Topic: Almost 2 months of FSB and FST testing

Hi,

I run tests of different strategies since August 2010. There are 1-minute and generated randomly with FSB. Hand picked only the best looking. Only 1 profitable from more than 20.
Please have a look:

http://screencast.com/t/VyrS6bWGti8

Re: Almost 2 months of FSB and FST testing

I agree...

It's very difficult to get a working strategy. Even an awesome backtesting results will not make the strategy profitable in the real world environment. Usually good backtested strategies will fail in forward testing.

Here are my demo testig results: http://www.mt4i.com/users/zenoni/stats

I have developed and tested over 60 strategies, but I have found only 4-5 working strategies.

3 (edited by acharyadas 2010-10-05 20:14:40)

Re: Almost 2 months of FSB and FST testing

Thanks for kind words, footon. 1M is fastest to find out if strategy is good or not. If I would use 1d they I would have to wait 2 years smile In August and September I have some 5m strats, but they are too slow.
Do you think that using 15m would be better or less random?

I used 30k bars for generating strategies and also OOS 30%.

Settings below (sometimes I used perm s/l with 50 or 60 pips)
http://screencast.com/t/et8aIUjxDH64
and
http://screencast.com/t/FSXuxumgXRb3

My next move would be to change the losers faster and not hope they will get better smile as you can see on the chart. I plan to gather few of such strategies and run some micro real account.

Winning strategy gives 6% a month which is not bad. If I would have at least 2 such strategies working for 1 year, it would be good.

Re: Almost 2 months of FSB and FST testing

zenoni wrote:

I agree...

It's very difficult to get a working strategy. Even an awesome backtesting results will not make the strategy profitable in the real world environment. Usually good backtested strategies will fail in formward testing.

Here are my demo testig results: http://www.mt4i.com/users/zenoni/stats

I have developed and tested over 60 strategies, but I have found only 4-5 working strategies.

Thanks - this is encouraging - you at least in green, but my overall result is -31.51%.

Re: Almost 2 months of FSB and FST testing

You guys make me feel like a chump. I was doing OK for about a week using some 1H strats. I abandoned those when they started to falter and the newer version of FSB came out. (Didn't bother to rebuild) I'm not having a lot of luck with over 100 different schema. I have used the generator, and I find that it works very well for helping to sreamline a "good idea". I have tried manual running through the ticks, and I've tried getting very paticular into the nuts and bolts. I have had some real downturns, but some significant results to keep my hopes up.
Here's the latest I was working on. I used someone's post from earlier and tweaked a bit. the pop up shows full lots and the test shows microlots from Sept 1 through today.
http://s4.postimage.org/igsUJ.jpg
COuld I be "over optimizing"?

Post's attachments

20101005 eurusd M15 ziet generate 003 {1346}.xml 6.98 kb, 64 downloads since 2010-10-06 

You don't have the permssions to download the attachments of this post.

Re: Almost 2 months of FSB and FST testing

I improved my backtesting results when I started to use slippage in FSB. I'm always use 1-1.5 pip slippage in FSB to get more realistic estimations.

Market -> Edit Instruments > slippage

Secons issue is related to the third-party custom indicators. I think that HTF indicators will give a false backtesting results. Use only standard indicators or those custom indicators that provided by Popov itself.

Re: Almost 2 months of FSB and FST testing

thanks to all. I will enter this as a forum note on the "indicators" forum, but any chance we can get a zip of the "good indicators"?

8 (edited by acharyadas 2010-10-06 19:01:26)

Re: Almost 2 months of FSB and FST testing

Update with my test:

http://screencast.com/t/ABczr2UHtdi

+5.41% in October so far

Anyone did 15m tests in comparison with 1m or 5m timeframe?

Re: Almost 2 months of FSB and FST testing

Another day:
http://screencast.com/t/Y8cInnlJPU

And the profit in this month:
http://screencast.com/t/vlqQNTNesLU1

Re: Almost 2 months of FSB and FST testing

End of 1-min charts for me:

http://screencast.com/t/euY0vUPy

And months:
http://screencast.com/t/qabczNBE

Moving to 15-min frame then.

Re: Almost 2 months of FSB and FST testing

Well, they were not consistent - in one time they had +15% in given month... just after the end of the month they went under 0. I just have to check 15-min TF for 1-2 months and see the difference.