Algorithm Optimisation Championship. - page 105
![MQL5 - Language of trade strategies built-in the MetaTrader 5 client terminal](https://c.mql5.com/i/registerlandings/logo-2.png)
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
I did not manage to find comments in the post at the link above - practical use and examples of tasks.
Often we need to find the maximum and minimum values (extremes) of something. For example it is extremely important for scalpers to know trading conditions, e.g. max. and min. spread per timeframe at a particular broker.
The spread is determined by the market condition as well as the politics of a certain broker. What algorithm the broker uses is anybody's guess. Suppose the min. spread on the timeframe is determined by three main factors - max. and min. price and time of the bar H, L, T.Also, spread= f(H,L,T) is not given by the formula but by array spread= double[ H,L,T]. The task is to send to the FF (i.e. algorithm) such an array at which the FF is minimum. In fact, there are much more factors determining spread and they are constantly changing.
Often we need to find the maximum and minimum values (extremes) of something. For example it is crucial for scalpers to know the trading conditions, e.g. max. and min. spread per timeframe at a particular broker.
The spread is determined by the state of the market as well as by the policy of a certain broker. What algorithm the broker uses is anyone's guess. Suppose the min. spread on the timeframe is determined by three main factors - max. and min. price and time of the bar H, L, T.Also, spread= f(H,L,T) is not given by the formula but by array spread= double[ H,L,T]. The task is to send to the FF (i.e. algorithm) such an array at which the FF is minimum. In fact, there are much more factors determining spread and they are constantly changing.
Your examples are not convincing at all, one might even say that the examples are about nothing -- at the very least, where in your example:
-- the need to use some special algorithm to find an extremum?
The question of practicality and examples is open.
Your examples are not convincing at all -- at the very least, where in your example:
-- the need to use some special algorithm to find an extremum?
The question of practicality and examples is open.
The algorithm may be a classical one. But it must be fast and find extrema of unknown functions.
The algorithm can be a classical one. But it must be fast and find extrema of unknown functions.
Why does it have to be fast?
You speak of some abstraction in general terms.
Can you give an example of a concrete problem so that it becomes clear -- yes, a fast optimization algorithm, about which there have been debates for two months, such an algorithm is needed.
The question of practical use emerged from the beginning, as soon as the topicstarter started to boldly babble about his championship and his superpowered skill - but the question was ignored and we have been talking about some abstractions, which are incomprehensibly applicable to practical trading tasks.
Your examples are not convincing at all, one might even say that the examples are about nothing -- at the very least, where in your example
-- the need to use some special algorithm to find an extremum?
The question of practicality and examples of problems -- open.
The point is that the algorithm for finding optimal values does not necessarily have to search for the maximum of a function. It is the choice of the organizer.
Actually, the algorithm must search for the optimum values of the system properties (parameters) at which the system works stably. That is, the values of the very parameters which, according to the problem, must be passed in an array into the FF. Their values define the state of the system property, which is returned by the FF as a value.
The analytical function reflects the relationship between the environment parameters and the state of the system property.
Assuming that the system is stable at the maximum value of the analytic function, we should look for the maximum, but it is more likely that the best property state is not the peak value of the function, but an intermediate value.
Why does it have to be quick?
You're talking about some abstraction in general terms.
Can you give an example of a real problem, so that it becomes clear -- yes, the fast optimization algorithm, about which there have been two months of discussions, such an algorithm is needed.
The question of practical use - it arose initially, as soon as the topicstarter started to smartly proclaim about his championship and his super-deep expert - but the question was ignored and we have been discussing some abstractions, which are inapplicable to practical trading problems for two months.
Sometimes the code needs to make a split-second decision, and to do this it needs to optimize something fast. I would be glad to talk with you in depth. But right now I'm busy writing a program by one classical method.
The point is that the algorithm for finding the optimum values does not necessarily have to search for the maximum of the function. It is the organiser's choice.
...
GIVE ME AN EXAMPLE. Now it's interesting to get an understanding of the practical usefulness in trading of "optimization algorithms".
And "organizer" (ifAndrey Dik is meant) and his choice doesn't interest us at all. I strongly doubt his competence in this regard. He made a polemic for two months - the result and benefit is MINUS NULL.
GIVE ME AN EXAMPLE. Now it is interesting to understand the practical usefulness of "optimization algorithms" in trading.
And the "organizer" (ifAndrey Dik is meant) and his choice does not interest us at all. I strongly doubt his competence in this regard. He made a polemic for two months - the usefulness and result is MINUS ZERO.
Trading, of course, narrows down the scope of the algorithm.
I think it comes down to finding the values of trading strategy parameters that give the best trading results (the highest profitability) at the tested interval of recorded history.
Elementary adjustment of the existing parameters of the trader's trading strategy for maximum profit at previous trading sessions in the hope that specifics of a certain session will repeat in the future and these parameter values will be useful.
Sometimes a code needs to make a split-second decision, and to do that you need to optimise something quickly. I would love to talk to you in a non-abstract way. But at the moment I am busy writing a program using one classical method.