Backtesting - Real Tick Data / Tross - page 5
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
Once used the pair is EUR.USD,and once take it is GBP.USD,
Not sure I understand what you mean.
Backtested on EurUsd, can't backtest GbPUsd.
I want to post a caveat here. The tick data from Disktrading is representative data... a compendium of data from various brokerages. The backtest results can be misleading when this kind of data is used. Please be careful when you use these results to base your live trading on.
Thanks to Tross for doing this laborious task.
Tross,
Have you done any forward testing yet? heave you seen any correlation with your backtesting?
Maji,
I have noticed that the disktrading data is more volatile than the IBFX feed, can you suggest an alternate source?
Not sure I understand what you mean. Backtested on EurUsd, can't backtest GbPUsd.
Thank you tross,
I mean i want bactest on GbpUsd too,
Thank's,
From Post 1
I wanted to have a place to post some real backtesting data.I have and am in the process of collecting real tick data for 2006.
If you have a great backtest for any EA, but you'd like to see it modelled with 99% then post it here and I will run it on my data.
To request a backtest, in your post include;
1. the mq4 (attach it)
2. the settings of Every option, or the detailed backtest report
2. the pair or pairs (check the list)
3. the timeframe
Pairs available as of Dec 29/06;
EURUSD
USDJPY
USDCAD
I will make the first post in a format I think should be used.Hehehe, exactly.
Thanks to you guys for the kind words and support.
I do hope we find some honest gem of an EA here.
No correlations with forward data yet.
Not sure how I'd go about that I guess.
Forward test an EA from Jan1 to Mar1 then run a backtest over the same period from tick data and compare results?
Hmm, I will do that, currently in the process of testing 3 pairs on 12 different indicators over 4 groups of time.
Someone do the math, sounds like I'm going to be backtesting for a few days.
Once I find the best set of indicators, I'm going to mix them with a Dynamic EA that I've been working on.
Then, omg, forward test and correlate.
Probably on a live account, because then we'll really know.
Tross
Craig,
There was a guy selling FXDD's tick data for EURUSD and GBPUSD on ebay a few weeks back. Check it out.
Also, you can go to the MQ4 forum and download the ticksaver program that will save ticks on your computer and create a data file for you to test with.
Another user named cubesteak has started a group to save tick data. You can contact him and see what he can do.
Good luck.
We all need a backester, which shows the same results as a forward test does for the same period and with the same broker. If that will not happen, mq4 for EA would be plaything only and no one lives long enough to develope something real from it and being able to test it through a forward test. Why is it no go? Is it just coincidental?
The algorithm that mq4 uses estimates the tick during any given period. (up to 90%)
If you use tick data, it doesn't have to estimate, it can use the actual tick. (99%)
The final problem is that for a backtester placing orders on tick data, it does not take into account requotes during rapid trend, or inability to open or close orders during that time.
It's always best case with a backtester. In my opinion.
Backtester is used to test your code for logical errors.
It allows you to see differences between different entry and exit strategies.
It doesn't predict the future, or know how well your broker is treating you.
Still useful.
Tross