Menu
Sister Sites

Options

Bookmark Thread

First Page First Unread Last Page Last Post

Print Thread

Similar Threads

Backtesting optimization with wrong parameter values 0 replies

Backtesting and Optimization mode: what? how? why? 2 replies

Backtesting/Optimization platforms 1 reply

Optimization question - Optimizing vs Backtesting? 2 replies

Backtesting/Optimization Q&A 3 replies

Attachments: Free Backtesting and Optimization Service
Exit Attachments

Free Backtesting and Optimization Service

Quoting forexsaint
Hi traderkid, AdamnN is right..with EA and indi's etc i wont be able to test it.
may be if i had those pogramming skills i may have tried ! but i dont !
May be you can start a Ea Making request Thread with all rules etc and some generous programmer may help you.
I tried to use Metatrader running on Crossover, the application that allows PC programs to run on Macs without using Windows. But it did not work. So, I am stuck operating my method manually. So, I set alerts to notify me when a price point is reached, then I must manually place or change orders. It's a hassle, particularly when alerted in the middle of the night, I have to get up and deal with it. Pain!! Thanks for your reply. I suppose I should post a request for MT4 programmer assistance. Again, many thanks, and I hope your trading is successful!!!!!!!
Alright
this thread is again open for anyone needing help with EA optimization and testing.
I have this server till may end and can test for anybody help.
Better News is that, the tests and optimizations will be QUICK as well as accurate>90% as i am testing this on tick data also.
Hope this helps.
Regards
100 Fold Challenge->Interested? ->https://www.forexfactory.com/thread/32152
ForexSaint,

I am interested in doing my own optimization and backtesting, and am looking to buy a new computer to do so. What would you recommend to look for in a computer for this purpose. For Example - Would a 4 core system be worth spending the money on vs and dual core, and what size of memory would I need, as I plan on doing this for several pairs 10+.

Any advice would be appreciated.
I would opt for quad core, Metatrader 5 allows for multi-core optimization.
Thanks RR for the input.
Well all depends upon usage.
new pc just for testing..doesn't make sense imo,
You can be better off taking a dedicated high end server for that and do whatever you want test without power bills and net issues from the comfort of your pc or pda(remoting it)

If you want that to be used for other purposes too.. (De)Mark your other requirements (lol) and then accordingly buy one.
Of course though quad core is better. I havent yet looked at intel i5 and i7 stuff . Speed wise they will be lot better!
Further , idk whether mt5 will have proper testing or crap like mt4 (except older versions with tick data usage in backtesting) ..Ronald knows more about it i think!

Just cores wont matter too...RAM speed..FSB and all stuff is important .. !
5200rpm HDD with ultra high RAm (DDR31333) +some 4Ghz multi-core processor +1333 Dualchannal multiple DIMM supporting MBoard +high end 2gb graphics card ...would be like putting water as fuel in ferrari ! This becomes esp important when data usage/read/write and stuff is involved. HDD is weakest/slowest point in every system ! Just an example.
An PShop lover would be at most loss working on a pc like above.

Think about it and then go for it.
Testing will also be less in future...after you will find the grail ..
except if you are like RR.. who will die if data is taken off him and who will one day write... "i trade with 100 EA and 100 are shared on FF" but most people still not making money off them because RR only knows the secret..he doesn't share that ..lol (i have no offensive intentions here... i know RR is laughin now)

Hope this helped (or waster your time)
whatever.. Regards
100 Fold Challenge->Interested? ->https://www.forexfactory.com/thread/32152
Quoting Ronald Raygun
If you read the article a little more closely, you would realize that the history center in metatrader (where the backtester gets its data) does not in fact store tick data. The most granular data it stores is M1 data. Once I discovered that, I moved my backtesting off of Metatrader and into a small program I had written in C# and I just feed that from an SQL database.

Is backtesting not accurate enough when only using M1 data?

Are there any other free alternatives to metatrader that I can test mql4 EA's in for free?

Forexsaint: what are you using to backtest which is allowing you to use ticks?
The greatest force on earth is compound interest.
Backtesting is okay when using M1 data. It could be better if you used tick data.
Quoting sebastionay
Is backtesting not accurate enough when only using M1 data?

Are there any other free alternatives to metatrader that I can test mql4 EA's in for free?

Forexsaint: what are you using to backtest which is allowing you to use ticks?
Ive collected 1 year of historical m1 data from alpari but can't get above 90% modelling quality with that. perhaps I can better quality with the oanada data.
The greatest force on earth is compound interest.
Quoting sebastionay
Forexsaint: what are you using to backtest which is allowing you to use ticks?
Some previous version of 2007, of mt4, used to do that. I got one from FF itself, i think.
100 Fold Challenge->Interested? ->https://www.forexfactory.com/thread/32152
Hi ForexSaint,

I have two new promising EAs which I am optimising atm but it takes such a long time as you say/know and I have not even completed 20% yet.

Can I send the first one to you pls?

If the EAs are proven successfull, I have no problem sharing them with you as you have made a big contribution by proving that the EA is robust enough to trade!

Let me know yoour thoughts pls?
  • Post #32
  • Quote
  • Edited 9:06am May 29, 2012 8:55am | Edited 9:06am
Posting to inform....
As i have re-hauled my backtesting stuff & data, this service is again live.




P.S.
- Testing available on major pairs only
- With 1-3 years of tick data only
- With both ranging and trending market conditions (mainly from 2008 to 2010)
100 Fold Challenge->Interested? ->https://www.forexfactory.com/thread/32152
Hi forexsaint,


Is the service still on? please check your PM.
wo-yoy! wo-yoy! wo-yoy! wo-yoi! wo-yoy-yoy-yoy!
i want to back test some strategies. can u plz report yur finding.
1st strategy-enter to the breakout of previous candle high or low only if previous candle difference between open and close is less then 50% of whole range,

  1. stop loss-low or high of previous candle.
  2. take profit-3 times yur stop loss
  3. time frame-daily

2nd strategy-high volatility breakout.enter on the breakout of 100% of previous day range. stop loss low or high of entry day. try different time frames

Hey guys.

I have coded an EA, and run the backtest, so this is the backtesting result. Then i did some minor amendment of the coding to see if i could get more pips. So in the end, i change back the code to its origin rules. However, the backtesting result is not longer the same as before. I am not sure what is going on. Can someone look and check or backtest this EA? I want the backtesting result back as below.

Attached Image (click to enlarge)
Click to Enlarge Name: Scalping V4.jpg Size: 104 KB
Attached File(s)
File Type: mq4 Scalping V4.mq4   22 KB | 352 downloads
File Type: rar Backtesting result.rar   178 KB | 297 downloads
Quoting invincible1
i want to back test some strategies. can u plz report yur finding. 1st strategy-enter to the breakout of previous candle high or low only if previous candle difference between open and close is less then 50% of whole range, stop loss-low or high of previous candle. take profit-3 times yur stop loss time frame-daily 2nd strategy-high volatility breakout.enter on the breakout of 100% of previous day range. stop loss low or high of entry day. try different time frames
Do you really think i (or even some1 else) has time to write(spend to get written) EA's of your strategy and then spend time on backesting...
that too for a 2 post newbie.????
I DONT !
Firstly you gotta read the first post carefully.
Secondly, last time the free offering was available for 2 months in May 2012 and you are replying after an year or so.


Quoting ~bull.bear~
Hey guys. I have coded an EA, and run the backtest, so this is the backtesting result. Then i did some minor amendment of the coding to see if i could get more pips. So in the end, i change back the code to its origin rules. However, the backtesting result is not longer the same as before. I am not sure what is going on. Can someone look and check or backtest this EA? I want the backtesting result back as below. {image} {file} {file}
Though i would have done that for you, i dont have good 2004-mid2005 tick data (thats d date it shows for your test in backtest statement)
So there is no point. Further try increasing backtest test quality(99% is achievable in MT4). How much this backtesting is worth is a different topic though.
100 Fold Challenge->Interested? ->https://www.forexfactory.com/thread/32152
Thanks for your reply. Forexsaint

Actually I also have tick data for backtesting, for model quality up to 99%. But what I want is someone here could backtest it and check if the logic is the same as listed in my backtest result, it doesn't matter for the entry exit price is the same. Importantly, it should work and profitable. I suspect that my Mt4 backtesting data is something wrong.

Appreciate that if you could check it. And one thing, if you have closed this service, please mention in first post. Thanks a lot!
Quoting ~bull.bear~
Thanks for your reply. Forexsaint Actually I also have tick data for backtesting, for model quality up to 99%. But what I want is someone here could backtest it and check if the logic is the same as listed in my backtest result, it doesn't matter for the entry exit price is the same. Importantly, it should work and profitable. I suspect that my Mt4 backtesting data is something wrong. Appreciate that if you could check it. And one thing, if you have closed this service, please mention in first post. Thanks a lot!
From whatever i can understand in your reply ( didn't get a lot of stuff)
and FWIW
here is a backtest of that ea
with settings as provided in backtest report
On GBPUSD
M15 TF
with 500:1 leverage broker
with data period as 2007.08.06 to 2010.03.27
with fixed spread of 2.5 pips
P.S. Log has a lot of order modify errors.(Prolly some current stopout level issues in Ea working)
Attached File(s)
File Type: zip 1st test data.zip   101 KB | 269 downloads
100 Fold Challenge->Interested? ->https://www.forexfactory.com/thread/32152
Quoting forexsaint
{quote} From whatever i can understand in your reply ( didn't get a lot of stuff) and FWIW here is a backtest of that ea with settings as provided in backtest report On GBPUSD M15 TF with 500:1 leverage broker with data period as 2007.08.06 to 2010.03.27 with fixed spread of 2.5 pips P.S. Log has a lot of order modify errors.(Prolly some current stopout level issues in Ea working) {file}
Thanks you very much!
Yes, it has many order modify errors. It is normal because SL or TP still remain the same price.
Quoting ~bull.bear~
{quote} Thanks you very much! Yes, it has many order modify errors. It is normal because SL or TP still remain the same price.
You are welcome.


@Others.
I spruced up the setup and this free service is again open for help for next 2 months
100 Fold Challenge->Interested? ->https://www.forexfactory.com/thread/32152
0 traders viewing now