Unhappy Weird results in backtest build 200.. need help! | Forex Factory

Options

Permalinks
Print Thread

Similar Threads

Weird, only M30 timeframe missing candles/data (v4 build 765) 5 replies

200 Ema system - help me build a system ! 19 replies

Visual backtest result different from non-visual backtest 0 replies

Need EA writer to help backtest simple system 7 replies

Backtest results! 7 replies

Attachments: Unhappy Weird results in backtest build 200.. need help!
Exit Attachments

Unhappy Weird results in backtest build 200.. need help!

  • #1
  • First Post: Nov 29, 2006 4:53pm Nov 29, 2006 4:53pm
Hi all.

I've been trying to test the attached EA and keep on getting incredibly bizarre results (you'll know if you test it). They change with every broker and it seems that it may all be messed up because of MT4 build 200. The results I get are in some cases good profit, other cases, total loss (wipes out account).

Could you help out in testing it and posting the backtests for the year 2006 here please? I've been trying all day to get it tested properly but couldn't. I'll have to do some forward testing as soon as I get some reasonably accurate backtest results. The ea is called superscalper. It's supposed to be tested (as most scalpers) in 1M timeframe and for EURUSD only. Also, make sure you test only for 2006 only (not earlier years). You can also use the default parameter of risk or increase it [it's integer value, no limit].

I warn you though. It's a superscalper, so you may end up with too many orders (tens of thousands?)

Thanks for your help..

i_me
Attached File(s)
File Type: zip SuperScalper-test.zip   11 KB | 636 downloads
  • #2
  • Nov 29, 2006 4:56pm Nov 29, 2006 4:56pm
make sure you have accurate 1 m data
  • #3
  • Nov 29, 2006 5:07pm Nov 29, 2006 5:07pm
Here it is. I tested it with my 1M data, which is from Alpari. The modelling quality is 25%, which is the highest possible using 1M data.

andersenws
Attached File(s)
File Type: zip StrategyTester.zip   162 KB | 574 downloads
"Youth is the trustee of prosperity." - Benjamin Disraeli
  • #4
  • Nov 30, 2006 5:37am Nov 30, 2006 5:37am
i_me...

I had the same trouble with the new build also. I redownloaded the data from Alpari after the build and my backtest results are 'very' different from what they were before, from a nice uplsoping profit to a nice downsloping loss. So, I'm starting it all over....Not sure what the cause was/is...
  • #5
  • Dec 4, 2006 8:50am Dec 4, 2006 8:50am
The build 200 downloads the history data as separate files and apparently the data is broken. Some of the days are missing from the history. I can't even backtest my own EAs reliably anymore. I hope they fix the problem very soon.

No matter you use this build or older Alpari data there is no way that you can reliably backtest an M1-based EA with an acceptable degree of reliablity. The simple reason is that MT4 does not keep the tick data. If you want to use such EAs I suggest to forward test them on a demo account for a long time (at least 3 to 6 months).

Good Luck,
Al
  • #6
  • Dec 4, 2006 9:10am Dec 4, 2006 9:10am
Quoting aparsai
The build 200 downloads the history data as separate files and apparently the data is broken. Some of the days are missing from the history. I can't even backtest my own EAs reliably anymore. I hope they fix the problem very soon.

No matter you use this build or older Alpari data there is no way that you can reliably backtest an M1-based EA with an acceptable degree of reliablity. The simple reason is that MT4 does not keep the tick data. If you want to use such EAs I suggest to forward test them on a demo account for a long time (at least 3 to 6 months).

Good Luck,
Al
You're absolutely right. I guess MT4's latest build had destroyed the backtester credibility (not that it was high any way). So indeed, from now on, one needs to use forward testing aggressively to know if an EA is good and profitable.

BTW, Thanks andersenws for the test results. This confirms the fact tha the backtester is less useful with this build particularly for 1M charts. But I'll start to do forward testing from today. It will take rather a long time though
  • #7
  • Dec 4, 2006 9:46am Dec 4, 2006 9:46am
Quoting investor_me
You're absolutely right. I guess MT4's latest build had destroyed the backtester credibility (not that it was high any way). So indeed, from now on, one needs to use forward testing aggressively to know if an EA is good and profitable.

BTW, Thanks andersenws for the test results. This confirms the fact tha the backtester is less useful with this build particularly for 1M charts. But I'll start to do forward testing from today. It will take rather a long time though
Good luck with forward testing. Although it takes more time, it's for sure much more reliable.
  • #8
  • Last Post: Feb 19, 2007 2:32pm Feb 19, 2007 2:32pm
I've setup Scalper v1.0 on 4 different brokers with new demo account and $2,000 in it but they all fail to make any orders.
I've seen some incredible backtesting results (build 202 on all instances of MT4) with all 4 brokers, but again, when I put it on live it fails to create orders. I didn't change any settings except for the lot size (instead of .1 I use 1).
Please help,
Mike
0 traders viewing now