Tester in new build 604 doesn't work correct - page 13

 
moneycode:

I did another install.

this time I found the terminal.exe. after I did new install to D drive it showed it there. Though ti still doesn't show up on the C drive anywhere .

I tried that portable too, but when I add /portable all I get is error. I changed target exactly as documentation was.

tester crawls, critical errors, results are a disaster -50% when they were + 150%. Unable to even get connection, so not sure what will happen at that point.

What a mess they've made. Why would anyone put out an update that doesn't work properly?

Going to be a while to sort all this out.


You have to place a space between terminal.exe and /portable.
 
This one will be better in the long run, but growth seldom comes without pain. If you want a better system then we just have to deal with this stage of it imo.
 
angevoyageur:
You have to place a space between terminal.exe and /portable.



okay thanks I didn't notice the space. That did fix the error that I pic'd

I can't test it fully unit I switch over to admin.

thanks

 
cbalta:

Hi SDC, my difference, 509 test: 5 minutes, 610 test: 5-6 hours, with long period indicators, like some iMA(1000)'s, some Band(300)'s, on M1 tick data test.

some iCustom indicator calling EA's, 509 test: 5 minutes, 610 test: 1-2 hours. i'would be happy with 27 second tests.. tests for year 2013, comparin 509 and 610, to prove migrated code is healthy.

The tests I posted are 1 year 2013/01/01 - 2013/12/31

Are you running your tests in visual mode or non visual mode ?

I'll change the indicator periods to 300 and 1000 and do a new test on build 610

1 Year
BackTest
1M Chart
MT4 build 610 Speed Test
Non Visual Mode
Using Test EA 600+
Using iMA(14) and iMA(21)
Test1
Test2
Test3
11.622 seconds
11.934 seconds
11.481 seconds
Average11.678 seconds

1 Year
BackTest
1M Chart
MT4 build 610 Speed Test
Non Visual Mode
Using Test EA 600+
Using iMA(300) and iMA(1000)
Test1
Test2
Test3
14.711 seconds
14.337 seconds
15.163 seconds
Average 14.737 seconds


The Strategy Test on build 610 was 26% slower with the long EMA's than with the short MA's.

But still a whole lot better than build 509 with the same long MA's:

1 Year
BackTest
1M Chart
MT4 build 509 Speed Test
Non Visual Mode
Using Test EA 509
Using iMA(300) and iMA(1000)
Test1
Test2
Test3
32.370 seconds
32.012 seconds
34.059 seconds
Average32.813 seconds

The new 600+ builds have been faster than 509 in all my strategy tester tests except in visual mode. 509 was faster in visual mode.

 

hi, i'm not using visual mode. my slowness is about x50 times, compared to 509. under 509, they are healthy and running.

i've some ea's that is good under 610, but they are not using M1, M5, M15, M30 etc. at the same time.

another slowness is coming from icustom ea's, in another ea work.

can your 610 shows indicators, after test, when you open chart? my 610 do not show related indicators on chart, after non-visual test, even period matches. oposite to 509.

 

Raptor,

I wanted to report back to you on this thread. You may remember I was getting different back-testing results with build 60x vs 509 with bots that appear to be trading ok in live with the new builds. Not too much of a problem for current/live set-ups but causing a complete halt to further development with the new build/tester and leaving us uncertain as to what the problem is/was.

Since the updating of builds has slowed and 610 has been the latest incarnation for over a week, I decided to upgrade to it on one of our test machines and re-run a like-for-like test that was previously giving us unexpected results. This time I immediately got the right/expected result. So I'm in the process of checking our other bots and different set-ups.

Whilst I have no idea which bug/issue was causing the problem, I thought it was worth letting you know that whatever it was, for us at least, appears to be fixed, provisionally. This is a big relief.

Reason: