MT5 terminal updated today and the "Optimisation" window does not show up during the test - page 17

New comment
 
Slava:

IndicatorRelease() in the tester didn't work before. does it work now?

 

@Slava

Problem: The button "Run Single Test" is inactive.

I'm doing optimization.

After optimization is complete, if I change dates (start and end of testing), then button becomes inactive: "Run Single Test".

The same situation is observed when I change the tick generation model.

Build: 1835.


This problem definitely did not exist a month ago!

Is this a bug or such an idea?

 
Evgeny Belyaev:

@Slava

Problem: The button "Run Single Test" is inactive.

I'm doing optimization.

After optimization is complete, if I change dates (start and end of testing), then button becomes inactive: "Run Single Test".

The same situation is observed when I change the tick generation model.

Build: 1835.


This problem definitely did not exist a month ago!

Is this a bug or such an idea?

I join in the question.
Put it back the way it was!

 
Сергей Таболин:
Now I'm waiting for the optimizer to run for a longer time. I'll see what happens.

Your case makes sense. Passes that fail are not written to the cache

FM      2       15:59:58.313    Core 5  pass 1396 tested with error "incorrect input parameters" in 0:00:00.218
KI      2       16:00:01.301    Core 8  pass 1664 tested with error "incorrect input parameters" in 0:00:00.188
MD      2       16:00:01.301    Core 8  pass 1665 tested with error "incorrect input parameters" in 0:00:00.187
EP      2       16:00:01.301    Core 8  pass 1666 tested with error "incorrect input parameters" in 0:00:00.203
RL      2       16:00:01.301    Core 8  pass 1667 tested with error "incorrect input parameters" in 0:00:00.203
LH      2       16:00:01.371    Core 3  pass 1413 tested with error "incorrect input parameters" in 0:00:00.187
EG      2       16:00:01.371    Core 3  pass 1414 tested with error "incorrect input parameters" in 0:00:00.187
KS      2       16:00:01.371    Core 3  pass 1415 tested with error "incorrect input parameters" in 0:00:00.188
PO      2       16:00:06.154    Core 4  pass 1556 tested with error "incorrect input parameters" in 0:00:00.218
LK      2       16:00:06.154    Core 4  pass 1557 tested with error "incorrect input parameters" in 0:00:00.219
DF      2       16:00:06.154    Core 4  pass 1558 tested with error "incorrect input parameters" in 0:00:00.203
RR      2       16:00:06.154    Core 4  pass 1559 tested with error "incorrect input parameters" in 0:00:00.219
MN      2       16:00:10.151    Core 6  pass 1812 tested with error "incorrect input parameters" in 0:00:00.203
GJ      2       16:00:10.151    Core 6  pass 1813 tested with error "incorrect input parameters" in 0:00:00.219
GI      2       16:00:10.151    Core 6  pass 1814 tested with error "incorrect input parameters" in 0:00:00.203
GE      2       16:00:10.637    Core 5  pass 1397 tested with error "incorrect input parameters" in 0:00:00.203

And only those passes which have gone through without errors are cached.

There is data in the cache at the same time

QK      0       16:05:34.655    Tester  cache file 'tester\cache\KrL_4.93.GBPCHF.M30.20170401.20170701.10.678BEE8FF38C7C7BF8515210F346F03E.opt' contains 441 records
HH      0       16:05:34.655    Tester  Experts\KrL_4.93.ex5 on GBPCHF,M30 from 2017.04.01 00:00 to 2017.07.01 00:00
RI      0       16:05:34.758    Tester  GBPCHF: history data begins from 2009.10.20 00:00
MP      0       16:05:34.758    Tester  GBPUSD: history data begins from 1998.01.02 00:00
PK      0       16:05:34.863    Tester  USDCHF: history data begins from 1998.01.02 00:00
LE      0       16:05:34.863    Tester  complete optimization started
RF      0       16:05:34.863    Tester  size of initial task batch is 512
MR      0       16:05:34.868    Tester  reading of 441 result records from cache...
QP      0       16:05:34.868    Tester  1 blocks of results read from cache in 0 ms
You have to think about what to do in cases like yours. The EA has failed on its own, not because of external factors and some errors
 
Slava:

If you only change the input parameters of the EA, a new cache is created. The previous cache is not deleted.

I do not change the input parameters, I change the optimisation parameters.
 
Dmitiry Ananiev:

I join in the question.
Put it back the way it was!

Back to

 
Slava:

Back to

is the 1835 build as slow as the 1816 or is it like the 17xx versions?

 
Slava:

Back to

Would you be so kind as to tell me.

Why did they remove the pause in"genetic algorithm" mode.

I suspect the pause caused problems when used on third party agents,

but for optimization on local agents, this feature was very useful.

 
Сергей Таболин:

I haveinput parameters check and output by INIT_PARAMETERS_INCORRECT there.

But the thing is, previous versions of optimizer didn't have such a reset.

Is it related to the revised cache?

And doesn't it turn out that all the "reset" passes go through a new one after the restart? As long as they are not written down?

Purely arithmetically:

If initially there were 1000 passes, stopping at 300, there are 700 left.

Suppose that INIT_PARAMETERS_INCORRECT had 100 passes.

After start, it started with 200. And 800 passes are left. We still need to pass the same 1000 passes.

It turns out we have 100 unnecessary passes. Although not very expensive, but still?

And as for me: out of 18760 passes, 1534 were "correct". It means that 17226 passes will be repeated! That's not very nice.

Yes. The reset passes are performed in a new way. In case of INIT_PARAMETERS_INCORRECT, these passes will fly by very quickly.

Think about how to mark such passes in the cache

 
Slava:

Yes. The reset passes are executed in a new way. In case of INIT_PARAMETERS_INCORRECT, these passes will fly by very quickly.

Think about how to mark such passes in the cache

And you can then return correct data about financial indicators at once - as zeros instead of crazy numbers?

1...10111213141516171819202122232425
New comment