Some signs of the right TCs - page 6

 
DARYIA SHAPOLOVA:
Your example, what example?! You were asked a question, what does it have to do with me?!

Dasha, you are so far - as Saber says. Not competent, i.e. buh....

 
onedollarusd:
It depends on who and where (through what program and with what settings) broadcasts these "connections" to you, whether the pattern will work) The "search" will have a limit - processing/execution time, response... Also, there's a strong dependence on volumes, whatever way you look at it. Imho.

Apparently, the not-so-good phrase "correct TS" obscures the understanding of the raised topic of the mathematical characteristics of the TS as a function of the price vector.

 
Fast235:

Saber, I can see the desperation in you,

I hope it seemed

It's strange to ask for your ideas from the public.

Voicing your thoughts on algotrading does not bring negativity, as it turns out.

 
Igor Makanu:

It's not the same at all. All I'm saying is that the TS must satisfy the conditions in the initial post of the thread.

The principle is that if you are trading a pattern, then the TS should not react to things that have nothing to do with the pattern.

 

fxsaber:

I would also add that evaluating the profitability/profitability of the TS is irrelevant to the topic. I'm sure that everyone who has risen well using algorithmic trading has used the wrong TS.

Exactly! And what for, then, those correct TS, if our goal is money? :)
 
trader_number_one:
Exactly! And what's the point of those correct TCs, if our goal is money? :)

A proper TS does not require manual adjustments. Moreover, if we have a wrong profitable TS, then this is a reason to examine your trading logic to see if there are any pieces in it that are in no way connected with the market pattern. That is, to try to get rid of particular features of the symbol.


The oblique language is off the charts, of course. Anyway, another incomprehensible topic started.

 

Somewhat reminiscent of the fundamental idea in mathematics and physics of invariance with respect to transformations (Galois theory, Lie groups, etc.)

As long as the TS is invariant, this is a very restrictive condition. For example, in the stock market the buy-and-hold strategy may work, but if we "reverse" the quotes, it will not work (we need to sell-and-hold). Provided that the TS also undergoes changes, everything looks quite trivial. But maybe - it's only at first glance and it's possible to find some errors that way.

I prefer the Monte Carlo approach to price modelling. A large number of series with characteristics similar to the initial series is formed on the basis of the initial price series and the TS behavior on them is checked. Something similar has already been mentioned in this thread. I do not insist on this approach, because the topicstarter quite sharply referred to this approach as an empty theorizing.

 
Correctly reviewed. If we can generate rows with "similar characteristics", it means that we know them very well (which is not the case in reality). And since we know them, we just sharpen the TC to them and don't suffer bullshit.
 
Similarity is an obvious consequence of not changing the original series too much (this can be done in a wide variety of ways). This is a standard method of at least some modelling of unpredictable future variability. But I won't even try to advocate this approach here, since it's mostly unconventional thinkers here.
 
This topic was duplicated on a popular exchange site (I will not give the link). There was an interesting dialogue on the subject. I didn't copy it here.