Where is the line between fitting and actual patterns? - page 33

New comment
 
Mathemat:
There's a great man called the Forex Director. He has been promoting these patterns for several years now, saying that it's not 50/50. But he seems to have never been able to explain how he does it :)

But how do these half-century parr calculations and lack of simple programming skills fit together?
 
joo:

Do you agree that it sounds very vague? I at least try to divide the TCs into two types, so that we can communicate in a meaningful way in the future. This is not a rebuke, but there are more questions than answers. Which TCs are more prone to "bad fit" and why? And other questions. And there are a lot of questions in general on this forum (the legendary forum of MTS developers, the place where the best minds on the planet gather - judging by the globality of the problems they solve, or at least try to) due to lack of clear and unambiguous interpretation of the terms, although often used, examples of misunderstanding each other are numerous and examples are in this thread.

It's vague, but it can't be otherwise, it's not an area where you can strictly divide everything into black and white. I actually understand the logic of such a division of types, but:

1. All this is very relative and pulled by the ear, the diversity of TC and principles used in them is much wider than their division into "good" and "bad" by trading time.

2. We, for our purposes, as far as I understood it, do not need this. I suggest to leave it to trader/developer to distinguish a "hopeless" TS from a TS capable of searching for patterns and using them. With some experience, it is easy enough to understand whether the TS has common sense or we are just playing with numbers. Let's consider that our TSs ;) are capable to discover regularities when used correctly. It is also suggested to consider that these regularities actually exist.

----

In the context of the branch I see our task as purely practical: to understand how to find with the highest probability a set of parameters, set, FC or whatever else, in which TS will use the patterns that it is built to use. This is relevant to me. There are 2 points here too.

а. How to train? (What to train on? How much to train on? etc.)

б. How to filter? (analysis of results).

----

So far there is no agreement :) Here are some people think, for example, that OOS steals profits, others have contrived to put OOS in the training sample, and as an explanation draw such schemes, that half a litre of cognac did not help me understand how you can use a control sample for training, that it has not lost its meaning. There is a surfeit of unsupported dissent in our ranks)

 
joo:
Mismatches may occur over time, but they are periodic, so they can be accounted for (knowing frequency, and knowing frequency means knowing duration, patterns for example, or the size of another TC indicator)
Not necessarily periodic, cyclic yes. For example have a planet where there are also 4 seasons summer, autumn, winter spring. But their duration varies and the sequence is constant. Winter may be a few days (in our calculation) or a few years for example. We do not know their duration in advance, but we know omens/signals that a change of season is coming - a change in temperature or the birds have flown in.) It is a non-periodic but cyclical process - we synchronize with it by folk or other signals. The duration of the season is not known in advance :)
 
Figar0: Some excess of unsupported logic dissent in our ranks)

Yeah, I have a third interpretation, different from Reshetov and lasso. And I think it's the right one, but let's not talk about it yet :) .

_____

By the way, as a fisherman with quite a long experience (summer exclusively) I have learned to predict thunderstorms in a few hours.

One more thing by the way -- several years in a row I've observed the picture -- pulling bins of freshly picked mushrooms on the bus in late January/February. That's never a reason to continue the mushroom discussion.

 
artmedia70:
And how do other currency pairs behave under each option? Is it worth looking there? And what candlestick combinations preceded the ones tested?
Well, that's the question... If there is a clear system, then something should be done about it. I was making statistics graphs for many symbols (about 50) and different timeframes. They are quite even. Accordingly it is possible to find out regularities in "good" times in correlations between symbols. My guess is that if some mechanism does it, it happens according to a certain algorithm.) Analysis is dead! Long live analysis!
 
Avals:
....The length of the season is not known in advance :)
I agree. But the duration of one season can't be longer than the duration of the whole cycle, can it? This means that in the limit the season is limited to one year (day, month, galactic day, etc., depending on what cyclicity we observe and investigate), this is a sign of the second type of TC - time limitation.
 
TheXpert:

I have a third interpretation, different from Reshetov and lasso. And I think it's the right one, but let's not talk about it yet :) .


It's a pity, we could do with a bit of constructiveness.)

***

And we should also think up some "single" publicly available adviser that we could "fit" into the glory, and which would be a measure of the different approaches to points a) and b). Otherwise we will drown in words. Does anyone have anything in mind? I'm behind the times, I haven't been here in almost a year.

 
Figar0:


It's a pity, we could do with a bit of constructiveness.)

***

And we should also come up with some publicly available advisor that we could "tinker" with and that would be a measure of the different approaches to a) and b). Otherwise we'll drown in words. Does anyone have anything in mind? I've been out of the loop for almost a year.

Profitable or what?
 
neama:


I confirm. somewhere in the code base there was even a script calculating the probabilities of candlestick patterns.

It worked out on a large enough 50/50 range with small offsets obviously not sufficient for building a TS.

i got quite a large 50/50 range with small offsets, clearly not enough to build the trader.

i.e. it may wager 10 times pocket and then 5 times profit.

It would be statistically reliable and normal. :)

On the other hand, there are a lot of variations in processing with Excel data. But there is still such an even correlation on instruments that are completely unrelated at first glance, that you start to believe in the "directors" of forex, nys and the like...
 
Figar0:

I see our task in the context of this thread as purely practical: to understand how with the highest probability to find a set of parameters, set, FN or whatever, at which the TS will use the patterns for which it is built. This is relevant to me. There are 2 points here too.

а. How to train? (What to train on? How much to train on? etc.)

б. How to filter? (analysis of results).

----

So far there is no agreement :) Here are some people think for example that OOS steals profits, others have contrived to put OOS in the training sample and as an explanation draw such schemes, that half a litre of cognac did not help me understand, how can the control sample can be used for training, that it would not lose its meaning. There is a surfeit of unsubstantiated dissent in our ranks)

Actually, in an attempt to answer these questions the division into two types was precisely conceived - we have the same goals.

Figar0:

And we should also think of some publicly available Expert Advisor that we could "tinker" with and that would be a measure of the different approaches to a) and b).

Yes, if anyone has some free time, maybe search the code base or write one and post it here?

let me remind you, for the sake of simplicity

1) Mashes and their many combinations (crossing)

2) typical representative of this type - candlestick analysis with decision making on each bar with forced time limit trade (and similar TS)


HH Who is a forex director?

1...262728293031323334353637383940...64
New comment
Reason: