Proper backtest info very important! - Page 3

stuve · #21 12-18-2008, 03:30 AM

Yes, Now it back to normal again!! puuh!

alan · #22 12-18-2008, 03:46 AM

Excellent...happy trading!

T$$ · #23 01-08-2009, 07:00 PM

Hi Earl(or anybody else who knows)
I am trying to do some backtesting and I keep getting the following error message:
"TestGenerator: no history data 'EURUSD1'"

I have looked on the internet and through the posts and help centers for a solution.
Haven't found it yet so I am posting here.
If anybody has some wisdom for me I appreciate it.

T$

alan · #24 01-10-2009, 12:16 AM

Have a look at this post I made a while back:

http://alansforexblog.com/2008/04/16...rategy-tester/

Basically it boils down to you needing to download history data for the eur/usd.

dreamy77 · #25 01-16-2009, 03:13 PM

I tried to backtest today but I cant get above 25% modeling quality. The strategy isnt even using the LRR function correctly. Anyone know how to get the backtest to be more accurate? (I believe i have the data..)

T$$ · #26 01-18-2009, 01:39 PM

Hi Earl -
I am having major trouble trying to backtest. I keep getting "no data" message in the journal. I am using FXDD's platform in my attempt to backtest.
Do you happen to know what I am doing wrong?
Thanks in advance for your reply!
T$

soldierboy · #27 01-23-2009, 05:25 PM

Quote:

Originally Posted by Earl

I opened this thread to stress how important your back test
info is.

Without the right info...you could be going the wrong way..

Make sure you have download it correctly, and update it often.

1) Click on Tools
2) click on Options
3) click on the charts tab
4) change both "max bars in history" and "max bars in chart" to "99999999999"
5) close options window
6)click on tools
7)click on History Center
8)click on EURUSD
9) double click on EURUSD M1
10) click download

hope this helps....

Hi mate.

First post here.

I've been using fapturbo for a couple of weeks, my broker has banned the use of the scalping strategy but not the long term one.

Anyway I've decided to run my own backtest of the scalping strategy.

I've downloaded MT4 platform from http://www.metaquotes.net/downloads

I removed all the old data files .hst then downloaded the 1M Data from alpari imported the data and then converted the data to 15M Time Frame.

I've then run the back test and i've discovered that the results are not as good as I would of expected. The test was done from the period 3Jan2005 to 23Jan2009

I'm a bit suspicious of backtesting, because it is very sensitive to the data you have. When comparing backtest to live results the trades taken do not always match.

Expected Payoff is 1.71

A 16% Return in the Bactested Period which is a 5 Year Period which equates to 3.2% a year. Not that impressive.

Seems that some serious optimization/curve fitting has been done which has cuased the equity curve to really jump in the last part of the graph otherwise it is relatively flat.

I disabled MM in this test and used a fixed lot size of 0.1

Note there are some drawdown periods on the chart. If using leverage in those periods then I would suggest an account would of been destroyed.

Overall the equity curve is positive in the tested period which is a plus. However using a high leverage amount could cause concern if going through a significant drawdown period.

Here is the full report

http://www.sendspace.com/file/k18ufi

dynamo · #28 01-30-2009, 08:43 AM

Quote:

Originally Posted by stuve

Alan:
yes, but iam not running any backtests at the moment..

I try to reinstall mt4

Hi Stuve

The problem you have is because after you run the backtests you need to turn the chart settings back to normal. If you leave the settings at 999999999 (or whatever MT4 has set as your maximum), the program will be reading and writing the data at the same time and so gets stuck in a program loop until it runs out of memory and overloads the processor.

If you bear with it and patiently change the settings back to a lower number of bars in the charts (it may take a while as you have to wait for your pc to catch up with your clicks) then shut down and restart MT4, it should go back to the usual speed.

I have had this before on Vista, and now my MT4 is running back at 20k memory, and my CPU is back to 14%

Ken Long · #29 02-15-2009, 12:58 PM

999999999 history might be imperitive for long term testing on the 1M bars, but it eats ram!! Likewise none of the programs need this much data to run. My main demo account has over 40 windows open. I keep the bars in history and bars in charts at 5000 each, just so it can run 24/5 while I use the computer for other things. My laptop only has 1g of ram and it cant handle this much data without overloading. I do all my backtesting on the weekends or when the programs are turned off, and reset the bar counts and clean my registry and defragment before restarting for the weeks operations. I find that backtesting clogs my registry with unnecessary and fragmented files, and that cleaning regularly helps keep it running smoothly.

mystikdragon7 · #30 02-16-2009, 09:45 AM

Quote:

Originally Posted by Ken Long

999999999 history might be imperitive for long term testing on the 1M bars, but it eats ram!! Likewise none of the programs need this much data to run. My main demo account has over 40 windows open. I keep the bars in history and bars in charts at 5000 each, just so it can run 24/5 while I use the computer for other things. My laptop only has 1g of ram and it cant handle this much data without overloading. I do all my backtesting on the weekends or when the programs are turned off, and reset the bar counts and clean my registry and defragment before restarting for the weeks operations. I find that backtesting clogs my registry with unnecessary and fragmented files, and that cleaning regularly helps keep it running smoothly.

You know I often wondered if there was a memory leak somewhere in the FAPS program. Even on my desktop with 4gigs of memory it will make the computer run like a pig. I usually never do defrags or cleaning, guess I'll try that and see what happens.