Help setting up SBB

Post by **milanese** » Sat Dec 06, 2014 7:25 pm

ariel » Sat Dec 06, 2014 7:22 pm wrote:Hello!
Can i do a back test in tester the Empty4?
I tried but dont work in the back test.
But EA work well.

Regards

Ariel :hi:

no, as the Empty4 backtester can not handle multipair EA's..

Cheers

Tommaso

Post by **retireme** » Sun Dec 07, 2014 2:55 am

ariel » Sun Dec 07, 2014 3:22 am wrote:Hello!
Can i do a back test in tester the Empty4?
I tried but dont work in the back test.
But EA work well.

Regards

Ariel :hi:

No you cannot for too many reasons to list here but the main one is... if you backtest to see if an EA is good or not, you are wasting your time. The other reason is that SBB is a multi-trader and backtest can only use one pair.

Here is the backtest scenario for NOOBS (and we were all one at some point):

Load EA into backtester, set preferences and run... it results in one of two paths:

1. EA make s zillions so you immediately open a live account and lose the lot
or
2. EA is an absolute stinker in backtester so is confined to the garbage before it ever sees a forward testing and possibly, you just threw away the Holy Grail...

In conclusion, FORGET backtesting

Rgds
RM

Post by **Jeuro** » Mon Dec 08, 2014 9:07 am

retireme » Sun Dec 07, 2014 4:55 am wrote:
In conclusion, FORGET backtesting

Rgds
RM

Hello Retireme.
I totally disagree with your conclusion.

The only way to know a strategy has potential is to run it 5 to 7 years with past history prices.
Obviously no guarantee that the market would behave exactly like the past, but at least you have an idea what to expect. The other option is to trade the strategy in a forward demo for 5 or 7 years... but then, after convinced that is ok.. no guarantee that the next 5-7 years market would act the same... so, no point in wasting all that time.

Problem is that 99% of trader do not know how to do proper back testing. It is like a science. It requires
way more knowledge then just trading. It requires knowing the strategy inside out, an also a great deal of all the details and specifics of the back testing features and limitations of the specific platform .

I have been doing multiple -pairs back testing a few years in Jforex platform... but only my own strategies.
Has anybody coded SBB in Java for Jforex? I really do not know if the strategy uses a lot of non-conventional indicators and if is very complicated, but would be a good thing if someone is up to the task.

Unluckily I am not a coder. I know just enough to make changes but not to build a code from scratch.

J.

Post by **retireme** » Mon Dec 08, 2014 9:14 am

Jeuro » Mon Dec 08, 2014 5:07 pm wrote:
Hello Retireme.
I totally disagree with your conclusion.

The only way to know a strategy has potential is to run it 5 to 7 years with past history prices.
Obviously no guarantee that the market would behave exactly like the past, but at least you have an idea what to expect. The other option is to trade the strategy in a forward demo for 5 or 7 years... but then, after convinced that is ok.. no guarantee that the next 5-7 years market would act the same... so, no point in wasting all that time.

Problem is that 99% of trader do not know how to do proper back testing. It is like a science. It requires
way more knowledge then just trading. It requires knowing the strategy inside out, an also a great deal of all the details and specifics of the back testing features and limitations of the specific platform .

I have been doing multiple -pairs back testing a few years in Jforex platform... but only my own strategies.
Has anybody coded SBB in Java for Jforex? I really do not know if the strategy uses a lot of non-conventional indicators and if is very complicated, but would be a good thing if someone is up to the task.

Unluckily I am not a coder. I know just enough to make changes but not to build a code from scratch.

J.

Whether you agree or disagree is of no consequence to me, it is your opinion and what I said is mine. Maybe I should have been more clear insofar as this is a Empty4 forum where scripts, indicators and EA's are coded for use in Empty4. When someone posts something in a thread, unless they state specifically that they mean another platform, then it is read as a Empty4 comment. If you are using Jforex, good luck to you... happy backtesting. My opinion still stands, backtesting in Empty4 is a waste of time when using it to evaluate performance.

Rgds
RM

AnotherBrian · Post by **AnotherBrian** » Wed Dec 10, 2014 2:42 am

What the hell, one more person to jump in and offer an opinion....

"Backtesting" is useful in many ways.
Empty4 backtester is shit.
That sums it up.

I only use strategy tester to test EA logic and work out bugs before I forward test on a demo, only to find more errors. BUT the EA has to be only dependant on the single pairs data you're testing.

Back testing is useful. Empty4's backtester is not very useful.

Post by **Jeuro** » Wed Dec 10, 2014 11:30 am

opinions , opinions...

at least in this forum we can amicable disagree.

Even in Empty4 the some of the statistics pulled out of from "some" coded strategies are ok and useful.

We all know that using good ticks past data solve 85% of the deviations found between forward and back-tests. (takes care of the freaking interpolation of the 1m data).

But digressing from the topic of this thread.

Anyway.. the main point of my post was to ask if anybody has coded or has intention to code SBB in Java for Jforex. Then would be easy to get a few years of stats for the sets some are already using.

J.

tcollings · Post by **tcollings** » Fri Dec 12, 2014 12:18 am

Hi guys,

I have been running SBB on a demo account for a couple of weeks now, and it been doing exceptionally well - generating almost 14,000 pips to date.

One thing I noticed that I am repeatedly receiving the same error in my logs.

3 17:25:21.131 Secret Bob's Basket XAUUSD.xp,H1: invalid takeprofit for OrderModify function

Has anyone encountered a similar error?

Thanks