ForumSections

Algorithm Optimisation Championship. - page 90

New comment
Andrey Dik:

What don't you like about this range?! What do you think it should be? And why can't someone else demand that it be changed in the same way?

You change your opinions and demands to the exact opposite almost every day.

I have explained that with protection, it can be freely accessed, without protection, only in the hands of a trusted person.

If it is freely accessible, who will make the table? You do not trust me, you do not trustKhatimlianskii either. What is your problem with testing algorithms without restrictions in the hands of a trusted tester?

Take it easy.

All right, let's have this range.

It's no big deal.

I have no reason to trust you or any other user I don't know.

I don't think this is a sign of paranoia. Just common sense.

However, if a person is trusted by the majority, I will trust him too.

All right!
Apart from me and Peter, would anyone else like to compete?
Andrey Dik:
OK!
Does anyone else want to compete besides me and Peter?

I think when all the technical issues are sorted out, some will come up.

Library issues with the FF, connection interface issues - all of that should be perfectly clear.

For example: is it possible to implement the solution in a script, or only in the library?

If only in a library, who will write the script that calls it, etc...

Реter Konow:

I think when all the technical issues are sorted out, people will come forward.

The questions of the library with the FF, questions of the connection interface - all this should be perfectly clear.

For example: is it possible to implement the solution in a script, or only in a library?

If only in the library, who will write the script that calls it, etc...

nothing has changed in the connection scheme, everything remains as before.
I will write the FF and the executing script, the source code will be publicly available. The FF will provide for the generation of a unique key, this should be done by the tester.
In a couple of days I will provide the source code of the FF and the script.
Andrey Dik:
Nothing has changed, everything remains as before.
I will write the FF and the executing script, the sources will be publicly available. the FF will include generation of a unique key, this should be done by the tester. after that, participants will hand in their algorithms to the tester and wait for the results.
In a couple of days I will provide the source code of the FF and the script.

Ok.

But the question of a candidate for the role of tester remains open.

Реter Konow:

OK.

But the question of a candidate for the role of tester remains open.

Maybe a tester won't be needed, I think I've figured out how to protect the algorithm while not limiting it in parameters.
Andrey Dik:
Maybe a tester won't be needed, I think I've figured out how to protect the algorithm without limiting the parameters.
Great, but do you still need a referee to generate the key?
Реter Konow:
Great, but do we still need a referee to generate the key?
Yes. Someone has to generate the key and compile the FF bible. This will ensure the same FF for all and at the same time the FF will not be known to the participants.

In general, I propose the following protection: the algorithm should check the date and check whether there is a connection to the broker's server(any).

In this way, a "trial" period can be set up for the testing period, e.g. 3 days from the day of publication. Thus, algorithms of all participants can be posted in the public domain and everyone can run a test on an FF prepared by a neutral person. In this case, there are no restrictions on FF parameters on the one hand, but on the other hand, it will provide a really open and transparent contest.

Are there ways to bypass such protection and use the algorithms after the trial period expires?

But still, someone will have to compile the test results into a table. I can do that. Anyone else can additionally compile a table of results and publish them.

Andrey Dik:

Are there ways to bypass such protection and use the algorithms after the trial period has expired?

Buy an MT server ;)

Bind yourself to MQ-Demo additionally.

1...838485868788899091929394959697...133
New comment