expert testing of strategies - page 2

New comment
 

Dear Figar0 The fact that you question the necessity of what is written on the forum and want to understand does you credit. As answers to your questions, let me ask mine:

Are you questioning

- that the quote chart you use is historically real (i.e. when modelling you use exactly those prices at which you would have traded in the past, because if not, how does such modelling differ from guessing by tarot cards) ? I personally sometimes observe the opposite...

- You do not have the source code of the platform, and you do not know how it works in different situations. Also, for example it is known that there is no quotation stream, but there is a BID quotation stream and ASK quotation stream, and therefore it matters in all kinds of intersection conditions. 3-4 points (spread between bid and ask) per trade - may cause significant deviation from reality. Besides, if price has not reached the specified level by 3 points, the level has not been crossed and vice versa, which affects the result. A simple example - if you use the BID chart, then the limit is triggered not only when the price has crossed the order line but also exceeds it by the value of spread, etc. Does your program take these nuances into account?

- The way your EA works? And what signals it gives to the platform coincide with what you would do manually?

If you answer all these questions without questioning the above, then you are an independent expert, and you will not be interested in my strip.

For KimIV - I will answer on Monday.

 
Artur, interesting questions you ask Figar0. They reminded me of my idea to test the MTS by swapping deals
Do you have something similar in mind or did I misremember?
 
Artur:

..... do you question.....

I have long removed rose-colored glasses, and my aspiration to get rich quick "degraded" in desire to steadily earn money). And of course on the final stage of development I subject my work to "tortures and torments", this includes testing of tester / demo (sometimes micro) with quotes of different brokerage companies, and simulation of requotes, and tool change, and restarts of terminal, and connection failure, but it is already almost paranoid. The final verdict is a normal account - minimum lot. And you are proposing to replace (or supplement?) this with something that you yourself still can't really describe, or want to?

Igor says the questions are interesting. But I think they are rather naive.

I do not question the need to test the expert for "stress resistance", I only doubt that the testing of expert's results by some "apparatus" with "unclear logic" (let's call it "BB"), can add me confidence in its reliability. What can be obtained from the input data "fed" to your BB? Well the profit factor, or the expected payoff, or the drawdown, or the average of testing interval, nothing that cannot be obtained within a couple of seconds using a calculator and its results. So then where does the "far-fetched" conclusions that this BB gives out come from?

I'm not AGAINST it, I'm FOR it!!!, it's just not clear how. And I don't think I'm the only one who doesn't understand.

 

It's all bullshit, there's nothing to get worked up about. The input information is too fluid to make even expert estimates based on it, let alone stable statistical ones. There is no information about the testing interval itself (not counting the duration), nor about the pair, nor about the distribution of trades over time. Can this be taken seriously?

P.S. Almost the only variant with a positive MO allowing drawing some definite conclusions about the strategy in general on the basis of this information is if the array of deals contains a very small percentage of deals that bring in the sum almost all profit. Then it is really difficult to consider the strategy stable.

 
I have a hunch - this whole BB - linear regression analysis - look angle and Pearson coefficient - just too little information for anything else.
 
Mathemat:

It's all bullshit, there's nothing to get worked up about. The input information is too fluid to make even expert estimates on its basis, let alone stable statistical ones. There is no information about the testing interval itself (not counting the duration), nor about the pair, nor about the distribution of trades over time. Can this be taken seriously?

P.S. Almost the only variant with a positive MO allowing drawing some definite conclusions about the strategy in general on the basis of this information is if the array of deals contains a very small percentage of deals that bring in the sum almost all profit. Then it is really difficult to consider the strategy stable.


I was wondering how long we will have to wait until someone "screw" this newcomer "guru" of automated trading together with his Perpetuum_Mobile :-)
 
Good day to you all !
I seem to have already been sent - banned from the forum..., had to start a new email and registration for this last post.

For Figar0, Mathemat, Itso, Xeon - you guys are all experienced people, all with higher education, programmers and mathematicians. No one here has tried to impose a dubious favour on you and certainly no one has pretended to be a guru. I am an equal participant here on this forum, just like everyone else. Honestly, that last remark, "Fuck the guru ... ..." even made me sad. In general, I do not understand how anyone can achieve anything, not even trying to understand new ideas or tools offered for free for testing by other participants. For some reason I'm more than sure that if you are offered to buy a miracle program for 500 quid, you will think about it, but you don't want to just take part in testing.

All I suggested, this is a joint testing of BB, so participants themselves will decide whether it is useful to them or not.

All who are interested in continuation please write to fxforum dog inbox dot ru. (Hackers please do not break the box, it is absolutely empty).
KimIV and Parabellum, write.

I give up the forum, not interested, and banned me (I am logged in, but I can not send messages). Like a forum type of intellectuals, and peasant aggression is somehow too much here. It was a hunt to break spears.
 
xeon:

I was wondering how long I would have to wait before someone "screwed" this new-found "guru" of autotrading, along with his Perpetuum_Mobile :-)
I can't do that, I recently read a book on anger management, and the psychoanalyst doesn't advise it ;)
 
Artuur:
Good day to you all !
I seem to have already been sent - banned from the forum..., had to start a new email and registration for this latest post.

You haven't been banned yet, but conclusions can already be drawn. Firstly, the forum has minimal protection against bots at the first level - you are allowed to leave a few posts without manual confirmation of registration by a moderator. That is, you haven't been confirmed yet, but you're already claiming a ban. Moderators are also resting on weekends, and you had the opportunity to leave multiple posts.

Secondly, you don't seem to understand MT4, but you offer to analyze something. All I see so far is disguised spam. Let's wait to see what forum members have to say.
 
xeon:
I was wondering how long I would have to wait before someone "sends" this new-fangled "guru" of autotrading, along with his Perpetuum_Mobile :-)
I thought the dispatch had already happened - subtle and intellectual, with a stretch (from a respected forum member). I even wrote the first reaction to that message, but quickly erased it, wanting to observe the denouement. It's true, though, that the first post in the thread looks a lot like spam.
1234567
New comment