Another Optimization issue.

New comment
[Deleted]  

Hello all, I'm trying like crazy to really understand the optimization process and have run into a stumbling block that is throwing me off.

When i run a basic optimization i get pretty nice results, I'm only running a 3-4 month time period just so i can see if I've worked out a few bugs. Now i didn't code these, they were given to me to back test/ optimize for a friend. I run my optimization and i get nice results, i then take the best results and run it on a back test to get the report and see how well it did. Once i run the back test it completely does the opposite from the optimization. It destroys the account and ends close to what i started with. Why would the results from the optimization vary so much on the back test? its the same time period and everything. It is a demo account I'm using, using Dukascopy data, using tick data suite, as far as i know I'm not missing anything, I've converted csv files to fxt i should be OK but this is really making me puzzled, it started happening about 2-3 days ago, before it would produce the same awesome results.


Any help would be great.

 

Did you let liveUpdate update your version of mt4 a few days ago ?

 
mbro777:

Hello all, I'm trying like crazy to really understand the optimization process and have run into a stumbling block that is throwing me off.

When i run a basic optimization i get pretty nice results, I'm only running a 3-4 month time period just so i can see if I've worked out a few bugs. Now i didn't code these, they were given to me to back test/ optimize for a friend. I run my optimization and i get nice results, i then take the best results and run it on a back test to get the report and see how well it did. Once i run the back test it completely does the opposite from the optimization. It destroys the account and ends close to what i started with. Why would the results from the optimization vary so much on the back test? its the same time period and everything. It is a demo account I'm using, using Dukascopy data, using tick data suite, as far as i know I'm not missing anything, I've converted csv files to fxt i should be OK but this is really making me puzzled, it started happening about 2-3 days ago, before it would produce the same awesome results.


Any help would be great.

I bet the spread changed between the tests. An adverse spread can wreck a strategy, especially a heavily optimized one.
[Deleted]  
SDC:

Did you let liveUpdate update your version of mt4 a few days ago ?


Yes, got the new updates. Thanks

Dabbler,

Should i leave the spread at 0? or 1? i'm doing the EUR/USD as my test.

 
Well the reason I asked, which may be irrelevent to your issue, a few days ago I discovered the update deleted all the chart history and the recent chart history was replaced with chart history that comes with the install, so I was wondering, when you run your backtest, are you using mt4 strategy tester and if you are, what chart history is strategy tester using ?
 
mbro777:

Yes, got the new updates. Thanks

Dabbler,

Should i leave the spread at 0? or 1? i'm doing the EUR/USD as my test.

I am not able to adjust the spread when doing a back test. It is whatever happens to be there at the time. I know that there are tricks to login to bogus accounts and run a script to change the spread but I don't do that very often. On my accounts I typically see a EURUSD spread of 1.6 pips. If you are able to set the spread then set it to what is normal for your account.
[Deleted]  

Maybe this is off topic but I'm in the strategy tester and on the far right is a box that say symbol properties, it says spread is 9, i don't know if this is an error or it.9? man there is so much to know about this stuff.


SDC,

i just deleted old data and ran new data, will see if this takes care of the problem, thanks for the tip.

[Deleted]  
forget the spread, its a 5 digit broker and it changes, jeesh i a moron.
New comment