Beta version of MetaTrader 4 IDE including new MQL4 compiler and editor - page 18
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
Indeed, I would argue.
And could you expand on the "you can't pick" idea ? just my opinion, in the MT5 strategy tester - you can pick literally anything, by any criteria...
1. Why not argue if you want to?
2. Let's not cut the context already [There is no way to make a rough selection from a large number of parameters.]
Ask for a selection of at least (ridiculously enough) a negligible number of 10000 parameters (at array size of 10 lakh, 10K is a negligible number), and you will answer your own question.
Indeed, I would argue.
Let me explain what I mean, and give you some background for the argument.
Take data storage format .avi, it is a compressed format for storing video, it is quite compact but does not allow you to achieve high quality.
The information stored in avi restorable, that is, from avi through processing can be extracted information on the 3D object, and paint Shtirlitz confirming that.
Now take our favorite bars. When slicing bars, compression is unrecoverable, hence the problem with building cagi, Renko, etc., I'm not talking about delta charts.
That's why all platforms have it easily implemented, but MQ has problems with such charts.
MQ has abandoned restorable tick compression and based the platform on non-restorable tick recording in the form of bars, thus cutting off a whole layer of research work and throwing its brainchild into the realm of kitchens and hamsters.
By the way, it also concerns MT4, but at least there is a custom history there that helps to make the situation smoother.
So we have a powerful trading platform with an excellent programming language, but we have no input data to process, about the same as an engine with 1024 horses, put on a bicycle.
Ask a selection of at least (ridiculously enough) a negligible number of 10000 parameters (at an array size of 10 lakh, 10K is a negligible number), and you will answer your own question.
I wonder... 10 thousand parameters - what kind of multi-parameter task is this ? Here, I think even more serious mathematical systems would have difficulty optimising such TS...
MQ, by abandoning recoverable tick compression and basing the platform on a non-recoverable tick record in the form of bars, has cut off a whole layer of research, and thrown its brainchild into the realm of kitchens and hamsters.
Debatable.
What kind of TS should be, that it would give a stable profit on ticks, and stably lost on ticks generation? In my opinion, generation of ticks in MT5 is quite adequate, and if TS uses ticks on generation - it will use it on real account.
Of course, this is an amateurish opinion, I've never tried to trade bellow M15, and now I'm inclined to buy D1 and even not to look lower than on H1... But, of course, I want to understand, what am I missing on ticks that is "cut off" during generation?
There's nothing controversial about "the first one". - The MT4 tester is considerably faster than the tester of its older colleague without the cloud.
And it is given by the modeling accuracy, but it is disputable that in MT5 the modeling accuracy is supposedly higher (this is exactly what is disputable),
Debatable.
But what kind of TS should be, that on ticks - it would give a stable profit, and on generation of ticks - stably lost? In my opinion, generation of ticks in MT5 is quite adequate, and if TS is losing on generation - it will lose on real account...
Of course, this is an amateurish opinion, I've never tried to trade bellow M15, and now I'm inclined to buy D1 and not to look lower than on H1... But, of course, I want to understand, what am I missing on ticks that is "cut off" during generation?
The tick generation is based on the scheme "a tick is a change of bid", but in fact many ticks happen without changes of bid, and often without any price changes (just changes of the block volume ).
But I'm not talking about generation of ticks, I'm talking about irreversible compression into bars. Simulate a numerical series consisting of several harmonics and you can easily predict it, because the series is deterministic, and now after each +-100 tick insert a false tick repeating the previous one (seems to be OK, a little extra noise), but the completely deterministic series will become unpredictable, because all frequencies will run.
If you have raw data, even noisy ones, you can reconstruct the original picture to a certain degree of quality, but if you have unrecoverable compression at your disposal, you can be happy with what you are shown and nothing more.
I gave above an example with avi, the restorable nature of the compression allows you to decompress the data and use a vector representation to describe 3D objects (which don't exist in a flat picture). But irreducible compression won't let you do that.
People in their masses don't notice how little things form the environment. Raising the temperature by just 1 degree is a problem, and recording ticks incorrectly is not a problem. Why not?
I feel it's not the only one. From the point of view of productivity, we see a very problematic use of functions of Copy* class (CopyBuffer, CopyRates, etc.). Saving on memory we dramatically reduce the performance, because the operations of data deployment into memory are very expensive, especially if they occur in giant loops, which we see in practice.
But I'm not talking about generating ticks, I'm talking about irreducible compression into bars. Simulate the numerical series by making it out of several harmonics and you can easily predict it, because the series is deterministic, but now after every +-100 tick insert a false tick repeating the previous one (it seems to be OK, a little extra noise), but the completely deterministic series becomes unpredictable, because all frequencies will float.
Hmmm... Really? It seemed to me, that one "extra" tick for hundred will just add low-amplitude harmonics, but it will hardly change the whole picture. Am I wrong? Not for the sake of objection, I am only superficially familiar with Fourier analysis, but still - why would the result be unpredictable?
People in their masses don't notice how little things form the environment. Raising the temperature by just 1 degree is a problem, and recording ticks incorrectly is not a problem. Why ?
Why ? It's 15 or 16 degrees outside - no difference.
Or is it about body temperature ? But then it is a very big difference as it should be measured from the temperature of the beginning of protein structure breakdown - 42 degrees... the difference, it turns out, is actually 20% - which is not a small difference.
But here I would like to have a substantive conversation with Stringo, I think only he can explain why MT4 tester is faster.
It's not faster than that. If you can prove your assertions, then provide us with comparative data.
And here I would like a substantive conversation with Stringo, I think only he can explain why the MT4 tester is faster.