Same Builder Config Produces vastly different generation rates depending where you load it?

See attached config. EURUSD H1 using tick data.



I have used this config to build strategies on the EURUSD for several months. It outputs around 300 accepted strategies per hour. This is very consistent, and very reliable.


I decided to use the same build config for the AUDUSD, EURGBP etc...however it produces only around 2-3 accepted strategies per hour. Odd. Spreads are different, but it shouldn't restrict the hourly output rate by that much, right?


So I did an experiment. Simply load the exact same config into the main builder. Same instrument, EURUSD, everything identical. Should obviously produce 300/hour right? Nope. 40 per hour.


How about a new project and load the config into the build there? Again, its identical. Everything, Only 40 per hour.


I cannot fathom this. Everything is absolutely identical. I even took screen shots of every setting screen to confirm everything was 100% identical. Yet strategies per hour is way down.


Is it possible that SQX saves the stats of builds along side projects, so if a project has been running for a couple of months, it has a pretty good idea which building blocks to focus on? Then if you load the same config anywhere else, those stats are reset?


I cannot understand the 300 per hour versus 40 per hour for the exact same build config on the same instrument. 



Attachments
Build strategies EURUSD H1 300 PER HOUR.cfx
(21.56 KiB)
  • Votes +1
  • Project StrategyQuant X
  • Type Bug
  • Status Fixed
  • Priority Normal

History

N
#1

nathan

06.03.2019 15:17

Task created

h
#2

hankeys

06.03.2019 20:59
you need to know that every building process will be different. there will be situations that you catch something that work and another time you will find nothing


secondly - you cant compare markets - EURUSD is the easiest

N
#3

nathan

06.03.2019 21:07
I agree fully. What threw me though and prompted me to post was that the absolute identical settings using the same instrument (EURUSD) using the same tick data and the same installation and version of SQX consistently produced 300 accepted strategies per hour in Project A, but in project B, C, D or main data builder this was always 40 per hour. Project A, B, C, D and main builder were all 100% identical in every way - same build settings, same instrument (EURUSD), same tick data. All were carbon copies. I purposely set them up as carbon copies to prove the variation. It is this that I cannot explain.
RL
#4

rickliao

10.03.2019 16:39
Voted for this task.
TT
#5

Tamas

11.05.2020 10:05

Status changed from New to Fixed


Votes: +1

Drop files to upload

or

choose files

Max size: 5MB

Not allowed: exe, msi, application, reg, php, js, htaccess, htpasswd, gitignore

...
Wait please