pm May 2, 2007 12:33pm
P.S. I'm not talking about the sticky note on this forum, just some general/specific tips you guys have encountered during this journey/process. Thanks!
  • #3
  • May 2, 2007 12:37pm May 2, 2007 12:37pm
Sure, personally, I think the MetaTrader backtest system isn't worth anything. I'd recommend learning a better back-test system like wealth lab or amibroker.

Quoting Skyhr
Can anyone help me with this? I'm planning to learn to program on MetaTrader's Editor so that I can backtest like 4years of data, instead of like 100 manual trades.

Can anyone with experience leave their two cents?
  • #4
  • May 2, 2007 1:36pm May 2, 2007 1:36pm
Backtesting i great, but only usefull to determine the long term stability of your system. Foward testing and live trading will both behave differently.
Keep it simple stoopid....
  • #5
  • May 3, 2007 2:47am May 3, 2007 2:47am
Quoting Kurka Fund
Backtesting i great, but only usefull to determine the long term stability of your system. Foward testing and live trading will both behave differently.
I know. But I really need to know how much of an edge the systems I create and will be using really do have.
  • #6
  • May 3, 2007 2:49am May 3, 2007 2:49am
Quoting mrmikal
Sure, personally, I think the MetaTrader backtest system isn't worth anything. I'd recommend learning a better back-test system like wealth lab or amibroker.
Really? Why not?

For the other two - what language do they use (easy language?) and do I have to pay for them? (I do not want to pay for them, I'd rather use the free Metatrader even if its backtesting system may not be as good. It'll also be good since I'll be using Metatrader as my charting and will be coding some indicators)
  • #7
  • May 3, 2007 2:56am May 3, 2007 2:56am
mt4 is not worth your time for backtesting

i have run an ea forward for weeks and then spent the time to get my modeling quality to 99% and then back tested the same forward test I ran.

it was off by over 200%

mt4 isnt designed for back testing, there is a logic flaw built into the software. I actually believe they are happy to give you better results then you would get actually to attract more people to the market.
  • #8
  • May 4, 2007 3:20am May 4, 2007 3:20am
Quoting witchazel
mt4 is not worth your time for backtesting

i have run an ea forward for weeks and then spent the time to get my modeling quality to 99% and then back tested the same forward test I ran.

it was off by over 200%

mt4 isnt designed for back testing, there is a logic flaw built into the software. I actually believe they are happy to give you better results then you would get actually to attract more people to the market.
What do you suggest I use that I can use for free - or something that's cheap? I cannot use Tradestation because I'm not 21 yet. I don't know much about this Wealth Lab thing, what do you personally use?
  • #9
  • May 4, 2007 4:00am May 4, 2007 4:00am
I have the same "bias" that MT4 backtesting suck/erronuous.

I was following a simple MACD indicator on D1 charts. Manually paper traded it and was making good pips. But backtesting shows otherwise?

I think I would use EA to auto enter trades, and I would manually manage it instead.

ie, its hard to be awake 24 hours to wait for the perfect setup. But surely, I can be alert enough when a position is already taken by the EA. This takes 50% of the effort out of the equation.
  • #10
  • May 5, 2007 9:35pm May 5, 2007 9:35pm
my group actually colllects ticks from live and demo and puts them in a database. from there we just use ruby or python scripts.

just make sure when you capture the ticks you get ask and bid
  • #11
  • May 5, 2007 9:40pm May 5, 2007 9:40pm
Quoting witchazel
my group actually colllects ticks from live and demo and puts them in a database. from there we just use ruby or python scripts.

just make sure when you capture the ticks you get ask and bid
Interesting that you use Python/Ruby (or any other programming language) for this backtesting.. Interesting!

I am a programmer as well. Perhaps I could do just the same.
  • #12
  • Last Post: May 6, 2007 7:31am May 6, 2007 7:31am
  •  iya
  • | Joined Oct 2006 | Status: always protect yourself | 129 Posts
Quoting witchazel
my group actually colllects ticks from live and demo and puts them in a database. from there we just use ruby or python scripts.

just make sure when you capture the ticks you get ask and bid
If you have a tick database, you can also use Metatrader for reliable backtesting. There was a thread somewhere for this project. If an EA only trades once every candle, backtesting with "open prices" only is also very reliable. One sure sign of a faulty backtest is, when several orders are opened and closed on the same candle.

That's not to say that Python and Ruby don't rock.
0 traders viewing now