Renko Live and Recording ticks

 
Hi everybody.
I have been playing around with EAs for quite a while. Had no great success with ones trading on standard charts. Then friend introduced me to Renko charts. I started to build EAs which were making terrific profits trading on Renko charts but only in the backtest. In live demos they were failing miserably.
Obviously simulated data of back tester was far away from real data.
Additionally Renko Chart software I was using was creating artificial Opens and Closes
which don't exists in real life.
Eventually I rewrote Renko charts software, I wrote my own tick recorders, I wrote my own fxt file generators to create Renko back-tests using real ticks.
After all that I tested my all renko based EAs and they all failed miserably. I knew that I was on right track.
I started to write new EAs for my Renko and couple proved to be promising.
The problem is that there is some loss of data in the system.
There is sometimes quite a difference between live charts running on different computers (I have VPS as well) as well as with charts recreated from recorded ticks.
Obviously data gets lost in Live Charts as well as in tick recorders.
So the results of the same EA with the same settings running on different computers (the same broker), as well as back-tests from recorded ticks are different.

Have anybody any experience in this area and maybe advice?

Jake
        

 
Yeah... It's Hopeless. This is the "Gorilla in the Room". The new guy comes in and startled as to why no-body's acknowledging what he's seeing. I was no different. Give it some time and you too would pay no attention to this Big Black Muscular Beast which would break every bone in your body later.
 
Jake2701:

Hi Jake and everybody!

I develop and test experts on renko charts (to tell you the truth, my favourite game is the MT4 strategy tester, i play with it all the time).

I do it in an MT4 which was downloaded from this site. It was never connected to any broker, so it sees spread to be 0.

I always test experts in this environment at first, spread calculation is added later only, if the strategy seems to work. I recommend this method, below you will see why.

If you backtest expert on renko blocks, you have to keep its specifications in mind.

I mean:

- all the blocks may have a wick which can be (2*blocksize) - 1 point. They are not visible on classic renko chart.

- (while we don't count with spread) the minimal stoploss and trailingstop (if used) on backtest must be 2*blocksize. Both of them always must be an integer number * blocksize, otherwise you won't get real results. All the orders have to stop out at a close value of a block. After making a backtest, press 'open chart' in strategy tester and check it.

- you get the open price of the block only when it closes. In testing you get the whole blocks like ticks. In trading Close[0] is the current price while it doesn't reach one of the close values of the current block.

Thats why i apply this method in code:

int i; double c = Close[i] ; o = Open[i]; if(Istesting) {i=0;} else {i=1;}

I use a modified version of RenkoLiveChart v3.2 to generate block chart, i don't know how others work. It's not perfekt, but better than any script what i tried.

I am not a programmer, only a gamer, i am trying to help, i hope it's not too late.

Sorry for my beginner English.

Best regards

Reason: