MetaTrader 5 Strategy Tester and MQL5 Cloud Network - page 3

To add comments, please log in or register
Jinsong Zhang
17686
Jinsong Zhang  

I do back testing with MQL5 cloud network, but some error "no history" occur, so if is the result of testing believable?

 

2011.06.04 14:06:49 MQL5 Cloud USA pass 38 tested with error "no history"
2011.06.04 14:05:59 MQL5 Cloud Cyprus pass 33 tested with error "no history"
2011.06.04 14:05:59 MQL5 Cloud Cyprus pass 20 tested with error "no history"
2011.06.04 14:05:59 MQL5 Cloud Cyprus pass 32 tested with error "no history"
2011.06.04 14:04:58 MQL5 Cloud USA pass 13 tested with error "no history"
2011.06.04 14:04:58 MQL5 Cloud USA pass 20 tested with error "no history"
2011.06.04 14:04:48 MQL5 Cloud Cyprus pass 20 tested with error "no history"
2011.06.04 14:04:48 MQL5 Cloud Cyprus pass 19 tested with error "no history"
2011.06.04 14:04:48 MQL5 Cloud Cyprus pass 22 tested with error "no history"
2011.06.04 14:03:52 MQL5 Cloud Cyprus pass 23 tested with error "no history"
2011.06.04 14:03:50 MQL5 Cloud Cyprus pass 22 tested with error "no history"
2011.06.04 14:03:49 MQL5 Cloud Cyprus pass 21 tested with error "no history"
2011.06.04 14:03:48 MQL5 Cloud Cyprus pass 20 tested with error "no history"
2011.06.04 14:03:48 MQL5 Cloud Cyprus pass 19 tested with error "no history"
2011.06.04 14:03:42 MQL5 Cloud USA pass 17 tested with error "no history"
2011.06.04 14:03:42 MQL5 Cloud USA pass 15 tested with error "no history"
2011.06.04 14:03:42 MQL5 Cloud USA pass 13 tested with error "no history"
2011.06.04 14:03:41 MQL5 Cloud USA pass 12 tested with error "no history"
2011.06.04 14:03:30 MQL5 Cloud Hong Kong GBPUSD: history synchronization completed [8 Kb]
2011.06.04 14:03:30 MQL5 Cloud Hong Kong GBPUSD: history for 2011 year synchronized
2011.06.04 14:03:12 MQL5 Cloud Europe GBPUSD: history synchronization completed [837 Kb]
2011.06.04 14:03:12 MQL5 Cloud Europe GBPUSD: history for 2011 year synchronized
2011.06.04 14:03:12 MQL5 Cloud Europe GBPUSD: history for 2010 year synchronized
2011.06.04 14:03:07 MQL5 Cloud USA common synchronization completed
2011.06.04 14:03:06 MQL5 Cloud Europe common synchronization completed
2011.06.04 14:03:06 MQL5 Cloud Cyprus common synchronization completed
2011.06.04 14:03:06 MQL5 Cloud Hong Kong common synchronization completed
2011.06.04 14:03:03 Core 1 GBPUSD: history synchronization completed [7 Kb]
2011.06.04 14:03:03 Core 1 GBPUSD: history for 2010 year synchronized
2011.06.04 14:03:03 MQL5 Cloud Singapore common synchronization completed
2011.06.04 14:03:03 Core 1 common synchronization completed
2011.06.04 14:03:03 Core 1 authorized (agent build 463)
2011.06.04 14:03:03 Core 1 connected
2011.06.04 14:03:01 MQL5 Cloud Cyprus authorized (agent build 463)
2011.06.04 14:03:01 MQL5 Cloud Europe authorized (agent build 463)
2011.06.04 14:03:01 MQL5 Cloud USA authorized (agent build 463)
2011.06.04 14:03:00 MQL5 Cloud Singapore authorized (agent build 463)
2011.06.04 14:03:00 MQL5 Cloud Cyprus connected
2011.06.04 14:03:00 MQL5 Cloud Europe connected
2011.06.04 14:03:00 MQL5 Cloud USA connected
2011.06.04 14:03:00 MQL5 Cloud Hong Kong authorized (agent build 463)
2011.06.04 14:03:00 MQL5 Cloud Singapore connected
2011.06.04 14:02:59 MQL5 Cloud Hong Kong connected
2011.06.04 14:02:59 MQL5 Cloud USA connecting to 2.agents.mql5.com:443
2011.06.04 14:02:59 MQL5 Cloud Singapore connecting to 5.agents.mql5.com:443
2011.06.04 14:02:59 MQL5 Cloud Hong Kong connecting to 4.agents.mql5.com:443
2011.06.04 14:02:59 MQL5 Cloud Europe connecting to 1.agents.mql5.com:443
2011.06.04 14:02:59 MQL5 Cloud Cyprus connecting to 3.agents.mql5.com:443
2011.06.04 14:02:59 Core 1 connecting to 127.0.0.1:3000
2011.06.04 14:02:59 Core 1 agent process started
2011.06.04 14:02:59 Tester size of initial task pack is 6
2011.06.04 14:02:59 Tester GBPUSD: preliminary downloading of history completed
2011.06.04 14:02:58 Tester GBPUSD: preliminary downloading of history started
2011.06.04 14:02:58 Tester complete optimization started
2011.06.04 14:02:58 Tester Experts\song4compete2011_9_Optimized.ex5 on GBPUSD,M5 from 2010.06.01 00:00 to 2010.12.31 00:00
2011.06.04 14:02:58 Tester genetic mode turned off due to 40 passes, to use it increase amount of optimized parameters
 

Jinsong Zhang
17686
Jinsong Zhang  
A problem of cloud network must be solve, 4 of the 5 networks finished their task but one of them is frozen, so it spend more time than not use cloud network.
MetaQuotes
Admin
28210
Renat Fatkhullin  
song_song:

I do back testing with MQL5 cloud network, but some error "no history" occur, so if is the result of testing believable?

MQL5 Cloud Network is not ready yet.

Franco
361
Franco  
Renat:

MetaTrader 5 client terminal is working only with own servers and their data. All detailed history (and M1 spreads) provided by trading server. No need to change chart history.

You can control quality of historical data from 'Results' page:

  • History Quality — this value characterizes the quality of price data used for testing. It is determined as a percentage ratio of correct and incorrect minute data. Incorrect minute bars are the ones having zero spread, or volume 1 when OHLC values are different. History gaps are also considered as incorrect data. Depending on size, the period of testing is divided into 1 — 199 intervals. Then the history quality is determined for each of them separately. Those time intervals are shown on the graphical indicator of the history quality (the lighter tint of green is the better is the quality, the red color represents intervals with the quality lower than 50%).

Hey Renat,

Thanks for the reply.  This is probably not the thread for this discussion but I will ask anyways:

Does the "History Quality" test for gaps in the history?  The last time I used MT5 was when I detected a 6 month gap in the GBPUSD data and I refused to use the tester further until a history import function is implemented.

The problem with using only one set of data is different brokers have different feeds, and because I don't trust the MQL data.  On timeframes < 1H this is a huge problem, and a strategy should be tested on different sets of data to test for any broker dependencies that the strategy might have, therefore a history center such as available in MT4 is a neccesity for any data mining application.

MetaQuotes
Admin
28210
Renat Fatkhullin  
Saidar:

Does the "History Quality" test for gaps in the history?  The last time I used MT5 was when I detected a 6 month gap in the GBPUSD data and I refused to use the tester further until a history import function is implemented.

  • History Quality — this value characterizes the quality of price data used for testing. It is determined as a percentage ratio of correct and incorrect minute data. Incorrect minute bars are the ones having zero spread, or volume 1 when OHLC values are different. History gaps are also considered as incorrect data. Depending on size, the period of testing is divided into 1 — 199 intervals. Then the history quality is determined for each of them separately. Those time intervals are shown on the graphical indicator of the history quality (the lighter tint of green is the better is the quality, the red color represents intervals with the quality lower than 50%).
Documentation on MQL5: Standard Constants, Enumerations and Structures / Objects Constants / Web Colors
  • www.mql5.com
Standard Constants, Enumerations and Structures / Objects Constants / Web Colors - Documentation on MQL5
Franco
361
Franco  
Renat:

  • History Quality — this value characterizes the quality of price data used for testing. It is determined as a percentage ratio of correct and incorrect minute data. Incorrect minute bars are the ones having zero spread, or volume 1 when OHLC values are different. History gaps are also considered as incorrect data. Depending on size, the period of testing is divided into 1 — 199 intervals. Then the history quality is determined for each of them separately. Those time intervals are shown on the graphical indicator of the history quality (the lighter tint of green is the better is the quality, the red color represents intervals with the quality lower than 50%).
Must have missed that part sorry for that :) 
supercomputer
55
supercomputer  

I see my MetaTester Agents Manager have some weird statistic...

 Local Statistic:

Test passes: 6596

Time taken: 7 hours 50 minutes

MQL5 Cloud Network Statistic:

Test passes: 17

Time taken: 0 hours 0 minutes

 I don't run optimization recently and I believe someone have access to my agents for testing. Could I suggest add a alert when metatester agent is used by third party!!!

MetaQuotes
Admin
28210
Renat Fatkhullin  
supercomputer:

I don't run optimization recently and I believe someone have access to my agents for testing. Could I suggest add a alert when metatester agent is used by third party!!!
Don't worry, we are testing MQL5 Cloud Network.
techsupport_
29
techsupport_  
is it possibe to anyone try the cloud network?

jamesw
6
jamesw  

This is indeed a big step as cloud computing has been the latest advancement in the field, though a lot would find it a little vague looking at its description.

I am intrigued how this would affect local developers who thrive on producing and marketing on client specific modules rather than a centralized function distributed across networks.
To add comments, please log in or register