Menu
Sister Sites

Options

Bookmark Thread

First Page First Unread Last Page Last Post

Print Thread

Similar Threads

MetaTrader Backtesting questions 4 replies

MetaTrader Expert Advisor BUILDER for backtesting and auto trading 4 replies

metatrader- bad backtesting or bad broker quotations? 0 replies

Tickdata during visual strategytester modus 0 replies

MetaTrader BackTesting, Reliable or Not?? 17 replies

Attachments: How to get 99% tickdata for metatrader backtesting?
Exit Attachments

How to get 99% tickdata for metatrader backtesting?

  • Post #1
  • Quote
  • First Post: Apr 16, 2011 1:57pm Apr 16, 2011 1:57pm
after lot of hardwork and study i created an EA and made 600 pips today on live testing.

however when i add the Ea to backtesting the same week i only made 25 pips.

so my live testing was better then backtesting. then i checked out the modeeling quality and saw that it was only 90% accurate.

i checked google and saw some info how to obtain 99% tickdata. i tried that method. it creates the fxt file, but the HSt files are much to small.

after i copied those over and did a backtest my modelquality went to 25%.

so my question is, how can i obtain 99% tickdata for metatrader to backcheck my EA?

and also what went wrong with the conversion?

i used the dukascopFXt script or the Jforex script to convert csv files into hst and fxt files. but it does not work.

somehow the script creates the fxt files correctly but does not add very small Hst files.

thanks

w
HST files are not used anywhere in this process.

  1. Create the fxt files for the TF you want to test on. These contain tick data. Move the files to ../tester/history/. They should be write protected.
  2. Prior to running strategy tester, run the 'birt's patch.ex4' script on the symbol you intend to run the test on. I think this will prevent MT4 from creating new fxt files from the hst files with simulated data.
  3. Run your EA in strategy tester now. It should use the fxt file created by birts scripts that contain actual tick data.

thanks a lot! Xaphod!

your the metatradergod on forexfactory!

pure genius!

bye the way sorry have no replied in other thread on the icustom stuff. but i have been so very busy and made so much progress since then.

suddenly all pieces came falling together. i can now make arrays, baskets etc.

when all is completely done i wil show you the result. i am going further working on this now.

thanks again... a lot!
Quoting trade4fun
thanks a lot! Xaphod!
....
suddenly all pieces came falling together. i can now make arrays, baskets etc.
You are welcome. I'm glad you have made progress with your programming skills.

/X
thanks. i didnt made 600 pips a day. it was in a week . now that i just reread my thread.

just so happy, that what i had in my mind i am now party able to do. need more study though. so that is what i am going to do

thanks
Quoting Xaphod
... Prior to running strategy tester, run the 'birt's patch.ex4' script on the symbol you intend to run the test on ...
Is there an alternative to Birt's patch for build 509?
Quoting Unit01
{quote} Is there an alternative to Birt's patch for build 509?
Yes, Tick Data Suite also by Birt but you have to buy it . . . also have a look at TickStory
20 pips a day isn't too much to ask . . .
anybody that knows knows that backtesting does not work.

it is like learning to drive a car while looking in the rearview mirror.

i have personally written code in metatrader's backtesting module that turns 1k into billions of dollars in only a matter of months.

any coder can fit their code to yesterday's data, omg.

but go ahead and try to make your flimsy functions fit tomorrow's prices.

go ahead.

happy trading,

zero/.
contact zero 510-684-1114
Quoting slam.dunk
anybody that knows knows that backtesting does not work. .
It really depends on what you use it for and how you use it . . . it most certainly does work and gives useful information. If you want it to tell you garbage then yes it can be made to do that, but why do you go looking for garbage ?
20 pips a day isn't too much to ask . . .
Quoting RaptorUK
{quote} It really depends on what you use it for and how you use it . . . it most certainly does work and gives useful information. If you want it to tell you garbage then yes it can be made to do that, but why do you go looking for garbage ?
Thanks Raptor : ) Got TickStory running!

Do you know if there is a way to move backwards in MT4's Strategy Tester by a tick/candle or to start the testing at a specific time of day?
I don't think it's implemented. You can move the chart back with Shift+F12 but that doesn't make the already drawn candles disappear.

Neither could I find a way to start backtesting e.g. at the middle of the day.
Hello, you should sign up for an account and download a trial FXCM platforms Ninjatrader free. FXCM account you use to sign in and open Ticks chart. Strengths: Quality of integrated options, Ticks automatically run continuously update and not have to download any historical data. Important: It's free.

233T eurjpy
Attached Image (click to enlarge)
Click to Enlarge Name: untitled.PNG Size: 85 KB
Trading in the zone
On the subject of price feed & modelling quality!

I have an EA that works perfectly over a 5 year testing period with forex.com brokers, it trades exclusively 4H GOLD.

However, I have tried it on 20 other broker's demos, it just makes zero profit, completely different results (strange)!

Does anyone have an explanation, why on earth an EA would perform perfectly with one Broker & crash on 20 other Brokers.

The prices & spreads of the Brokers look fairly similar, nothing to cause major concern. Also, I may add the EA is not a

scalper, so not a particularly spread sensitive EA.


Can someone explain in detail: the difference in Broker prices! How they are arrived at! Importantly, how & why they effect EA performance!


MANY THANKS
Quoting GreatForex
On the subject of price feed & modelling quality! I have an EA that works perfectly over a 5 year testing period with forex.com brokers, it trades exclusively 4H GOLD. However, I have tried it on 20 other broker's demos, it just makes zero profit, completely different results (strange)! Does anyone have an explanation, why on earth an EA would perform perfectly with one Broker & crash on 20 other Brokers. The prices & spreads of the Brokers look fairly similar, nothing to cause major concern. Also, I may add the EA is not a scalper, so not a particularly...
Your very first concern when you test on H4 is that brokers may have different candle time.
MT4/MT5 EAs/Indicators/Alerts coder
''Candle Time'' interesting can you expand, on your suggestion please?

By the way, my EA relies on the ratio of Green to Red Candles to predict the next set of candle.
Quoting GreatForex
''Candle Time'' interesting can you expand, on your suggestion please? By the way, my EA relies on the ratio of Green to Red Candles to predict the next set of candle.
I don't think it's a suggestion more of a concern. D1 and H4 bars/candles are not reliable, Brokers using different timezones will have different D1 start times and also possibly different H4 start times . . . H1 and below will always have the same start times.
20 pips a day isn't too much to ask . . .
So, the obvious question is, if I am having success with forex.com how can I find a similar broker?
What is the criteria?
Who has the same candle time?
Quoting GreatForex
So, the obvious question is, if I am having success with forex.com how can I find a similar broker? What is the criteria? Who has the same candle time?
You have no guarantee that another broker will give the same result. Anyway, the only valid result (even on Forex . com) will be when you trade LIVE.

Your time question: you must find a broker with the same candle time (ie. GMT offset).
MT4/MT5 EAs/Indicators/Alerts coder
Quoting GreatForex
Can someone explain in detail: the difference in Broker prices! How they are arrived at! Importantly, how & why they effect EA performance! MANY THANKS
Check testing statements. Total trades number. Profit-loss series. Any difference in these statements can help you to find the reason.
0 traders viewing now